What Is Enactivism?

Abstract

‘Enactivism’ is usually taken to be an umbrella term encompassing any approach that emphasises the importance of recurrent loops of sensorimotor activity to adequate philosophical and scientific understandings of how minds fit into the natural world. Under this umbrella different variants can then be distinguished. I argue that this is a distortive and unhelpful understanding insofar as it ignores three central commitments of the enactive approach as originally set out in Varela, Thompson & Rosch’s The Embodied Mind. These commitments to life/mind continuity, experiential primacy, and the reciprocity between science and self-understanding set the enactive approach apart from other variants of embodied cognition in science and philosophy, and do so in ways that make it distinctively timely and valuable.

1. Introduction

What on earth is ‘enactivism’? Surely we should know the answer to this question by now, 35 years after Varela, Thompson and Rosch introduced the ‘enactive approach’ in The Embodied Mind, with a deep well of special issues, edited volumes, articles and books on the topic to draw on. And perhaps we do know: ‘enactivism’ is an umbrella term encompassing several overlapping frameworks that take some degree of inspiration from The Embodied Mind (Varela et al., 2017 henceforth TEM), united by their shared emphasis on the centrality of embodied, sensorimotor activity to adequate philosophical and scientific understanding of how minds fit into the natural world. There are three main variants: ‘autopoietic’ enactivists who emphasise the importance of organisational properties of the biological processes involved in the self-maintenance of living things (e.g. Thompson, 2007; Di Paolo et al., 2017); ‘sensorimotor’ enactivists, who care less than ‘autopoietic’ enactivists about biological organisation, but more than them about identifying the specific patterned relations between sensation and movement that characterise different modalities of conscious perception (e.g. Hurley, 1998; Noë, 2004; O’Regan, 2012); and ‘radical’ enactivists who don’t care so much about any of that stuff, but do care about criticising research in the mind sciences that appeals to representational states, and about arguing for a distinction between a ‘basic’, contentless kind of mindedness characteristic of nonverbal cognition and a richer mode of contentful mindedness that comes with enculturation into language (e.g. Hutto & Myin, 2012, 2017).

The lump-and-split taxonomical strategy just summarised – lumping approaches that emphasise embodied sensorimotor interaction under the heading ‘enactivism’, then making a rough tripartite division between them – has become common currency (Ward, Silverman, & Villalobos, 2017; Hutto, 2023; Shapiro & Spaulding, 2025). But, I argue here, it fails to capture what is distinctive, original and important in the enactive approach as sketched in TEM and developed by Evan Thompson and others. A valid assessment of the prospects and potentially distinctive value of enactive cognitive science requires a clear picture of just what ‘enactivism’ is, and what demarcates it from neighbouring approaches.

My aim here is to provide such a picture and point towards some of its payoffs. A distinctively enactive approach to study of mind should, I argue, be understood in terms of the structural role that three interwoven commitments play in its theorising. First, mind-life continuity: the agential, experiential and meaningful properties distinctive of our minds depend on organisational properties distinctive of living things. Second, the primacy of lived experience: the way in which perception, agency and dealings with meaning feel to living creatures like us is centrally important to enactivism, both as an explanandum and as a source of evidential constraint. Third, the reciprocity of the relationship between our practices for theorising about the mind, and our self-understanding: the experiences of ourselves that we are trying to theorise and our culturally instituted practices and vocabularies for theorising about them are not insulated from each other. The practices and vocabularies draw on the experiences as their source material and subject matter, but also pre- and re-shape the experiences themselves.

These emphases on mind-life continuity, experiential primacy, and science/self-understanding reciprocity are sketched out in TEM, developed in later work by Varela (e.g. Varela, 1996; Weber & Varela, 2002), Thompson (2007, 2014; Frank et al., 2024) and others (e.g. Di Paolo et al., 2017, 2018), but usually absent in other work classified as ‘enactive’. The lump-and-split taxonomy of enactivism fails to capture what is original and distinctive in TEM’s project because it lumps TEM’s enactivism together with approaches that ignore or eschew commitments to continuity, primacy and reciprocity. By distracting attention from these distinctive commitments, the lumping obscures the nature and potential of enactivism as a distinctive and timely theoretical approach. In an effort to be the change I’d like to see in the world, in what follows uses of ‘enactivism’ or ‘enactivist’ unqualified by modifiers or scare-quotes henceforth refer to views that combine these emphases on continuity, primacy and reciprocity. ¹

The next two sections sketch two ways of carving up the space of approaches to theorising about the mind. Section two contrasts ‘cognitivist’ approaches which emphasise skull-bound computations with ‘embodied’ approaches that emphasise structures that span brain, body and world. The problem with this carving is that it obscures an important dimension of variation among ‘embodied’ approaches: whether they hold that psychological states or capacities should be understood in terms of information-processing or pickup, or in terms of their relationship to properties distinctive of living things. The third section explains this contrast between ‘informational’ and ‘organicist’ approaches. Section four shows that it is orthogonal to the ‘cognitivism’/‘embodied cognition’ contrast, because many ‘embodied’ approaches are agnostic about or resistant to organicism. This puts us in a position to see what makes enactivism a distinctive position in theoretical space. Section five explains the commitment to life/mind continuity that makes TEM’s enactivism organicist, and section six explains the commitments to experiential primacy and science/self-understanding reciprocity that further set it apart from other embodied approaches. The final section points towards some payoffs of this understanding of enactivism. Most important is that it lets us see that enactivism offers a distinctive vision of the relationship between scientific practice and self-understanding, one which has only become more timely in the years since TEM was first published.

2. Cognitivism or Embodied Cognition?

The lump-and-split taxonomy which has three semi-distinct ‘enactivisms’ (autopoietic, sensorimotor, radical) emerging from TEM distorts our understanding of enactivism because it obscures the emphases on mind/life continuity, experiential primacy, and science/self-understanding reciprocity that make the project sketched out in TEM distinctive. Before arguing for this claim, it is helpful to zoom out and consider the motivations for lumping these views together in the first place. Despite their differences, it has seemed to most proponents and critics alike that they are different variants of an overarching view, ‘enactivism’. Why?

The lumping makes sense in the context of a reaction to a particular kind of computational cognitive science that took shape and became influential in the 20^th century. This is the ‘cognitivist’ project of modelling psychological processes in terms of the kind of production and rule-governed manipulations of symbolic states characteristic of early digital computation, coupled with the natural assumption that, when it comes to human psychological capacities, the processes most central to this production and manipulation are located in the human brain and nervous system (see Haugeland, 1978, for an early articulation and critique, Chipman, 2017; Friedenberg et al., 2022 for contemporary introductions to cognitive science in these terms). By the late 20^th century the vision of mind encouraged by this approach was codified in pop-science bestsellers informing readers that they were ‘survival machines – robot vehicles blindly programmed to preserve the selfish molecules known as genes’ (Dawkins, 1982/2016, p. xix), and that ‘[t]he mind is a system of organs of computation designed by natural selection… The mind is what the brain does; specifically, the brain processes information, and thinking is a kind of computation’ (Pinker, 1997, p. 21).

