{
  "id": "enactivism",
  "title": "Enactivism",
  "lang": "en",
  "source_lang": "en",
  "summary": "Enactivism explains cognition as meaningful activity arising through an agent’s embodied engagement with its environment.",
  "category": "philosophy",
  "subcategory": "metaphysics-epistemology",
  "type": "concept",
  "aliases": [
    "Enaction",
    "Enactive approach to cognition"
  ],
  "generated_at": "2026-11-06",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/enactivism/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/enactivism/",
      "json": "https://aiwiki.page/zh/enactivism.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Enactivism is a family of approaches in cognitive science and philosophy of mind that explains cognition through the active engagement of embodied agents with their environments. Its central claim is that perceiving, understanding, and acting are not adequately explained as the internal reconstruction of an independently specified world. Cognitive agents instead establish meaningful relations with their surroundings through their activities, bodily organization, and histories of interaction. “Enacting a world” does not ordinarily mean creating physical reality at will; it means bringing forth a domain of significance in which things matter to a particular kind of agent.\n\nThe approach acquired its influential modern formulation in The Embodied Mind (1991), by Francisco Varela, Evan Thompson, and Eleanor Rosch. It subsequently developed into partially overlapping biological, sensorimotor, social, and radical approaches. These share an emphasis on embodied cognition, but differ over whether cognition necessarily depends on living organization, whether mental representations should be rejected, and how language and abstract thought arise. Enactivism consequently functions less as a single established theory than as a research program linking biology, perception, social interaction, experience, and the philosophical foundations of cognitive explanation."
    },
    {
      "heading": "Intellectual origins and historical development",
      "text": "Enactivism emerged against the background of twentieth-century debates about how minds relate to bodies and environments. Classical computational approaches treated cognition as operations over internal structures whose contents represent objects or states of affairs. Connectionist approaches replaced many explicit symbolic rules with distributed activity in networks, but could still describe cognition primarily as an internal mapping from inputs to outputs. Enactivists questioned the shared assumption that cognition should begin with a world already divided into relevant features and an organism that must reconstruct those features internally.\n\nA major biological source was the work of Chilean researchers Humberto Maturana and Francisco Varela. Their account of living organization, developed during the early 1970s and presented in English in Autopoiesis and Cognition (1980), described organisms as self-producing systems. Their later book The Tree of Knowledge, published in Spanish in 1984 and English in 1987, connected biological organization with cognition and the emergence of a lived world. This work grew partly out of cybernetics, although it redirected attention from externally assigned control functions toward the organization through which a system constitutes and maintains itself.\n\nA second source was phenomenology, especially the accounts of embodiment associated with Edmund Husserl and Maurice Merleau-Ponty. In Phenomenology of Perception (1945), Merleau-Ponty described the body not merely as an object in space but as the practical means through which a world becomes accessible. Skilled movement, bodily orientation, and perceptual meaning were therefore not secondary additions to intellectual judgment. This offered enactivists a vocabulary for explaining the experienced body without reducing experience to detached observation of bodily mechanisms.\n\nThe Embodied Mind brought these biological and phenomenological strands into dialogue with cognitive science and Buddhist analyses of experience. Its authors used “enactive” to emphasize that cognition depends on the kinds of experience made possible by a body with particular sensorimotor capacities, situated in biological, psychological, and cultural contexts. Later work diversified the program. Thompson’s Mind in Life (2007) developed its biological and phenomenological foundations; Alva Noë’s Action in Perception (2004) advanced an action-centered account of perception; and Daniel Hutto and Erik Myin’s Radicalizing Enactivism (2013) argued against content-involving explanations of basic cognition. These developments established related but nonidentical positions rather than a single doctrinal succession."
