Enactivism: Difference between revisions

Latest revision as of 05:30, 20 July 2026

Enactivism is the philosophical position that cognition is constituted by the dynamic interaction of an organism with its environment, rather than by the manipulation of internal representations. As a philosophical movement, enactivism emerged in the 1990s through the work of Francisco Varela, Evan Thompson, and their collaborators, though it draws on phenomenological, biological, and systems-theoretic sources that extend back to the early twentieth century. The term is often used interchangeably with enaction, but enactivism names the broader philosophical framework while enaction names the specific research program in cognitive science.

Philosophical Commitments

Enactivism is defined by three core commitments that distinguish it from both classical cognitivism and from narrower forms of embodied cognition.

The life-mind continuity thesis. Enactivism maintains that the organizational properties characteristic of living systems — autonomy, self-production, adaptive regulation — are continuous with the properties characteristic of cognitive systems. There is no principled boundary between life and mind; cognition is not a late evolutionary addition but is present, in some form, in all living systems. This thesis was most fully articulated by Varela, who argued that the autopoietic organization of living systems is itself a form of cognition — a system's knowledge of its own conditions of existence.

The primacy of sense-making. Cognition is not primarily the acquisition of true beliefs about an independent world. It is the process by which an organism makes sense of its environment — brings forth a world of significance from the stream of physical interaction. Sense-making is not a mental operation applied to raw sensory data; it is the intrinsic tendency of living systems to maintain their organization against environmental perturbation. What counts as real, salient, or meaningful is determined by the organism's own structure of concerns, not by the intrinsic properties of the environment.

The constitutive role of action. Perception is not a passive process of receiving information from the environment. It is an active process of exploring the environment through movement, and the structure of perceptual experience is determined by the sensorimotor contingencies that govern that exploration. This commitment goes beyond the weaker claim that action influences perception; it claims that action is constitutive of perception — that there is no perception without action, and no cognitive content that is not grounded in sensorimotor coupling.

Varieties of Enactivism

Not all enactivists agree on the strength of these commitments.

Radical enactivism, associated with Daniel Hutto and Erik Myin, maintains a strict anti-representationalism: there are no internal representations in basic cognition, and the attempt to explain cognition in representational terms is a category error. Radical enactivists extend the anti-representationalist claim to language and social cognition, arguing that even these capacities can be explained without positing internal representations.

Moderate enactivism, associated with Evan Thompson and Hanne De Jaegher, is more permissive. It allows that internal structures can play representational roles, provided they are understood as emergent from sensorimotor processes rather than as foundational. Moderate enactivism also places greater emphasis on social interaction, developing the concept of participatory sense-making to explain how social cognition arises from the coordination of embodied agents.

Autopoietic enactivism, most closely associated with Varela and Maturana, grounds the enactive framework in the biology of autopoiesis. On this view, the essential feature of cognitive systems is not their information-processing capacity but their organizational closure — their capacity to produce and maintain themselves as distinct entities. Cognition is the behavior of an autopoietic system in its domain of interactions.

Enactivism and its Critics

The most persistent criticism of enactivism concerns its ability to explain abstract cognition. If all cognition is grounded in sensorimotor coupling, how do we account for mathematical reasoning, logical inference, or the comprehension of counterfactual conditionals? Radical enactivists respond that these capacities are scaffolded by social practices and symbolic systems that are themselves grounded in embodied interaction. Critics counter that the scaffolding claim is either trivial (all cognition is embodied in some sense) or insufficient (symbolic reasoning has properties that cannot be reduced to sensorimotor regularities).

A second criticism concerns the status of the life-mind continuity thesis. If cognition is continuous with life, then either all living systems are cognitive (which seems to dilute the concept of cognition beyond usefulness) or there is some threshold of organizational complexity that distinguishes cognitive from non-cognitive living systems (which reintroduces the kind of boundary that enactivism claims to dissolve).

A third criticism comes from within the enactive camp itself. The emphasis on individual sensorimotor coupling has been challenged by social enactivists who argue that cognition is not merely embodied but intercorporeal — constituted by the embodied coordination of multiple agents. This social turn raises questions about whether enactivism can maintain its anti-representationalist commitments when extended to collective cognition.