As the cognitivist vision of minds as skull-bound symbol-crunchers sunk into popular consciousness, it attracted a growing backlash. Philosophers like Hubert Dreyfus (1972; also Haugeland, 1995; Grene, 1995) drew on readings of phenomenological philosophy to argue that the most fundamental kind of mindedness is embodied in a skilful, unreflective attunement to a meaningful world, rather than the rule-respecting inferential processes that are cognitivism’s stock-in-trade. And several existing or emerging strands of cognitive scientific research appeared to fit nicely with this emphasis on embodied activity. Ecological psychologists like James Gibson argued that the environments of mobile organisms were replete with rich, behaviour-specifying sources of information that can drive adaptive behaviour without requiring intervening computational or inferential processes (Gibson, 2015; Lee & Reddish, 1981). Some AI researchers, harking back to early cybernetic models of intelligence as sensorimotor control, took to building whole, embodied systems that exhibited semi-flexible adaptive behaviour despite lacking the kind of centralised representation and control structures emphasised by cognitivism (Beer, 1997; Brooks, 1991; Webb, 1995). Proponents of ‘distributed cognition’ explored how the information-processing loads could be eased by, or distributed over, relationships obtaining between brains (or central-processors), bodies and environments (Ballard et al., 1997; Clark, 1997; Hutchins, 1995). Finally, work on ‘dynamical systems’ approaches to cognition showed how sensorimotor capacities and developmental trajectories could be effectively modelled using the theoretical and mathematical vocabulary of complex systems theory (Kelso, 1994; Prigogine & Stengers, 1984; Van Gelder and Port, 1995). To some researchers, this suggested a rigorous and mathematically precise way of studying mind as embodied interaction, a viable alternative to the computational mathematical toolkit of cognitivism.

This is an important aspect of the context in which TEM was written and received: a broadly cognitivist orthodoxy and a backlash from various theoretical and empirical sources, loosely united by an emphasis on the importance of body/world systems and interactions to the proper study of minds. ² Despite differences in subject matter, methodology and theoretical vocabulary between these sub-movements, viewed at a suitable level of abstraction they could be – and were – usefully lumped together as a unified happening in cognitive science and its philosophy. For cognitivism, the study of mind was primarily the study of computational processes sandwiched between sensory input and motor output (Hurley, 1998). Bodies, environments and their relationships with this computational core were of secondary importance at best, as causal channels into and out of where the real cognition happens. ‘Embodied cognition’ came to be used as a broad umbrella term for any theoretical or empirical approach that rejected this ‘classical sandwich’ approach to the mind (Shapiro, 2004; Wilson, 2002). For ‘embodied’ approaches, it was the intra-cranial happenings privileged by cognitivism that were of secondary importance, deriving their cognitive significance from their place within a wider web of brain/body/world interactions.

This global lumping of anti-cognitivist approaches under the heading ‘embodied cognition’ is the context within which the more local lumping together of ‘sensorimotor’ and ‘radical’ approaches with TEM’s enactive approach occurred. Once the many movements comprising ‘embodied cognition’ were bundled together, it was natural to wonder about their interrelations. Why did so many of them involve words beginning with the letter ‘e’ (ecological, embedded, embodied, emotional/affective, enactive, extended, emergent, etc.)? Were the ‘e’s mutually compatible? Did one or more of them have primacy over the others? The lump-and-split taxonomy arose as part of wider attempts to clarify such matters. ‘Enactive’ approaches, the idea went, were a sub-class of approaches within embodied cognition, distinguished by their central commitment to the claim that ‘cognition emerges from sensorimotor activity’ (Shapiro & Spaulding, 2025). Other flavours of embodied cognition either appeared incompatible with this commitment (e.g. Lakoff & Johnson, 1999; Clark & Chalmers, 1998), or did not accord it central importance (e.g. Gibson, 2015; Port & van Gelder, 1995). ³ The presence or absence of a central commitment to conceptualising and studying cognition as emergent from cycles of sensorimotor interaction thus seemed a useful way of partitioning the cluttered space of ‘embodied cognition’ theorising.

So the lump-and-split taxonomy invites us to partition the space of theorising about minds like this: start by choosing between ‘cognitivist’ and ‘embodied cognition’ stances on the structures on which cognition depends. Are they skull-bound computations (cognitivism); or more encompassing aspects of brain/body/world systems (‘embodied cognition’)? ‘Enactivism’ is then specified as one way of articulating which kinds of brain/body/world-spanning structures matter: cycles of sensorimotor interaction. Varieties of ‘enactivism’ can then be distinguished in terms of the different ways they attempt to spell out that claim – the structural properties of the cycles of interaction they emphasise, the specific claims they make about how and which cognitive structures emerge from these cycles, and the other conceptual, empirical or formal resources they draw on to develop and implement their views.

3. Informationalism or Organicism?

Here is a different way of partitioning the space of theorising about minds. Start by asking: what kinds of properties do the meaning or significance of mental states of a subject or system – its perceptions, beliefs, urges, and willings, for example – fundamentally depend on? The first candidate answer is: on their informational properties. The second is: on their organic properties; that is, on properties that are distinctive of living things. Informationalists are those who give the first answer; organicists are those who give the second. This organicist/informationalist axis is an orthogonal dimension of variation to the cognitivism/embodied cognition axis (see Figure 1 in §4), and we will see that appreciating this helps illuminate the distinctive features of TEM’s enactivism.

‘Informationalism’ and ‘organicism’ are fuzzy terms of art in need of explanation. The idea of ‘informational properties’ appealed to here is intended to be broad and inclusive, encompassing any view that treats cognition as the processing of, or dealing with, information. ⁴ Cognitivism is a paradigmatic case of such a view, in which the content or significance of psychological states depends on underlying symbolic states and the computational processes they figure in. But many other kinds of approaches work with an informationalist conception of the properties in virtue of which psychological states have their meaning: connectionist approaches (Buckner, 2018; Churchland, 1995) appealing to information encoded in distributed states of a network; neo-cybernetic or predictive processing approaches (Clark, 2015; Seth, 2014) appealing to the way in which control systems approximate processes of Bayesian inference, or to the information encoded in a hierarchically stacked sequence of constantly updating generative models; ecological approaches appealing to behaviour-specifying information present in a perceiver’s sensory array (Lee & Reddish, 1981), or sociomaterial environment (Bruineberg, Chemero & Rietveld, 2019); dynamical systems approaches appealing to meaningful control parameters and state-space variables that govern how a system’s behaviour unfolds (Chemero, 2000; Dreyfus, 2007; Kelso, 1994); or ‘wide computationalist’ approaches that take the computations supporting cognition to be distributed across brain/body/world systems (Kersten, 2024; Wilson, 1994), to name a few.

Some of these approaches cut across each other in important ways (‘wide computationalism’, for example, is a claim about where computations underlying cognition are located; cognitivism and connectionism make competing claims about what kind of computations underlie cognition), and disagree on others – notably, on just how the notions of processing, carrying, or dealing with information should be cashed out. But they share a pair of core commitments: first, that cognitive systems’ dealings with meanings should be explained in terms of the ways in which those systems process, encode or deal with information; second, that the notion of ‘information’ in question can be made sufficiently formally and mathematically precise to serve as a foundational load-bearing concept in cognitive science. Endorsing these commitments makes an approach ‘informationalist’ according to the taxonomy I am sketching.

The ‘organicist’ alternative to ‘informationalism’ about mind and meaning holds that the meaning or significance of psychological states fundamentally depends on the way in which those states are embedded within living activity. The ‘organicist’ label is typically used for approaches in theoretical or philosophical biology that combine two commitments: first, organisms have emergent, system-level properties that cannot be adequately explained or understood solely in terms of the properties of their parts and their interactions – the powers and properties of the whole are greater than the sum of their parts; second, the existence of such emergent powers and properties is compatible with a materialist or physicalist conception of what organisms are made of Gilbert & Sarkar (2000) and Rupik (2024). Applied to the mind sciences, we can use this label to pick out approaches that hold that the dealings with meaning involved in perception, agency and understanding must be explained in terms of their relationship to organisational properties distinctive of living things.