    },
    {
      "heading": "Living organization, autonomy, and autopoiesis",
      "text": "Biologically oriented enactivism begins with autopoiesis, literally self-production. An autopoietic system consists of processes that produce components which, through their interactions, regenerate the network of processes producing them and constitute the system as a distinguishable unity. A living cell provides the central example: chemical reactions maintain components of the system, including a boundary that helps sustain the conditions under which those reactions continue. The boundary is not simply a container imposed from outside; its production belongs to the organization of the system itself.\n\nAutopoiesis must be distinguished from energetic or material isolation. Organisms depend on flows of matter and energy, and their persistence requires ongoing exchanges with their surroundings. The relevant closure is organizational: processes depend recursively on one another in a network that maintains its own organization. Accordingly, an organism can be operationally closed while remaining thermodynamically open. “Closure” in this setting does not mean that external events have no effects, or that an organism lacks contact with a real environment.\n\nThe associated concept of autonomy concerns the system’s capacity to establish and maintain its own identity through interdependent processes. An external observer can describe a machine as fulfilling a purpose, but an organism’s continued existence itself supplies a basis for distinguishing conditions that sustain or threaten it. Enactivists use this distinction to explain why biological activity can have an intrinsic orientation rather than only a function assigned by a designer. The organism’s identity is precarious: the processes sustaining it must keep operating, and their interruption can destroy the organization they maintain.\n\nLater work distinguished self-maintenance from adaptivity. In influential formulations developed by Ezequiel Di Paolo, an adaptive system regulates its activity with respect to conditions affecting its continued viability. Merely remaining within a viable range is different from detecting and compensating for movement toward its limits. A bacterium’s regulation of movement in relation to chemical gradients illustrates the kind of organization at issue, although the details of bacterial behavior require specific biochemical explanation rather than a generic appeal to autonomy.\n\nThese ideas support the proposed continuity between life and mind: cognitive capacities elaborate forms of self-maintenance already present in living organization. The strongest formulations approach the claim that living is cognitive; more qualified versions distinguish minimal biological sense-making from perception, learning, and reflective understanding. The distinction matters because autopoiesis alone does not specify every feature normally associated with cognition. Whether additional capacities such as adaptivity, agency, or learning are necessary remains a central issue within biological enactivism."
    },
    {
      "heading": "Sense-making, significance, and the enacted world",
      "text": "Sense-making names the process through which environmental conditions acquire significance relative to an agent’s organization and activity. A sugar gradient is physically describable independently of a bacterium, but its significance as a possible nutrient depends on the bacterium’s metabolic requirements and capacities for interaction. The same substance can be beneficial, irrelevant, or harmful to different organisms, or to the same organism under different conditions. Meaning in this minimal biological sense is therefore relational rather than an intrinsic property of the stimulus alone.\n\nThis account links cognition to normativity: some interactions count as going better or worse for the agent. Such norms need not be consciously formulated rules. They can arise from the conditions under which an autonomous system maintains itself. An organism’s activities can succeed or fail relative to those conditions even when no observer has assigned it a task. Enactivists regard this as an important step toward explaining intentionality, the directedness of cognition toward something, without beginning with internal symbols that already possess meaning.\n\nThe environment, on this account, is neither an undifferentiated physical surround nor a private mental image. An organism’s sensory capacities, bodily dimensions, needs, and acquired skills determine which differences it can encounter and exploit. A flower presents different possibilities to a bee, a grazing mammal, and a human botanist. These possibilities are constrained by actual structures and causal relations, but their relevance depends on the agent. Enactivist talk of “bringing forth a world” concerns this organization of significance, not the claim that flowers or chemical compounds cease to exist when unperceived.\n\nSense-making also develops across time. Recurrent interactions can establish habits that shape what becomes noticeable and how an agent responds. Human hunger, for example, involves metabolic regulation, but eating is additionally organized by learned practices, familiar objects, social schedules, and linguistic categories. Human significance therefore cannot be read directly from biological survival. Enactivists have increasingly emphasized habits, identities, and social practices as relatively autonomous sources of concern, while debating how their norms remain connected to organismic life.\n\nA persistent difficulty is the transition from biological evaluation to semantic correctness. Conditions can be favorable or unfavorable for an organism without thereby making its activity true or false. Explaining how practical significance develops into judgments with determinate content requires further accounts of learning, communication, and social normativity. Enactivism offers a proposed foundation for such an explanation, but its different branches disagree about whether minimal sense-making already possesses semantic content or whether content appears only with more complex cultural practices."