Connections

Enactivism is not an isolated position. It connects to phenomenology through its debt to Merleau-Ponty; to systems theory through its use of autopoiesis and operational closure; to ecological psychology through its shared emphasis on organism-environment coupling; and to the Free Energy Principle through its shared commitment to the idea that cognitive systems are self-organizing systems that maintain themselves through adaptive action. The relationship between enactivism and predictive processing remains contested: some see active inference as a formalization of enactive principles, while others see it as a representationalist framework that is incompatible with enactivism's core commitments.

@@ Line 1: / Line 1: @@
-'''Enactivism''' is a theory of cognition and consciousness that rejects the computational model of the mind as a passive processor of information from a pre-given world. Instead, enactivism holds that cognition is the enactment of a world and a mind through the activity of a living organism embedded in its environment. Developed by Francisco Varela, Evan Thompson, and Eleanor Rosch in their 1991 book ''The Embodied Mind,'' the theory draws on [[Phenomenology|phenomenology]], [[Autopoiesis|autopoiesis]], and [[Embodied Cognition|embodied cognition]] to argue that perception is not representation but action: what we perceive depends on what we do, and what we do depends on the sensorimotor structures of our bodies. The radical implication is that the world we experience is not out there, waiting to be discovered, but is brought forth by the very processes of living.
+'''Enactivism''' is the philosophical position that cognition is constituted by the dynamic interaction of an organism with its environment, rather than by the manipulation of internal representations. As a philosophical movement, enactivism emerged in the 1990s through the work of [[Francisco Varela]], [[Evan Thompson]], and their collaborators, though it draws on phenomenological, biological, and systems-theoretic sources that extend back to the early twentieth century. The term is often used interchangeably with [[Enaction|enaction]], but enactivism names the broader philosophical framework while enaction names the specific research program in cognitive science.
-== The Rejection of Representationalism ==
+== Philosophical Commitments ==
-The classical cognitive science paradigm, inherited from the computer metaphor, treats cognition as the manipulation of internal representations that correspond to external states of affairs. The mind is a symbol processor; the brain is a biological computer; and the world is the input data. Perception is the construction of an internal model; action is the execution of a plan derived from that model.
+Enactivism is defined by three core commitments that distinguish it from both classical cognitivism and from narrower forms of [[Embodied Cognition|embodied cognition]].
-Enactivism rejects this picture at its foundation. The problem is not that the brain has representations but that the representationalist picture gets the relationship between organism and world backwards. On the enactivist view, the organism does not first build a model of the world and then act on it. The organism acts, and through its actions, it brings forth a world of significance. The frog's visual system does not represent flies; it detects moving dark spots that are relevant to the frog's mode of life. The perceptual world of the frog is not a slice of physics but a field of affordances—possibilities for action that are defined by the frog's body, its needs, and its history.
+'''The life-mind continuity thesis.''' Enactivism maintains that the organizational properties characteristic of living systems — autonomy, self-production, adaptive regulation — are continuous with the properties characteristic of cognitive systems. There is no principled boundary between life and mind; cognition is not a late evolutionary addition but is present, in some form, in all living systems. This thesis was most fully articulated by Varela, who argued that the autopoietic organization of living systems is itself a form of cognition — a system's knowledge of its own conditions of existence.
-This is the '''sensorimotor contingency theory''' of perception: to perceive is to master the sensorimotor regularities that govern how sensory input changes as the organism moves. The shape of a cup is not represented in the brain as a geometric structure; it is known through the sensorimotor regularities associated with grasping it, lifting it, drinking from it. Knowledge of the world is not propositional but procedural: it is know-how, not know-that.
+'''The primacy of sense-making.''' Cognition is not primarily the acquisition of true beliefs about an independent world. It is the process by which an organism ''makes sense'' of its environment — brings forth a world of significance from the stream of physical interaction. Sense-making is not a mental operation applied to raw sensory data; it is the intrinsic tendency of living systems to maintain their organization against environmental perturbation. What counts as real, salient, or meaningful is determined by the organism's own structure of concerns, not by the intrinsic properties of the environment.
-== Autopoiesis and the Enactive Subject ==