Informationalists can disagree about what is involved in carrying, processing, encoding or picking up the information on which the meanings of psychological states depend; organicists can disagree about which organisational properties of living things give rise to dealings with meaning, and how they do so. Most closely associated with enactivism is the claim that the self-producing dynamics of metabolic organisation give rise to primitive forms of teleology and normativity (Di Paolo, 2005; Godfrey-Smith, 2016; Jonas, 2001; Thompson, 2007). But other ways of specifying distinctively organic organisational properties and linking them to meaning are possible. Perhaps whole animals (Ginsburg & Jablonka, 2019; Godfrey-Smith, 2020; Merleau-Ponty, 1963), or animal/environment ecosystems (Lala et al., 2024; Reed, 1996), or cultural systems (Bruner, 1990; Stotz, 2010), have emergent organisational properties not found in the inorganic realm, and dealings with meaning obtain in virtue of these. However these properties are spelled out, organicists emphasise distinctive organic properties, ones that set living things apart from inorganic nature. This goes beyond merely appealing to organisational properties that are characteristic of living things – properties that must be appealed to in studying their minds, but which are also present in inorganic nature. More on this contrast below.

‘Cognitivism’ and ‘embodied cognition’ disagree about the spatial boundaries of the processes underpinning cognition. In characterising the informationalist/organicist partition in terms of a disagreement about what the meaning or significance of psychological states ‘fundamentally depends on’, my aim is to contrast it with this disagreement about spatial boundaries. Fuzzy talk of ‘fundamental dependence’ can be given stronger and weaker readings. Stronger readings imply commitment to a metaphysical claim about what meaningful psychological states essentially are – about what sorts of properties or processes, informational or organic, constitute genuine dealings with meaning. But weaker pragmatist or instrumentalist readings are possible: metaphysics aside, good cognitive scientific practice involves explaining the cognitive phenomena of interest via appeal to organisational properties distinctive of living things. The weaker reading is necessary and sufficient for membership of the ‘organicist’ grouping of approaches I have in mind here – some organicists may additionally wish to endorse stronger metaphysical claims about what minds and meanings essentially are, but this is an optional add-on to the weaker, methodological claim about good scientific practice.

4. Informational or Organic Embodied Cognition?

There is ample room for debate about how to place specific approaches with respect to the informational/organicist partition just sketched. In particular, different views about what ‘information’ or ‘life’ are (of which there are many) will give different distributions of approaches. The point of the partition is not that it automatically yields a neat and unequivocal sorting of methodologies for theorising about the mind; it is that adds a valuable extra dimension to our cartography of theoretical space, making us ask different questions and think about different things when placing approaches with respect to each other. In doing so, it will help show why the lump-and-split taxonomy fails to capture much of what is distinctive about the project set out in TEM. Crucially for present purposes, that project is organicist, locating the rudiments of mind in organisational properties of living things (more on which below). Most other embodied cognition approaches – including ‘sensorimotor’ and ‘radical’ enactivism – are at best agnostic on the organicist/informationalist split. Consequently, many approaches lumped together with TEM’s enactivism when we ask ‘does cognition causally depend on internal symbol-manipulation or brain/body/world relationships?’ are separated from it when we ask ‘do the meanings of mental states and processes fundamentally depend on informational or organic properties?’.

The differing consequences of focusing on the cognitivism/‘embodied cognition’ versus the informationalist/organicist partition are particularly clear when we consider ‘wide computationalist’ (Kersten, 2024; Wilson, 1994) or ‘extended functionalist’ (Clark & Chalmers, 1998; Clark, 2024) approaches. These retain a computationalist vision of mind but hold that the causal basis of those computations can span brain, body and world. They are classed alongside enactivism and other ‘embodied cognition’ views when we ask the first question, but belong with fellow computationalists like Fodor (1975) and Pinker (1997), split off from organicist approaches, when we ask the second (see Figure 1, below).

They are joined on the informationalist side of the partition by ‘embodied predictive processing’ views (Clark, 2024; Gallagher & Allen, 2018; Seth, 2014, 2021). These appeal to information flow within a layered stack of ‘generative models’ which gets its cognitive purchase on the world via each layer’s attempts to reduce the mismatch between its inflowing signal and its prediction of that signal’s structure. Many predictive processing fans have claimed kinship with enactivism due to their stress on the way in which cycles of bodily interaction can help align signals with predictions, distributing the prediction error minimisation process across brain, body and environment (Clark, 2015; Gallagher & Allen, 2018; Kirchhoff & Kiverstein, 2019). But, according to standard embodied predictive processing accounts, these bodily, active and worldly factors are important only insofar as they help implement the hierarchical pattern of information flow upon which cognition ultimately depends. Such accounts thus belong on the informational, not the organicist, side of the partition.

Those are two kinds of ‘embodied’ approaches that do not share the organicist side of the ledger with enactivism. What company does enactivism have on that side? Taking a long view, plenty of scientists and philosophers prior to the advent of cognitive science held that living organisms had organisational properties that rendered them qualitatively distinct from mere mechanisms, and that this fact was key to understanding their mindedness (see Harrington, 1996; Riskin, 2016; Rupik, 2024, for surveys). Moreover, descendants of such organicist views are increasingly visible in contemporary biology and its philosophy, as a new wave of approaches reacts to perceived limitations of the 20^th century’s mechanistic ‘modern synthesis’ biology by emphasising distinctive biological organisational properties (Bich, 2024; Fábregas-Tejeda, Baedke, Prieto, Radick, & Refstyled, 2024; Nicholson & Dupré 2018; Walsh, 2015). A recurring motif in this new wave is the claim that these organisational approaches do justice to a crucial fact that modern synthesis biology ignores, abstracts away from, or simply denies: organisms are agents, sources of directed activity (Ball, 2023). When these approaches hold that sui generis organisational properties explain organismic agency, or dealings with meaning or significance, they constitute an organicist approach to an aspect of mind.

Finding other clear companions for enactivists on the organicist side of the partition presents at least two kinds of challenge. Organicist approaches, I stipulated, hold that the meaningful aspect of psychological states fundamentally depends on organisational properties distinctive of living things. Many approaches appeal to life in their explanations of cognition, but leave room for debate about whether this appeal is fundamental to their explanation, and whether they are singling out properties distinctive of living things.

Consider the first challenge, of adjudicating the issue of ‘fundamental dependence’. Some approaches appeal both to properties of biological organisation and of information-processing or pickup. How one sorts these approaches with respect to the partition depends on which properties one takes to be in the explanatory driving seat. Organicists who appeal to information hold that information-processing is meaningful only in virtue of the way it inheres in living activity; informationalists who appeal to biology hold that the biological properties are relevant to cognitive science only insofar as they implement a kind of information-processing. But (having better things to do) theorists often leave their exact stance on these philosophical niceties unclear, leaving it indeterminate where their approach sits within my taxonomy. Categorisation as organicist or informationalist requires taking a stance on whether a particular theorist or approach takes biological or informational properties (or neither) to be more fundamental in their explanation of cognition.

We find examples of this challenge in the lively field of ‘basal cognition’ research (Lyon, 2006; Lyon et al., 2021). In an early touchstone paper, Pamela Lyon (2006) distinguishes anthropogenic from biogenic approaches to cognition. The former assume that distinctively human psychological capacities such as believing, reasoning, and categorizing are the paradigmatic instances of cognition; the latter assume that the rudiments of cognition are found in biological processes of self-organization, self-regulation, and self-perpetuation, and that the study of cognitive systems should thus begin with these processes. Lyon characterises anthropogenic approaches partly in terms of their assumptions that cognition is a matter of information-processing and the production and manipulation of content-bearing representational states (2006, pp. 14-15), contrasting them with biogenic emphases on biological organization and maintenance as the grounds of capacities for minimal forms of agency, meaning, value and anticipation. Anchoring cognitive scientific research in the study of these capacities as manifested in the simplest organisms that possess them is the key animating commitment of basal cognition (Lyon et al., 2021), and the biogenic/anthropogenic contrast that motivates it appears to align nicely with the organicist/informationalist contrast. However, most basal cognition researchers are sanguine about characterising cognition in informational terms, with Lyon and others (2020, Lyon et al., 2021) even writing it explicitly into a definition of the ‘minimal’ cognition that is the movement’s subject matter:

Cognition comprises the sensory and other information-processing mechanisms an organism has for becoming familiar with, valuing, and interacting productively with features of its environment [exploring, exploiting, evading] in order to meet existential needs. (Lyon, 2020, p. 416; original square brackets)

Whether an instance of basal cognition research leans organicist or informationalist is determined by which of the properties in this definition are taken to be in the explanatory driving seat: is an organism’s information-processing ‘minimally cognitive’ only because it inheres in the organism’s biological self-maintenance – the processes via which an organism satisfies its existential needs? Or are capacities for biological self-maintenance only ‘minimally cognitive’ insofar as they implement a form of information-processing? The former option is organicist, the latter informationalist. Different basal cognition researchers have different stances here, not always made explicit. The important point for present purposes is that one’s choice here determines the side of the informationalist/organicist partition on which one sits. Basal cognition research is compatible or incompatible with organicism insofar as it is compatible or incompatible with the former option.