    },
    {
      "heading": "Perception as skilled sensorimotor activity",
      "text": "Sensorimotor enactivism explains perception through the organized relations between sensory change and possible action. Moving the eyes, turning the head, touching a surface, or walking around an object changes sensory stimulation in structured ways. The relevant regularities are called sensorimotor contingencies. Seeing a cube, for example, involves practical sensitivity to how its visible surfaces would change with movement, rather than simply receiving a complete internal picture of all its sides.\n\nJ. Kevin O’Regan and Alva Noë’s 2001 account of vision emphasized mastery of such regularities. Different senses involve different patterns of dependence: visual stimulation changes systematically with eye and head movements, whereas tactile stimulation changes through contact and bodily exploration. Perceptual differences can therefore be explained partly through differences in the structure of sensorimotor engagement. “Mastery” here need not mean explicit theoretical knowledge; it can be a practical capacity expressed in competent activity.\n\nThis approach distinguishes the sensory stimulation currently available from the wider environment perceptually present to an agent. A tomato can appear as a whole object although only part of its surface is visible. A room can seem richly available even though attention samples only a small portion at any moment. Enactivists explain this presence partly through the agent’s ability to obtain further detail by looking or moving. Change-blindness experiments, in which observers fail to notice substantial scene changes under particular conditions, challenge the idea that ordinary vision maintains an exhaustive, continuously accessible internal image.\n\nResearch on sensory substitution provides another important example. In devices that translate camera input into tactile or auditory patterns, users may acquire the ability to discriminate spatial layouts or locate objects through active exploration. Early work associated with Paul Bach-y-Rita helped establish the importance of training and action in such capacities. These findings are relevant to sensorimotor accounts because successful use depends on learning systematic relations among movement, device output, and environmental structure. They do not by themselves establish that every trained user has ordinary visual experience.\n\nThe strongest claims remain contested. Neural mechanisms clearly contribute to perception, and perceptual experience can occur without ongoing overt movement, as in dreams or imagery. Sensorimotor theorists therefore distinguish current movement from a history of learning and a standing capacity for sensorimotor engagement. Critics question whether this distinction explains experience or merely redescribes capacities associated with it. A further issue is whether action is necessary for perception in general, constitutive only of particular perceptual abilities, or simply a powerful contributor to their development and successful operation."
    },
    {
      "heading": "Representation and radical enactivism",
      "text": "Many enactivists oppose explanations that treat cognition as the manipulation of mental representations, but the extent of this opposition varies. A representation ordinarily stands for something, can be accurate or inaccurate, and plays a role in explaining behavior through its content. An internal neural state does not automatically count as a representation merely because it correlates with an environmental variable. The substantive question is whether its representational content is necessary to explain what the system does.\n\nBiological and sensorimotor enactivists often argue that basic adaptive behavior can be explained through coupled activity without positing an internal model of the world. A person adjusting posture on a moving surface, for example, continuously responds through neural, muscular, vestibular, and environmental interactions. Such coordination can sometimes be modeled without attributing a detailed representation of the surface’s motion. This does not imply that the nervous system is inactive or that internal organization is unimportant; it concerns the explanatory interpretation of those mechanisms.\n\nRadical enactivism, especially associated with Hutto and Myin, makes a stronger distinction between basic minds and content-involving cognition. Their “hard problem of content” asks how semantic content could be derived from merely informational relations understood as covariance or reliable dependence. Smoke can covary with fire without making a claim about fire, and a neural response can reliably track a stimulus without thereby possessing truth conditions. Radical enactivists argue that explaining basic cognition through content may introduce a property that has not itself been adequately naturalized.\n\nThis position does not deny every form of representation. Language, pictures, diagrams, and socially regulated symbolic practices provide clear cases in which something is used to stand for something else. Radical enactivists propose that some human cognitive activities acquire content through participation in these practices. The challenge is to explain the developmental and historical transition from embodied, non-contentful activity to capacities involving assertions, descriptions, and correctness governed by shared norms.\n\nCritics argue that representational explanations remain indispensable for navigation, memory, planning, and reasoning about absent situations. Evidence concerning place cells and other spatially selective neural activity, for example, supports sophisticated accounts of internal spatial organization, though it does not settle which theory of representation is correct. Enactivists must show more than that behavior can occasionally be redescribed without representational terminology. Conversely, defenders of representation must specify what explanatory work content performs beyond identifying a useful correlation. The dispute turns on competing accounts of mechanism, content, and explanatory adequacy rather than a simple choice between accepting or rejecting internal brain processes."