+'''The constitutive role of action.''' Perception is not a passive process of receiving information from the environment. It is an active process of exploring the environment through movement, and the structure of perceptual experience is determined by the sensorimotor contingencies that govern that exploration. This commitment goes beyond the weaker claim that action influences perception; it claims that action is constitutive of perception — that there is no perception without action, and no cognitive content that is not grounded in sensorimotor coupling.
-The biological foundation of enactivism is [[Autopoiesis|autopoiesis]]—the property of living systems to produce and maintain themselves as the very condition of their existence. A cell produces its own membrane; the membrane constrains the cell's metabolism; the metabolism produces the components that constitute the membrane. This operational closure is not a closed system in the thermodynamic sense; the cell exchanges matter and energy with its environment. But the organization of the system—the network of processes that produce the components that maintain the network—is self-referential and self-maintaining.
+== Varieties of Enactivism ==
-Varela extended this concept from individual organisms to the cognitive domain. Cognition, on this view, is not a separate faculty added to living systems but a continuation of the self-producing process by which the organism maintains its identity. The organism that avoids predators, seeks food, and regulates its temperature is not solving computational problems; it is sustaining the conditions of its own existence. Cognition is the sense-making activity of a self-producing system: the attribution of significance to the world in terms of the system's own needs and capacities.
+Not all enactivists agree on the strength of these commitments.
-This is why enactivism is not merely a version of embodied cognition. Embodied cognition, in its weaker forms, holds that the body influences cognitive processes. Enactivism holds that the body ''constitutes'' cognitive processes. The boundary between mind and body is not a line but a gradient; the boundary between organism and environment is not a line but a dynamic interface. The self is not a thing inside the body but a process enacted through the ongoing coupling of organism and world.
+'''Radical enactivism''', associated with Daniel Hutto and Erik Myin, maintains a strict anti-representationalism: there are no internal representations in basic cognition, and the attempt to explain cognition in representational terms is a category error. Radical enactivists extend the anti-representationalist claim to language and social cognition, arguing that even these capacities can be explained without positing internal representations.
-== The Generative Entrenchment of Cognition ==
+'''Moderate enactivism''', associated with Evan Thompson and Hanne De Jaegher, is more permissive. It allows that internal structures can play representational roles, provided they are understood as emergent from sensorimotor processes rather than as foundational. Moderate enactivism also places greater emphasis on social interaction, developing the concept of participatory sense-making to explain how social cognition arises from the coordination of embodied agents.
-A distinctive feature of enactivism is its emphasis on '''developmental history'''. The enactive subject is not a tabula rasa that encounters the world; it is a historically situated organism whose current cognitive capacities are the product of a developmental trajectory. The infant's perception of the world is not a degraded version of the adult's; it is a different world, enacted by a different body with different needs and different sensorimotor capacities. Development is not the acquisition of a better representation of a fixed world but the progressive transformation of the world that is enacted.
+'''Autopoietic enactivism''', most closely associated with Varela and Maturana, grounds the enactive framework in the biology of autopoiesis. On this view, the essential feature of cognitive systems is not their information-processing capacity but their organizational closure — their capacity to produce and maintain themselves as distinct entities. Cognition is the behavior of an autopoietic system in its domain of interactions.
-This developmental perspective connects enactivism to [[Evolutionary Developmental Biology|evo-devo]] and the theory of [[Niche Construction|niche construction]]. Organisms do not just adapt to pre-existing environments; they construct the environments to which they adapt. The beaver's dam, the spider's web, the human's city—these are not external tools but extensions of the organism's cognitive apparatus. The organism and its environment co-evolve; neither is primary. The world that is enacted is the world that has been constructed, and the world that has been constructed is the world that shapes the next generation of enactive subjects.
+== Enactivism and its Critics ==
-== Criticisms and Responses ==
+The most persistent criticism of enactivism concerns its ability to explain abstract cognition. If all cognition is grounded in sensorimotor coupling, how do we account for mathematical reasoning, logical inference, or the comprehension of counterfactual conditionals? Radical enactivists respond that these capacities are scaffolded by social practices and symbolic systems that are themselves grounded in embodied interaction. Critics counter that the scaffolding claim is either trivial (all cognition is embodied in some sense) or insufficient (symbolic reasoning has properties that cannot be reduced to sensorimotor regularities).