The second kind of challenge for placing views with respect to the informationalist/organicist divide is presented by approaches that depict themselves as explaining cognition in terms of properties of living things while denying that living things have sui generis organisational properties. As I put it above, such views claim that the meaning or significance of psychological properties is explained by organisational properties that are characteristic but not distinctive of life. Consider cybernetic views of cognition (Ashby, 1960; Maturana, 1975; Seth, 2014), the forerunners of the contemporary predictive-processing approaches mentioned above. These emphasise recursive loops of sensorimotor feedback and self-regulation, and often present themselves as illuminating the properties of living organisation which give rise to cognition. But these views do not belong on the organicist side of the ledger insofar as they hold that there is nothing distinctively organic about such properties. Living systems, cyberneticists tend to think, are cognitive systems because they involve self-regulating cycles of feedback that are found throughout organic and inorganic nature, cycles which can just as well be implemented by non-living systems like autonomous drones, or suitably constituted and embedded LLMs. They thus aspire to explain psychological properties in terms of organisational properties that are characteristic rather than distinctive of living systems, and so are not organicists. This again runs counter to the lump-and-split taxonomy in which, because of their emphases on brain/body/world-spanning cycles of sensorimotor feedback, cybernetic approaches are often lumped together with TEM’s enactivism, then split off as a ‘sensorimotor’ subvariant due to their indifference to many of TEM’s key commitments (e.g. Ward et al., 2017). ⁵

Finally, the ecological psychology of James and Eleanor Gibson and their followers presents both challenges – determining whether organic or informational properties are in the explanatory driving seat, and determining whether proponents take the organisational properties of living things to set them apart from inorganic nature. Ecological psychology builds a reference to life into itself, since ecology is a branch of biology dealing with organism/environment interrelations. And many of its guiding commitments sit well with organicism: its emphasis on ‘natural vision’, a mobile, exploratory sensorimotor engagement that is ‘the kind of vision we need in life’ (Gibson, 2015, p. xiv) in opposition to ‘aperture’ or ‘snapshot’ conceptions of vision that model perceivers as inert camera-like mechanisms; and its insistence that the meanings in the environments of animals can only be understood relative to the ‘ways of life’ (Ibid, p. 3) in which they figure. But these built-in emphases on the active lives of animals do not yet amount to the organicist insistence that psychological phenomena fundamentally depend on organisational properties distinctive of living things. Whilst most ecological psychology is not incompatible with these claims, most ecological psychologists do not commit to them. In the opening pages of The Ecological Approach to Visual Perception, Gibson writes:

Biology begins with the division between the non-living and the living. But psychology begins with the division between the inanimate and the animate, and this is where we choose to begin. (Ibid, p. 3)

For Gibson (2015), the important things about living, embodied perceivers are that they move (hence their perception depends on more than the snapshot stimulation of their sensory apparatus), and that they have goals and interests (hence their perceptible environment is populated with perceiver-relative meanings: affordances and obstacles for their ways of life). These aspects of animal life are characteristic, but arguably not distinctive of, living things. For most practitioners of ecological psychology its methods apply with equal validity to inorganic mobile robots with sensors and pre-programmed goals as to a living organism.

These strands of ecological psychology are not organicist, but others are. As Heft (2023) notes in a useful survey, Edward Reed (1996) ‘explicity grounded his ecological psychology writings in Darwinian evolution and developmental biology’, (Heft, 2023, p. 54) in contrast with other theorists who look for ecological psychological first principles and explanatory vocabulary in thermodynamic physics (see e.g. Turvey, 2019). Organicist ecological psychology must follow Reed’s lead, drawing on concepts and methods distinctive of the biological sciences in its study of mind.

How about the other organicist requirement – that psychological capacities fundamentally depend on these distinctive properties, rather than on informational or other properties? As with basal cognition, ecological psychological explanations make frequent reference to dealings with information – though usually while stressing that this is a proprietary kind of ecological information which is ‘picked up’ rather than processed by perceivers (Gibson, 2015; Lee & Reddish, 1981; Turvey, 2019). Once again, determining the organicist credentials of a given strand of ecological psychological research depends on whether informational or organic properties are in the explanatory driving seat – is the pickup of information a genuine psychological property only because it is embedded in an organism’s life? Or are living organisms psychological subjects only because they are the kinds of systems that can pick up ecological information? The former option is organicist; the latter informationalist. I won’t attempt the forensic analysis required to pin down the commitments of individual ecological psychologists on this issue (Bruineberg et al., 2024 is a helpful start). Suffice to say that an organicist ecological psychology is possible, and that Reed’s work is a good place to seek its outlines.

The broader point is that, once again, attending to the informationalist/organicist partition reveals disagreements and choice points among ‘embodied cognition’ approaches that are obscured when we focus only on the location and structure of the processes on which cognition depends. The latter focus is helpful for seeing the emphases on embodied activity and organism/environment relations that unite enactivism with other embodied approaches, but not the organicism that divides it from most of them. The partial landscape of approaches to theorising about the mind as I’ve sketched it here looks like this (Figure 1):

Figure 1.

A partial landscape of approaches to theorising about the mind

Approaches in the upper-left quadrant represent cognitive scientific business-as-usual, explaining psychological states and processes via appeal to computational processes in the brain. The upper-right quadrant encompasses most work standardly lumped together under the heading ‘embodied cognition’. These approaches insist that good cognitive scientific explanations involve reference to some aspect of brain/body/world relationships, but they hold that the significance of these relationships for cognition obtains in virtue of the information that is carried or processed within those relationships. The ‘organicist’ approaches in the lower-right quadrant disagree, holding that good cognitive scientific explanation must appeal to distinctive organisational properties of living things – embodied interactions are cognitively significant in virtue of the living organisation in which they are embedded, not in virtue of any informational properties they might have. ⁶

The problem with ‘embodied cognition’ as a label is that it lumps together a diverse range of approaches with very different methodological and conceptual commitments. About 30 years ago such a label was useful, highlighting a skepticism of the ‘classical sandwich’ conception of the mind shared by different strands of anti-cognitivism. Pitting ‘cognitivists’ against fans of ‘embodied cognition’ evoked an appealing image of a motley band of rebels united in struggle against a cognitivist empire. But the power of the cognitivist empire (if there ever was one) has long since waned, as the strategy of modelling minds as rule-governed manipulators of discrete symbolic states has lost ground to work involving more powerful and lucrative flavours of computational modelling and machine learning (Chirimuuta, 2024). Simultaneously, the space of ‘embodied cognition’ theorising has become more crowded, with the rise of embodied predictive processing, basal cognition and organicist approaches, and the increasing sophistication of work in these and other ‘embodied cognition’ subfields has made their incompatible methodological commitments easier to see. The ‘empire vs. rebels’ mental model is no longer useful – there is no ‘cognitivist empire’, and there is no prospect of a unified ‘embodied cognition’ rebel alliance.

