Loading field map
For much of the twentieth century, the dominant picture in philosophy of mind treated cognition as a computational process running on neural hardware. The body was a peripheral input-output device; the environment was a source of sensory data. Embodied cognition emerged in the early 1990s as a direct challenge to that picture, arguing that the body and the world are not merely inputs to a central processor but are constitutive of cognition itself. This shift raised a new set of questions: Where does the mind end and the world begin? What role do sensorimotor capacities play in thinking? And can cognition be understood without appealing to internal representations? Four frameworks have shaped the answers to these questions, each with its own emphasis and internal tensions.
In 1991, two landmark works appeared within months of each other: Francisco Varela, Evan Thompson, and Eleanor Rosch's The Embodied Mind and Jean Lave's Situated Learning. Though independent, both drew on converging critiques of classical computationalism, drawing inspiration from phenomenology (especially Merleau-Ponty), ecological psychology (Gibson), and robotics (Brooks). Together they launched the first wave of embodied cognition.
Enactivism holds that cognition is enacted through the sensorimotor coupling of an organism with its environment. A living being does not passively receive information and then compute a response; instead, it brings forth a world through its actions. Perception is not the construction of an internal model but a skill of guided exploration. Enactivism is explicitly anti-representational: it denies that cognition requires internal representations that stand in for absent states of affairs. The mind is not a mirror of nature but a dynamic activity of sense-making.
Situated Cognition shares the rejection of the disembodied, decontextualized view of mind but places greater emphasis on the social and material scaffolding that supports cognitive activity. Cognition is not just embodied but embedded in a physical and cultural environment that shapes what counts as a problem and what tools are available for solving it. Rodney Brooks's subsumption architecture in robotics—where intelligent behavior emerges from simple, layered control systems without central representation—became a powerful illustration. Later, Edwin Hutchins's work on distributed cognition showed how cognitive processes can be spread across individuals, artifacts, and social practices. Over time, Situated Cognition was largely absorbed into the broader traditions of distributed and embedded cognition, which treat the environment as a genuine part of the cognitive system rather than a mere context.
Enactivism and Situated Cognition thus co-emerged with overlapping concerns but different emphases. Enactivism focused on the organism's lived, sensorimotor engagement with the world; Situated Cognition focused on the external structures—tools, language, social norms—that enable and constrain thought. Both rejected the computationalist assumption that cognition is a purely internal, representational process, but they did so from different angles.
Seven years later, Andy Clark and David Chalmers published "The Extended Mind," which pushed the boundary question in a new direction. Their parity principle states that if a part of the world functions as a process that, were it done in the head, we would have no hesitation in recognizing as part of the cognitive process, then that part of the world is part of the cognitive process. Otto's notebook, used to store beliefs about museum locations, is as much a part of his mind as a biological memory trace.
The Extended Mind Thesis (EMT) shares with Enactivism the idea that cognition can extend beyond the skull. But it differs sharply on representation. EMT retains a representationalist framework: Otto's notebook contains internal representations (written symbols) that play the same functional role as neural representations. Enactivism, by contrast, denies that cognition is fundamentally representational at all. This disagreement has been a central fault line. Proponents of EMT argue that the parity principle offers a clear criterion for cognitive extension without abandoning the explanatory power of representations. Critics from the enactivist camp reply that the parity principle smuggles in a computationalist view of mind, treating cognition as the manipulation of symbols regardless of their medium.
EMT also differs from Situated Cognition in its focus on the individual's cognitive system rather than on distributed social practices. While Situated Cognition emphasizes how environments shape cognition, EMT argues that the environment can literally become part of the cognitive agent. The two frameworks are compatible in practice—many researchers draw on both—but they make different claims about where the boundaries of the mind lie.
Running alongside these philosophical debates, Dynamical Systems Theory (DST) emerged in the early 1990s as a formal mathematical approach to cognition. Drawing on work in developmental psychology (Esther Thelen, Linda Smith) and nonlinear dynamics, DST models cognitive processes as trajectories through a state space, governed by attractors and control parameters. A reaching movement, a developmental stage transition, or a perceptual decision can be described without invoking representations or discrete computational steps.
DST is not a competing philosophical thesis about the nature of mind but a methodological framework that offers a different explanatory style. Where Enactivism, Situated Cognition, and EMT argue about what cognition is, DST provides tools for modeling how cognition unfolds in real time. It is compatible with all three in principle, but in practice it has been most closely associated with anti-representationalist approaches. Because DST describes cognition as a continuous, self-organizing process, it fits naturally with Enactivism's emphasis on sensorimotor coupling and with Situated Cognition's focus on the dynamics of agent-environment interaction. It is less easily reconciled with EMT's representationalist commitments, though some have attempted to combine them.
The four frameworks are not a settled consensus but an ongoing conversation structured by three persistent disagreements.
Representation. The deepest divide is between those who think cognition requires internal representations (EMT, many versions of Situated Cognition) and those who think it does not (Enactivism, DST when used anti-representationally). This is not a minor technical dispute; it concerns what kind of explanation is appropriate for cognitive phenomena. Enactivists argue that representational explanations are a relic of computationalism; representationalists reply that without representations, it is impossible to explain planning, memory, or abstract thought.
Boundaries. Where does the mind end? EMT draws a functional boundary: anything that plays the right causal role is part of the mind. Enactivism draws a biological boundary: cognition is the sense-making activity of a living organism, and while it is shaped by the environment, it is not literally extended into it. Situated Cognition and DST tend to be agnostic about boundaries, focusing instead on the dynamics of coupled systems.
Methodology. DST offers a formal, continuous, non-computational modeling framework. The other three rely on conceptual analysis, phenomenological description, and qualitative or experimental methods. The methodological pluralism of the field is both a strength and a source of friction: DST can seem too abstract to capture the lived experience that Enactivism prizes, while Enactivism can seem too vague to generate testable predictions.
All four frameworks remain active today, but they occupy different niches. Enactivism has been especially influential in the philosophy of consciousness and emotion, where its emphasis on lived embodiment and sense-making resonates with phenomenological approaches. The Extended Mind Thesis continues to generate debate in philosophy of mind and cognitive science, particularly around the ethics of cognitive enhancement and the nature of belief. Dynamical Systems Theory is a standard tool in developmental psychology, motor control, and robotics, where its mathematical precision is valued. Situated Cognition, while no longer a distinct research program, has been absorbed into distributed cognition, ecological psychology, and human-computer interaction.
What the leading frameworks agree on is that cognition cannot be understood by studying the brain in isolation. The body, the environment, and the history of interactions are not optional extras but constitutive elements of cognitive processes. Where they disagree is on how far that insight should be pushed: whether it requires abandoning representation, redrawing the boundaries of the mind, or adopting a new mathematical language. These disagreements are not signs of weakness but of a field still working out the implications of its central insight. Embodied cognition has permanently reshaped the philosophy of mind by forcing it to take seriously the fact that thinkers are always, irreducibly, bodies in a world.