    },
    {
      "heading": "Social interaction and participatory sense-making",
      "text": "Social enactivism challenges the assumption that understanding another person always begins with inferring hidden mental states from externally observed behavior. In ordinary interaction, people respond to posture, gesture, gaze, speech rhythm, and emotion within an unfolding activity. A conversation, a shared task, or a disagreement can organize participants’ actions in ways that are not adequately described by considering each participant separately. Social understanding may therefore occur partly through interaction itself.\n\nHanne De Jaegher and Di Paolo introduced participatory sense-making in 2007 to describe how individual sense-making is transformed through coordinated engagement. Interaction can acquire a relatively autonomous organization: its patterns sustain themselves, constrain participants, and create possibilities neither participant fully controls. Two pedestrians repeatedly moving to the same side while trying to pass illustrate how coordination can momentarily perpetuate an unwanted pattern. In conversation, turn-taking and mutual adjustment can support a shared direction that neither speaker specified in advance.\n\nThis approach does not eliminate individual autonomy. Participatory sense-making concerns relations among agents who retain their own concerns and capacities. Interaction may support those concerns, disrupt them, or generate new ones. Its significance lies in the reciprocal relation between the organization of the encounter and the activities of the participants. A social process is not simply a larger organism, and coordination by itself does not establish a shared mind or common intention.\n\nExperimental work using perceptual-crossing arrangements has explored minimal forms of interaction. In a well-known 2009 study by Malika Auvray, Charles Lenay, and John Stewart, pairs of participants moved cursors in a simplified virtual environment and received tactile stimulation when encountering relevant objects. Participants could interact with another user’s avatar as well as nonresponsive or differently responsive objects. The arrangement examined how reciprocal responsiveness could support encounters with another agent under severely restricted perceptual conditions. Such results are informative about interaction dynamics, but do not independently explain the full range of human social understanding.\n\nSocial enactivism also extends to language. Linguistic Bodies (2018), by Di Paolo, Elena Cuffari, and De Jaegher, develops an account connecting embodied agency, participatory activity, and linguistic practices. Speaking is treated not only as transmitting preformed content but as acting within norms of coordination, interpretation, and repair. Difficult questions concern disagreement, deception, institutional authority, and asymmetrical power: interaction need not be harmonious, and its organization can restrict rather than enlarge a participant’s possibilities. These cases require attention to historical and social structures beyond immediate face-to-face coordination."
    },
    {
      "heading": "Experience, phenomenology, and neurophenomenology",
      "text": "Enactivism treats lived experience as a central subject of cognitive science rather than as an optional addition to behavioral and neural explanation. The body is both an observable biological system and the lived perspective through which perception and action occur. An account of grasping, for example, can describe joint angles, muscle activity, and neural control while also attending to the experienced readiness of an object for handling. These descriptions concern different aspects of an embodied activity and need not compete as alternative substances or worlds.\n\nPhenomenological analysis contributes distinctions that are easily obscured by treating experience as a collection of internal images. The sense of bodily orientation, the difference between acting and being moved, and the temporal organization of an unfolding experience all affect what cognition is like. Enactivists also distinguish the body as an object of attention from the prereflective bodily capacities through which attention is directed elsewhere. Such distinctions help specify the phenomena that experimental explanations are intended to address.\n\nVarela proposed neurophenomenology in 1996 as a research program linking disciplined descriptions of experience with investigations of neural dynamics. Its guiding idea was mutual constraint: first-person distinctions could help organize experimental analysis, while physiological findings could refine descriptions of experience. This did not mean accepting every introspective report as infallible. Reports must be elicited, compared, and interpreted through explicit procedures, and their relation to the target experience must remain open to examination.\n\nA frequently discussed study by Antoine Lutz and colleagues, published in 2002, combined reports of participants’ experiential states before a visual task with measurements of brain activity. The analysis examined whether differences in reported preparation were associated with differences in neural dynamics and task performance. Its importance for neurophenomenology lay in using experiential distinctions to inform the organization of physiological data. It was not a demonstration that subjective reports provide direct access to neural mechanisms, nor a general solution to the problem of consciousness.\n\nThe historical engagement with meditation and Buddhist thought also served methodological and philosophical purposes. The Embodied Mind drew on analyses of the instability of a supposedly fixed self and on practices intended to refine attention to experience. These connections do not make enactivism a religious doctrine, and Buddhist traditions themselves contain diverse accounts of mind. The scientific difficulties remain substantial: training can alter what is reported, expectations can shape descriptions, and relationships between reports and neural measurements may depend on analytic decisions. Enactivist research therefore faces ordinary requirements of experimental control and reproducibility alongside its distinctive concern with experiential accuracy."