The problem with the ‘lump-and-split’ taxonomy of ‘enactivism’ is a microcosm of this problem with the ‘cognitivism’ vs. ‘embodied cognition’ partition. Focusing on shared emphases on embodied sensorimotor interaction and a shared antipathy to ‘cognitivism’ is outdated and obscures crucial differences. In particular, it obscures the fact that the organicism of the approach set out in TEM prevents it from being assimilated by informationalist approaches such as embodied predictive processing, unlike its putative ‘sensorimotor’ and ‘radical’ variants (see Seth, 2014; Hutto, 2018, for illustrative assimilations), and obscures the emphases on primacy and reciprocity that further demarcate it from other embodied approaches. So far I have only alluded to or asserted these claims about the distinctiveness of TEM’s enactivism. Now I will try to explain them.

5. Continuity

The namesake commitment of the enactive approach is that cognition does not fundamentally consist in representing a world whose properties are essentially external to and independent of the cognizing subject, but in enacting a realm of subject-relative significance. It is this commitment, and the way that it is unpacked and developed in TEM and subsequent works, that determines enactivism’s distinctive position in theoretical space. To see how and why, consider a canonical enactivist example – the hungry bacterium swimming up the sugar gradient (Thompson, 2007, 2011).

A unicellular bacterium as the enactivist views it is, like any living thing, a network of processes that continuously creates and sustains its own demarcation from its environment as a persisting entity. The bacterium’s metabolic activity, for example, involves taking in nutrients that drive the processes that replace the constituents of the cell wall that demarcate it from its surrounding molecular soup. What makes enactivism organicist is its focus on further specifying this organisational structure in dialogue with the biological sciences, and in putting that specification into contact with accounts of the bases of minimal forms of agency, awareness and cognition. In TEM this takes the form of an emphasis on the autonomy of living systems, where autonomy is a matter of the ‘closure’ of the network of processes that constitute a living system: each process in the network depends on at least one other to keep itself going. In the case of the bacterium, the processes via which it repairs its cell wall depend on the process of absorbing nutrients from its environment, which in turn depends on the processes that move the bacterium to sucrose-rich portions of its surroundings, which in turn depend on the processes that keep the bacterium’s cell wall and flagella intact (and so on). This emphasis on autonomous organization grows out of Varela’s own engagement with the biosciences: his earlier work with Maturana on the self-producing or ‘autopoietic’ structure of cellular life (Maturana & Varela, 1973/1980), and his attempts to generalise these insights into overarching ‘principles of biological autonomy’ (Varela, 1979/2025).

All this is relevant to cognitive science because enactivists hold that autonomy brings with it simple kinds of purposiveness and directedness that are the first hints of mindedness in the natural world. This is the sense in which life/mind continuity is at the heart of enactivism. Motivated by an interest in ‘staying close to biological reality’ (TEM p. 213) in their cognitive scientific framework, enactivists draw on organisational biology to argue that life consists in the ongoing maintenance of the closed network of processes that constitutes an organism. They then argue that this organizational structure brings with it cognitive properties and capacities. Just how and whether this crucial argument for life/mind continuity works is complex and controversial (see Weber & Varela, 2002, Di Paolo, 2005, Thompson, 2007, 2011, for different reconstructions; Barrett, 2017, Sachs, 2023, Nave, 2025, for critical discussion) but here is one way it can be sketched.

As an integrated network of processes, an organism like a bacterium constitutes its own locus of causal power – some happenings in the world are the results of its integrated activity, not just of the running of one or other of its constituent processes. ⁷ An autonomous entity is, by definition, a self-perpetuating one – and so the autonomous bacterium’s integrated activity of facilitating its own persistence is something it, the bacterium, is doing. Qua autonomous entity, maintaining this autonomous organisation is the bacterium’s raison d’être – no more autonomy, no more bacterium. Focusing on the organisational property that makes it alive thus reveals a package of cognitively salient properties of the bacterium that are otherwise invisible: it is doing something (keeping itself around), and its integrated activity is oriented towards this end. Moreover, this activity brings with it a minimal kind of sense or significance – because sugar facilitates the ongoing autonomy of the bacterium, it has an organism-relative meaning. It is good for the bacterium which needs to have it, and the bacterium’s activity unfolds in a way that is sensitive to this value and this need. Post-TEM, Varela and other enactivists referred to this generation of meaning or significance from the integrated activity of a living system as ‘sense-making’. And so, for the enactivist,

cognition is grounded on the sense-making activity of autonomous agents—beings that actively generate and sustain themselves, and thereby enact or bring forth their own domains of meaning and value. (Thompson & Stapleton, 2009, p. 23)

Different enactivists will add different details to this quick sketch. Most will also stress that the network of processes that constitutes the bacterium is precarious (i.e. without the continuous integrated functioning of the network its constituent processes would run down and the bacterium would dissipate into its environment), and adaptive (i.e. the network can modulate its own activity in ways sensitive to its conditions of viability) (Di Paolo, 2005; Thompson, 2011). Some will also argue that the link between cognition and biological organization must be grasped phenomenologically rather than merely deductively, informed by our self-awareness of the living, embodied aspects of our own subjectivity (Weber & Varela, 2002; Thompson, 2007; cf. Sachs, 2023). The important point for present purposes is that, however the details are spelled out, enactivists hold that cognition fundamentally depends on distinctive organizational properties of living things, and hence are organicists.

This appeal to biological organization is also key to understanding the stances on embodiment and information that make enactivists bottom-right quadrant theorists. The most basic forms of cognition are embodied in the self-sustaining give-and-take between organism and environment that comprises the sense-making activity of autonomous agents. Enactivists thus hold that ‘cognition is embodied action’, (TEM p. 213) inhering in the unfolding relationship between organism and environment, rather than a process that takes place solely within the organism. They thus belong on the right-hand side of the cognitivism/embodied cognition divide. And whilst TEM’s enactivism is sanguine about the potential explanatory utility of informational or computational models of cognition (more on which below), it insists that these models abstract away from the living sense-making on which cognition fundamentally depends (see e.g. TEM p. 101, 140). Informational and computational properties are relevant to cognition only in virtue of their inherence in living activity, so even enactivists who partly embrace informational or computational models of cognition belong on the organicist side of the organicist/informationalist partition.

6. Primacy and Reciprocity

Enactivism’s commitment to life/mind continuity is already enough to set it apart from most of the variants of ‘embodied cognition’ surveyed above, as an unequivocally organicist approach. Its commitments to what I have called ‘primacy’ and ‘reciprocity’ set it apart even further. These commitments are expressed in TEM’s first sentences:

This book begins and ends with the conviction that the new sciences of mind need to enlarge their horizon to encompass both lived human experience and the possibilities for transformation inherent in human experience. Ordinary, everyday experience, on the other hand, must enlarge its horizon to benefit from the insights and analyses that are distinctly wrought by the sciences of mind. (xv)

Lived experience has a kind of ‘primacy’ for enactivism in that it plays a distinctive and central role both as an explanandum and as a source of evidence. ‘Lived experience’ here should be understood broadly, encompassing every aspect of our felt awareness of our unfolding lives, in contrast to the narrow focus on the felt qualities of sensory perception that tends to characterise anglophone philosophy and science of consciousness. Your first-person awareness of your unfolding life involves, for example, a felt sense of yourself as an agent and a knower, not just as a perceiver. Enactivism’s emphasis on the way in which lived experience and cognitive science can mutually inform and transform each other means that the reciprocal influences between experience and the mind sciences are central to it.