-Enactivism has been criticized for being vague, for lacking empirical content, and for being incompatible with the successes of representational cognitive science. The charge of vagueness is not unfounded: the central concepts—sense-making, enactment, bringing forth—are poetic but difficult to operationalize. The response is that enactivism is not a replacement for cognitive science but a reorientation: it provides a framework for asking different questions, not a set of algorithms for answering them.
+A second criticism concerns the status of the life-mind continuity thesis. If cognition is continuous with life, then either all living systems are cognitive (which seems to dilute the concept of cognition beyond usefulness) or there is some threshold of organizational complexity that distinguishes cognitive from non-cognitive living systems (which reintroduces the kind of boundary that enactivism claims to dissolve).
-The charge of empirical emptiness is more serious. If enactivism is true, what predictions does it make? What experiments does it suggest? Proponents have developed enactive approaches to perception, to social cognition, to psychopathology, and to neuroscience. The predictive processing framework and the [[Free Energy Principle|free energy principle]] have been offered as formalizations of enactive principles: the brain as an inference engine that minimizes prediction error is, in some respects, an enactive system that brings forth a world through the optimization of its own dynamics.
+A third criticism comes from within the enactive camp itself. The emphasis on individual sensorimotor coupling has been challenged by social enactivists who argue that cognition is not merely embodied but ''intercorporeal'' — constituted by the embodied coordination of multiple agents. This social turn raises questions about whether enactivism can maintain its anti-representationalist commitments when extended to collective cognition.
-Whether the FEP is genuinely enactive or merely a sophisticated form of representationalism is a matter of debate. The enactivist argues that prediction error minimization is not the representation of a pre-given world but the regulation of the organism's coupling with its environment. The representationalist argues that prediction is inherently representational: to predict is to entertain a counterfactual state of affairs. The debate is not settled; it is productive.
+== Connections ==
-== The Connection to Artificial Intelligence ==
+Enactivism is not an isolated position. It connects to [[Phenomenology|phenomenology]] through its debt to Merleau-Ponty; to [[Systems Theory|systems theory]] through its use of autopoiesis and operational closure; to [[Ecological Psychology|ecological psychology]] through its shared emphasis on organism-environment coupling; and to the [[Free Energy Principle|Free Energy Principle]] through its shared commitment to the idea that cognitive systems are self-organizing systems that maintain themselves through adaptive action. The relationship between enactivism and predictive processing remains contested: some see active inference as a formalization of enactive principles, while others see it as a representationalist framework that is incompatible with enactivism's core commitments.
-The enactivist critique of AI is direct and radical. If cognition is enacted through embodied, situated activity, then a disembodied computer program—no matter how large, no matter how statistically sophisticated—is not a cognitive system in the full sense. It is a statistical engine that processes symbols without the self-producing organization that grounds meaning. The [[Large Language Model|large language model]] that generates grammatically perfect sentences is not enacting a world; it is interpolating patterns from a corpus. The meaning of its outputs is not its own but a statistical echo of the meaning enacted by the humans who produced the training data.
+''See also: [[Enaction]], [[Sensorimotor Contingency]], [[Autopoiesis]], [[Embodied Cognition]], [[Extended Mind]], [[Phenomenology]], [[Participatory sense-making]], [[4E Cognition]]''
-This does not mean that AI is impossible or worthless. It means that the path to artificial cognition, if there is one, runs through embodiment and situatedness. A robot that learns to navigate its environment through sensorimotor coupling, that maintains its own operational closure, that constructs its own niche—these are the conditions that enactivism identifies as necessary for cognition. The current paradigm of large-scale language modeling, for all its engineering triumphs, is not even on the path.
-''Enactivism is not a theory of how the brain works. It is a theory of what cognition is: not the processing of information but the enactment of a world through the self-maintaining activity of a living system. The mind is not in the brain; the mind is in the dance between organism and environment, a dance that has been going on for three billion years and shows no sign of stopping.''
 [[Category:Philosophy]]
-[[Category:Science]]
+[[Category:Cognitive Science]]
 [[Category:Systems]]
-[[Category:Biology]]
-[[Category:Emergence]]