    },
    {
      "heading": "Dynamical models and empirical investigation",
      "text": "Enactivist explanations often use dynamical systems theory to describe how cognition unfolds through continuous reciprocal interaction. Instead of treating the environment as a source of inputs followed by internal processing and outputs, such models describe mutually dependent changes in brain, body, and surroundings. A schematic formulation is:\n\n\\[\\dot{x}=f(x,b,e), \\qquad \\dot{b}=g(x,b,e), \\qquad \\dot{e}=h(x,b,e),\\]\n\nwhere \\(x\\) denotes neural or internal states, \\(b\\) bodily states, and \\(e\\) relevant environmental states. The equations express reciprocal dependence, not a complete theory of cognition. Specific scientific claims require identifying variables, measuring interactions, and justifying the functions used.\n\nDynamical models can characterize attractors, transitions, coordination, and the stability of patterns. In skilled movement, a stable pattern may arise from interactions among neural activity, bodily mechanics, and task constraints rather than from a detailed instruction specifying every movement. Work on coordination dynamics, including the Haken–Kelso–Bunz model introduced in 1985, showed how changes in a control parameter could produce transitions between patterns of rhythmic movement. Such research is not exclusively enactivist, but it supplies tools for explaining organization at the level of the coupled system.\n\nRobotics offers a complementary method. A robot’s behavior may depend heavily on its body shape, material properties, sensor placement, and environmental contact. Designers can sometimes obtain adaptive performance by exploiting these relations rather than constructing a comprehensive internal world model. This is relevant to enactivist claims about embodiment, although successful robot behavior does not establish biological autonomy. A machine whose goals and maintenance conditions are externally supplied may model sensorimotor coordination without modeling an organism’s self-constituted identity.\n\nEmpirical investigation can also manipulate the normal relation between perception and action. Altered visual feedback, virtual environments, and sensory-substitution devices make it possible to study adaptation when familiar sensorimotor regularities change. Interaction experiments can compare reciprocal engagement with recorded or otherwise nonresponsive stimuli. These designs test more specific hypotheses than the broad claim that cognition is embodied, and they can help identify whether reciprocal coupling contributes something beyond exposure to similar sensory information.\n\nThe central methodological issue is underdetermination. A successful dynamical model does not automatically show that the modeled process lacks representations, while a successful computational model does not establish that cognition is confined to computation inside the brain. Models must be compared through predictions, intervention, generalization, and explanatory scope. Biological, sensorimotor, and social enactivism also target different phenomena, so no single experiment confirms or refutes the entire family. Progress depends on connecting broad commitments to mechanisms and discriminating tests rather than treating compatibility with an observed result as decisive support."