A key motivation for TEM and subsequent enactivist works (see especially Thompson, 2007, 2014; Frank et al., 2024) is that the mind sciences as mostly currently practiced are failing to adequately encompass lived human experience. One manifestation of this is the alleged ‘explanatory gap’ between conscious experience and the explanatory toolkit of cognitive science. Now, as when TEM was published, most agree that the computational and informationalist models of mainstream cognitive science seem to provide little or no explanatory purchase on the felt texture of experience. Now, as then, one ‘naturalist’ response to this situation is simply to deny that there is any fundamental explanatory mismatch here: despite the compelling appearance of an explanatory gap, if we could simply think harder (Dennett, 1991, 2018), make more scientific progress (Churchland, 1996; Seth, 2021), or appreciate the way in which our cognitive machinery tricks us into misunderstanding our own experience (Dennett, 1991; Frankish, 2016), we would see that there is nothing left to explain. Parallel issues arise for our felt sense of ourselves as agents – autonomous loci of causal power, able to shape some happenings in the universe according to our wills. Now as then, many hold that science unmasks this sense of ourselves as free agents as an illusion – it is an inescapable byproduct of forces over which we have no control, as is everything we think, feel, and do (e.g. Pinker, 1997; Sapolsky, 2023). Finally, parallel issues arise for our conception of ourselves as knowers, thinkers, or understanders. Since the advent of cognitive science, sceptics have argued that informationalist models of intelligence miss something important about genuine intelligence or understanding (Goddu et al., 2024; Searle, 1980; Turing, 1950), whilst optimists have dismissed these worries as luddite or mystical. Now, as generative technology meets or exceeds ever more attempted benchmark operationalisations of intelligence, many are keen to assure us that turbo-charged versions of the fundamental processes underpinning human knowledge, thought and understanding are present in a shiny new product they would happily sell us (Agüera y Arcas, 2025).

In each of these domains – consciousness, agency, intelligence – there is an apparent mismatch between our lived experience and the explanatory toolkit of informationalist cognitive science. And there is a standoff between those who take that mismatch to point to a defect in our explanatory toolkit that needs remedied, and those who don’t, holding instead that the mismatch is something we must stoically accept, or is only apparent or temporary, or is not there at all. Enactivists are on the former side of this standoff. They hold that informationalist models and explanations abstract away from the living processes of sense-making on which minimal subjectivity, agency and cognition depend, and that understanding how our lived experience depends on distinctive organizational properties of our metabolic, animal and cultural nature is required to address the mismatch. Whether they are right in holding this is not something I aim to establish here – it is enough to note that they are distinctive in holding this, set apart from informationalist ‘embodied’ approaches and from many organicist ones.

One sense in which lived experience has ‘primacy’ for enactivism is thus that it is a primary explanandum: enactivism is motivated by an ambition to do better than informationalist cognitive science in encompassing lived experience. A second sense is the evidential or epistemic primacy of experience. Our lived experience of the world provides the data that our theorising aims to systematise and explain – it is the anchor that keeps our theorising moored to our lived reality. As Thompson (2014) puts it, ‘without consciousness there’s no observation, and without observation there are no data’ (p. 14). This kind of primacy provides the organising theme of Thompson’s most recent co-authored book The Blind Spot – lived experience is the foundation of scientific explanation, but the structure of scientific practices that are oriented towards quantification, prediction and control commit them to abstracting away from the felt character of experience as we live it (cf. Chirimuuta, 2024). To a certain kind of hard-headed scientific realist, this commitment to lived experience as evidential bedrock will be enough to disqualify enactivism as a credible position. Surely a sensible naturalistic worldview requires taking science and its posits to be a better guide to reality than the nebulous, quantification-resistant ‘lived experience’ that enactivists privilege? Enactivists disagree. Like feminist, critical, and pragmatist theorists of science (Brancazio, 2018; Chang, 2022; Longino, 1990), enactivists understand scientific disciplines as bundles of theory and practice whose existence and structure depend on the particular human historical and intellectual contexts in which they arose, as employing simplifications and abstractions that reflect contingent and shifting human capacities and needs, and as continually evolving in tandem with our lives and experiences. For enactivists this view of the sciences as embodied in human practices and institutions, each with their own contingent histories and trajectories, is more naturalist and realist than taking science to be a fixed and direct conduit to the true, unfiltered nature of things. As Frank, Thompson & Gleiser put it:

all science is always our science, profoundly and irreducibly human, an expression of how we experience and interact with the world. But our science is also always the world’s science, an expression of how the world interacts with us. Science strives to be a self-correcting narrative. A successful scientific narrative is made from the world and our experience of it evolving together. (Frank et al., 2024, p. xiv)

As the final sentence of the above quote hints, however, the primacy that enactivism ascribes to experience must be understood in the context of the reciprocity between our theorising about the world and our experience of it. Experience has ‘primacy’ for enactivism not as an immutable fixed point or infallible source of knowledge, but as the domain to which all our theorising is ultimately responsive. ⁸ As mature, reflective humans, the structure of the lived experience that we are trying to understand is malleable and shaped by whatever contingent and evolving cultural apparatus for making sense of things currently surrounds us. This includes scientific concepts. Every cognitive scientific concept is an abstraction that ignores some aspects of the rich lifeworld on which it is founded whilst accentuating others. Employing these concepts to make sense of ourselves and our experiences can subtly alter the structure of the lived experience itself – conceptualising yourself as depressed, extroverted, dopamine-addled, or insecurely attached foregrounds some aspects of experience, backgrounds others, and selectively nudges you towards particular patterns of thought and action. Enactive cognitive science essentially involves recognising and attempting to work within this ‘fundamental circularity’ (TEM) or ‘strange loop’ (Frank et al., 2024) whereby our experiences and our scientific and other tools for making sense of them continuously reshape each other in an ongoing dialogue.

As an illustration, consider enactivism’s evolving engagement with Buddhist philosophy and contemplative practices. In TEM the ‘Buddhist tradition’ and its relationship to enactivism is introduced as a ‘hands-on, pragmatic approach with which to complement science’, which shows that ‘experience itself can be examined in a disciplined manner and that skill in such an examination can be considerably refined over time’ (TEM p. xviii). But, as Thompson notes in his introduction to the revised edition, some passages in TEM imply that Buddhist mindfulness practices are uniquely placed to yield direct and privileged insights into objective structures of lived experience, and that these insights should dictate the explananda of enactive cognitive science. These framings are in tension with enactivism’s aspiration to ‘forge a mutually enlightening and transformative relationship between cognitive science and human experience via a pragmatic and open-ended phenomenology of embodiment’ (TEM p. xxi, my emphasis). Giving Buddhist articulations of subjectivity the first and last word as to cognitive scientific explananda leaves little scope for Buddhist phenomenology to learn from cognitive science, nor for transformation or open-endedness in the structures of experience it alerts us to.

The corrective is to stress ‘another, better conception of mindfulness meditation’, (TEM p. xxiv) also present in TEM, according to which ‘mindfulness practices should be understood as skillful ways of enacting certain kinds of embodied states and behaviors in the world, not as inner observation of an observer-independent mental stream’ (TEM pp. xxiv-xxv). One strand of Thompson’s subsequent work (e.g. Thompson, 2020) explores this more thoroughgoing enactivism about mindfulness practices, emphasising the concrete historical and cultural contexts in which they arise, and their entanglements with co-evolving cultural, philosophical and soteriological frameworks. Reports of mindfulness practitioners should be taken not as privileged insights into primal experiential truths, but as elements within an ongoing process via which new patterns of experiencing and understanding are enacted amidst a particular historical and cultural context. And what is true of the vocabulary in which Buddhist doctrines or practices invite us to understand ourselves is equally true of the vocabularies offered by phenomenologists, cognitive scientists and other cartographers of subjectivity. No theoretical or descriptive vocabulary can offer us a comprehensively accurate map of the structure of own experience for two reasons. First, because a map is not the territory – it is an artefact designed for particular users with specific goals and needs, in the service of which it selectively attenuates and accentuates aspects of the terrain. Second, because the experiential terrain we are mapping shifts under our feet as our activity gradually sculpts its structure, tramping down new paths and entrenching or breaking away from old ones. The enactivist aspiration is to elucidate and study this evolving tangle of processes – living, experiencing and thinking amidst a changing world – whilst recognising our own current place within it.

Writing in 2004, reflecting on Varela’s passing and the early reception of TEM, Thompson noted the centrality of these themes of primacy and reciprocity to Varela’s life and thought, and their comparative lack of uptake among both friendly and critical readers (Thompson, 2004, pp. 380-2). These emphases on primacy and reciprocity further separate TEM’s enactivism from most other embodied approaches, including its putative radical and sensorimotor siblings. None of these make the reciprocal and reflexive relationship between lived experience and our available theoretical frameworks so central; nor do they insist on the primacy of lived experience, as the shifting ground to which all our theorising is ultimately responsive.

Before turning to the payoffs of this enactive conception of the relationships between life, experience and theory, note a final important consequence. Because enactivism seeks to understand lived experience, the culture-borne modes of theory and practice that shape it, and the interaction between these domains, its explanatory vocabulary should not be restricted to concepts drawn from the study of simple biological organisms. In characterising ‘organicism’ above, I said that organicists may focus on various levels of organic nature – whole animals, animal/environment ecosystems, or cultural systems may have the kinds of distinctive organisational properties to which organicists appeal. Varela, Thompson & Rosch already recognise the importance of understanding the relations of emergence and interaction between these different levels of organization (see e.g. TEM pp. 173-179). TEM’s project commits it to attempting to understand how culture-borne systems of meaning emerge from cycles of interaction between living things and their environments, and how those emergent cultural patterns reconfigure the lives and experiences of their participants. This daunting task is likely one reason why enactive emphases on primacy and reciprocity have had less uptake than those on life/mind continuity and sensorimotor interaction; it is easier to tweak or point approvingly to existing cognitive scientific work stressing life, embodiment or sensorimotor interaction than to explicitly work towards a comprehensive vision of the relationships between biological, animal and cultural life. But enactivists needn’t start from scratch here. Copious work from anthropology, history, critical and feminist theory, and other disciplines speaks to these themes. Recent years in particular have seen much important work on organicist-friendly approaches to animal and cultural life drawing on the biological sciences (Bich, 2024; Ginsburg & Jablonka, 2019; Godfrey-Smith, 2020; Lala et al., 2024). None of this work, however, proceeds by merely re-applying theoretical tools forged for describing distinctive organizational properties of simple organisms to animal and cultural levels. One reason for lobbying for plain, unqualified ‘enactivism’ as a label for TEM’s approach is that modifiers standardly used to pick it out (‘autopoietic’, ‘autonomous’, ‘biodynamic’) obscure the need for a pluralist approach that does not restrict itself to simple biological concepts in its characterisation of the full breadth of human life and experience.

7. Payoffs

Enactivism as outlined in TEM is an organicist approach to theorising about the mind that combines emphases on life/mind continuity, the primacy of experience, and the reciprocal relationships between experience and culture-borne practices for understanding ourselves. The claim that minds are enacted – emergent from cycles of interaction between living things and their environments – is the thread running through these emphases. Putting our lived experiences into satisfying contact with the depictions of ourselves offered by the mind sciences requires understanding the relationships between experience and the distinctive organizational structure of the processes that we, as living beings, fundamentally are. These processes involve the active production, or enaction, of the meanings that populate our lived experience. For humans these life processes outrun the self-producing dynamics shared by all living things, and the richer sensorimotor structures shared by all animals. Iterated cycles of interaction between embedded human animals bring forth culture-borne meanings, such as those that structure the self-understandings suggested by cognitivism, or embodied predictive processing, or Husserlian phenomenology, or Madhyakama Buddhism. These ways of understanding ourselves are founded in a lived experience that they also reshape by nudging us into new lifeways. Enactivism tries to understand our own nature as participants in this complex processual web.

Above I have made the case that this combination of commitments separates enactivism from other variants of ‘embodied cognition’ with which it is often lumped together. But why care about how to carve up the space of embodied and enactive theorising? Readers who have made it this far likely have their own answers to this question. But there are several payoffs of clarifying the shape and scope of TEM’s enactivism in this way, and I’ll conclude by briefly noting some of them.

Most simply, current debates about enactivism and embodied cognition would often benefit from greater clarity about what is distinctive about TEM’s approach. Lumping that approach together with others that are indifferent or hostile to claims about life/mind continuity, experiential primacy and science/self-understanding reciprocity impedes progress within the philosophy and science of embodied cognition. If we want to know whether enactivism is merely a new name for old or outdated ideas (Crippen, 2025), a degenerating research program (Meyer & Brancazio, 2023), a philosophy of nature (Gallagher, 2018), or something else entirely, it pays to be clear what we are asking about. The answers will differ according to whether, following the ‘lump-and-split’ taxonomy, we take ‘enactivism’ to label any approach that emphasises the fundamental importance of sensorimotor interactions to the study of mind, or whether we identify it with the distinctive suite of commitments summarised above.

Consider, for example, recent debate over whether enactivism should be construed as a ‘philosophy of nature’ (Gallagher, 2018; Meyer & Brancazio, 2022, 2023). The latter term was introduced by Peter Godfrey-Smith as part of an attempt to clarify the status of Developmental Systems Theory (Oyama, 2000; Griffiths & Gray, 1994), itself a close cousin of enactivism insofar as it stresses the active contributions organisms make to their own evolution and development and the neglect of these contributions by informationalist approaches within biology. Godfrey-Smith distinguishes between (i) a minimal notion of ‘scientific research program’, which involves commitments to a proprietary set of concepts for doing and describing scientific work, claims about what kinds of scientific work will and will not prove fruitful, and a set of core empirical claims, and (ii) a ‘philosophy of nature’, a commentary on the ‘overall picture of the natural world that science, and perhaps other types of inquiry, seem to be giving us’ (2001, p. 284).

Is enactivism a scientific research program or a philosophy of nature? If ‘enactivism’ is taken to involve the commitments to continuity, primacy and reciprocity I have outlined here, then it involves elements of each project. It has properties that Godfrey-Smith associates with a scientific research program: it is committed to the necessity of concepts that can capture the distinctive organizational properties of living processes on which sense-making depends, and to the claim that construing cognition as a matter of information-processing or pickup is not up to this task; it is committed to the claim that scientific work that ignores or abstracts away from the organic properties on which sense-making depends will bear, at best, limited fruit in illuminating awareness, agency and meaning as they figure in our lived experience; and it is committed to at least the empirical claims that living things have distinctive organizational properties, that these cannot be adequately captured in informational terms, and that these properties underpin the most minimal forms of cognition in living things.

Enactivism also, however, comprises the kind of ‘overall picture of the natural world’ inspired by science and other sources that make it a ‘philosophy of nature’ in Godfrey-Smith’s sense. TEM’s enactivism draws on organisational biology, phenomenology and Buddhist thought and practice to depict the natural world as a processual flux, out of which temporarily stable patterns of organisation emerge at metabolic, sensorimotor and cultural levels. The most minimal instances of these self-organising patterns are the basis of agency, meaning and experience in living things. The meanings that structure our human lives, however, are not limited to those emerging from the basic self-sustaining dynamics of biological organisation. Patterns of human interaction bring forth culture-borne meanings that shape our subjectivity, including the meanings of the scientific and other vocabularies in which we try to understand ourselves.