    },
    {
      "heading": "Relations to ecological, extended, and predictive approaches",
      "text": "Enactivism belongs to the wider field of embodied cognition, but the terms are not interchangeable. An embodied theory may hold only that bodily states influence otherwise conventional internal processing. Enactivism generally makes the stronger claim that active embodiment helps constitute cognition or its domain of significance. Similarly, being environmentally embedded can mean merely that context affects performance, whereas an enactive account asks how agent and environment jointly organize the activity to be explained.\n\nA close neighbor is ecological psychology, associated especially with James J. Gibson. Gibson’s The Ecological Approach to Visual Perception (1979) described perception as access to information available in an organism’s surroundings and emphasized affordances, the possibilities for action offered by an environment relative to an animal’s capacities. Both traditions reject the assumption that perception must reconstruct a world from impoverished sensations. They differ in emphasis: ecological accounts often foreground the structure of available information, while enactivists emphasize autonomous organization and the enactment of significance. Whether these perspectives are complementary or rest on different conceptions of information remains debated.\n\nThe extended mind thesis, influentially defended by Andy Clark and David Chalmers in 1998, holds that some cognitive processes can include resources beyond the biological individual. Their well-known notebook example asks whether a reliably available external memory aid can play a role comparable to biological memory. Enactivists likewise challenge an automatic boundary at the skull, but their reasoning need not depend on functional equivalence between inner and outer resources. Some instead emphasize how tools transform embodied activity, while biologically oriented approaches ask whether the autonomy of the organism places principled limits on cognitive extension.\n\nRelations with predictive processing are especially complex. Predictive theories explain perception and action through mechanisms involving predictions and the reduction of prediction error. Active inference additionally treats action as contributing to the fulfillment of predictions or the maintenance of expected conditions. These approaches share enactivism’s interest in active engagement and self-maintenance. However, many predictive accounts employ hierarchical generative models understood as representations of hidden causes, placing them in tension with anti-representational enactivism. Attempts at reconciliation often depend on whether models are interpreted as internal descriptions or as formal characterizations of embodied dynamics.\n\nEnactivism also has affinities with pragmatism, particularly its emphasis on action, habit, and the practical conditions of meaning. John Dewey’s 1896 criticism of the reflex-arc concept anticipated objections to separating stimulus, processing, and response into independent units. These historical affinities do not erase differences among the traditions. Enactivism’s distinctive contribution lies in bringing questions about autonomous life, sensorimotor organization, and lived experience into a connected research program, rather than merely asserting that action is important to thought."
    },
    {
      "heading": "Applications, explanatory limits, and unresolved questions",
      "text": "In education, enactivist approaches have informed work on learning through bodily engagement, shared activity, and the transformation of practical understanding. Mathematical learning, for example, can involve gesture, manipulation, and coordinated use of diagrams rather than only the acquisition of explicitly stated rules. The relevant claim is not that physical movement automatically improves learning, but that particular forms of activity can help organize concepts and skills. Educational outcomes still depend on the task, instructional design, prior knowledge, and the forms of assessment used.\n\nIn artificial intelligence and robotics, enactivism raises questions about what would distinguish genuine agency from externally interpreted task performance. A system can classify images or generate text without maintaining the biological organization emphasized by autopoietic accounts. Conversely, a robot that adapts through sensorimotor interaction may possess some capacities of interest without satisfying every proposed criterion for autonomous cognition. Different branches therefore reach different conclusions about artificial minds: biological approaches may stress constitutive self-maintenance, while sensorimotor and radical approaches can focus on embodied capacities and interaction histories.\n\nThe explanation of abstract and “offline” cognition remains a major challenge. People can reason about nonexistent objects, formulate mathematical proofs, imagine remote futures, and think while temporarily disengaged from immediate surroundings. Enactivist responses appeal to the development of bodily skills, reenactment, linguistic practices, and the use of external symbols. These resources are substantial, but they must explain the structure and systematicity of particular abilities, not merely establish that those abilities have embodied developmental origins. Showing that symbolic reasoning is learned through social activity does not by itself explain how a person later carries out a complex inference.\n\nThe relation between life and subjectivity is another unresolved issue. Self-maintenance may explain why events matter to a system, but it does not straightforwardly establish that the system feels anything. Some enactivists propose a deep continuity between organismic regulation, affectivity, and experience; others remain cautious about attributing experience to minimal living systems. The hard problem of consciousness therefore persists in a distinctive form: why should autonomous embodied organization be accompanied by experience, and which organizational differences matter?\n\nA final dispute concerns the explanatory status of interaction. An environmental factor can causally influence cognition without being part of the cognitive process itself. Establishing constitution requires more than demonstrating strong coupling or behavioral dependence. Enactivist accounts must specify the relevant system, its boundaries, and the temporal scale of explanation, especially when processes span metabolism, individual skill, interpersonal coordination, and cultural practice. These questions determine whether an account identifies a distinctive mechanism of cognition or supplies a broader description compatible with several competing theories."
    }
  ],
  "links": [
    "cognitive-science",
    "philosophy-of-mind",
    "embodied-cognition",
    "cybernetics",
    "phenomenology",
    "buddhism",
    "cell",
    "energy",
    "intentionality",
    "language",
    "consciousness",
    "meditation",
    "reproducibility",
    "active-inference",
    "pragmatism",
    "education",
    "artificial-intelligence",
    "hard-problem-of-consciousness"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}