Enactivism’s commitment to working within this ‘fundamental circularity’, wherein we try to catch ourselves in the act of bringing forth the complex form of enculturated subjectivity we are attempting to understand, has important consequences. I noted above that it places enactivists alongside feminist, critical and pragmatist theorists of science in understanding scientific disciplines as shifting bundles of theory and practice, ones which employ simplifications that reflect the particular contexts in which they arose, and the practical purposes that they serve. This conception of scientific theory and practice as enacted has an underappreciated upshot. ‘Enactivism’ (construed according to the lump-and-split taxonomy) and other sub-types of embodied cognition are often associated with a strong aversion to representational or informational talk in cognitive scientific explanation. But the position set out in TEM is more nuanced. I noted above that enactivism is so called due to its conviction that mental processes arise from the enaction of a realm of subject-relative significance, rather than consisting in patterns of information-processing or representational states instantiated by those patterns. Importantly, however, this conviction is compatible with acknowledging the indispensability of informational or representational explanations in many contexts. The authors of TEM, for example, do not attempt to ban or explain away talk of computations over symbolic states in cognitive scientific explanations; instead, ‘the need for a symbolic level is acknowledged’, (p. 101) but in the context of an insistence that symbolic computational characterisations of cognition are ‘approximate macrolevel descriptions’ (p. 102) which abstract away from the living sense-making on which cognition fundamentally depends. Similarly, they advocate not for a rejection of information-processing talk in cognitive science, but of the idea that the information appealed to in such explanations ‘exists ready-made in the world’, (p. 140) as opposed to existing as an emergent structure of the interactions between living things and their environments. ⁹ In this way the considerable explanatory and predictive successes of informationalist approaches need not be denied or downplayed; the enactive approach instead attempts to place those successes within a wider context, one that is sensitive to the idealisations they involve and interests they serve, and to their relationship to the living processes on which dealings with meaning are ultimately held to depend.

This attitude to informationalist concepts exemplifies an inclusive and pluralist attitude to different methodologies and explanatory vocabularies in the mind sciences which is entailed by the structure of enactivism as I have presented it here. Because of its commitment to reciprocity – to emphasising the ways in which lived experience and cognitive science mutually inform and transform each other – enactivism cannot consistently be read as offering a final vocabulary for our cognitive science or for our own self-understanding. By enactivism’s own lights, any such vocabulary reflects at best a partial take on some aspect of lived experience, indexed to a particular time and place. This applies even to enactivism’s own characterisations of biological organisation and its links to the sense-making activity it takes to be the basis of living subjectivity. According to a consistent enactivism, claims about the basis of minimal subjectivity in adaptive autonomous living systems are not immutable truths but partial perspectives on the relationships between mind and nature. This modest attitude to its own claims is nonetheless consistent with a commitment to organicism over informationalism: enactivists can and should hold that their conception of cognition as living sense-making has particular value at a time when the dominance of informationalist modes of science and self-understanding obscures the links between lived experience and our living embodiment. In the last section, I briefly noted the apparent gaps between our experiences of consciousness, agency and meaning on the one hand, and informationalist accounts of these phenomena on the other. I also noted that enactivism claims to offer bridges across those gaps unavailable to informationalists. If enactivists are right about this, and if we want a science of subjectivity that makes satisfying contact with our experience as we live it, then we have good pragmatic reasons to hold that the enactivist treatment of subjectivity enjoys a kind of priority over informationalist alternatives. Although the enactivist theory of subjectivity as living sense-making is, like all such theories, partial and provisional, it reveals and clarifies important things about ourselves that informationalist alternatives conceal.

Let me try to elaborate on the above by noting two final payoffs of the understanding of enactivism presented here. First, return to the question of whether enactivism is better construed as a research programme or as a philosophy of nature. I said above that, according to Godfrey-Smith’s definitions of these terms, enactivism embodies elements of each. We are now in a position to see that it also problematises the distinction between them in a way that other embodied approaches do not. We have just seen that, whilst it gives priority to organicist over informationalist approaches to subjectivity, enactivism should not be construed as attempting to ban or straightforwardly refute alternative explanatory vocabularies or modelling strategies. And we have seen that this is so, in part, because of the conception of scientific knowledge and practices as themselves enacted that it recommends. But whilst these features of enactivism appear to speak in favour of construing it as a philosophy of nature, we have also just seen that its commitment to the reciprocal circulation of ideas between science and self-understanding will, in certain contexts, lead it to favour some scientific vocabularies and methodologies over others – if we value a scientifically-informed self-conception that does not alienate us from our lived experiences of consciousness, agency and meaning, we must retain a place for the kind of organicist account of subjectivity that enactivism recommends, and understand the ways in which informationalist accounts of minds abstract away from the crucial structures of living sense-making. Enactivism, because of its commitment to reciprocity, thus blurs the boundaries between empirical research program and philosophy of nature. Inspired by the science of self-organising systems and the shortcomings of cognitivism, TEM works toward an overall picture of how subjectivity and (eventually) the theoretical and scientific projects of reflecting on and studying it emerge in the world – a philosophy of nature. But that picture should reciprocally influence our scientific practice. Insofar as science aims to help us understand ourselves and our place in the world, it should prioritise organicist frameworks for understanding subjectivity, and should remain attentive to the ways in which informationalist approaches abstract away from the living basis of our minds – a constraint on empirical research programs. The project of TEM thus essentially involves an attempt to reduce the tension between the way in which we live, experience and understand our lives and the way in which science invites us to understand ourselves. The consonance of our scientifically-informed self-conception with our lived and common-sense self-understandings is an ideal that we should work towards, and a distinctive virtue of enactivism is that it suggests a way of going about this.

There is room for disagreement about the desirability and feasibility of this animating ideal. Perhaps the idea that science aims to contribute to self-understanding is outdated romanticism; its proper aims are only to facilitate prediction, control, and economic growth. Perhaps there is no real gap to close between our experiences of ourselves as conscious agents living meaningful lives and the informationalist vision of the universe as a vast networked churn of zeros and ones. ‘Consciousness’, ‘agency’ and ‘meaning’ as we spontaneously understand them might be mere self-reflexive illusions that clumps of information-processing like us are prone to undergo. Or they might be genuine phenomena, but comprehensible only to a system with greater computational power than our lowly living minds – we ourselves are stuck with a gap between informationalist science and self-understanding that must be stoically accepted as unbridgeable for us. Or perhaps enactivists are right that there are gaps between consciousness, agency and meaning as we live them and the kinds of processes that can be instantiated by inorganic information-processing. But since they are also right that the structure of our lived experience gradually shifts in response to the changing cultural apparatus for self-understanding that surrounds us, these gaps will take care of themselves. As the informationalist vision of the world becomes ever more economically and ideologically dominant we can simply lean in to it. Eventually we will collectively forget that anything ever seemed to be missing from this vision; vestigial notions of ‘lived’ subjectivity will fade away as we relax into the techno-rapture.

All these options are well-represented in current thinking about the relationship between human subjectivity and information-processing. The final payoff of the conception of enactivism offered here is that it offers a clear alternative, one that has only become more timely since TEM was written. Since then our relationship to computation and networked information has become more entangled and fraught. As information technology increasingly permeates and structures our lives, our cultural context makes it ever more inviting to understand ourselves, our minds, and our place in the world in informational terms. As the conception of ourselves as sophisticated information-processing devices has gained traction, questions of the relationships between sentience, agency, value, and organic processes have taken on greater urgency as debates about the nature and social role of ‘artificial intelligence’ intensify, and as networked global capitalism’s effects on the biosphere become ever more pronounced.

Enactivism suggests a different perspective on the relationships between mind, life and culture from the one encoded in most contemporary cognitive science and its philosophy. The picture that emerges looks roughly like this: the project of understanding human subjectivity and agency as dependent on complex patterns of information-processing flowing through nervous systems becomes available and appealing within a specific social, ideological and technological context, one that we can articulate and scrutinise. This scrutiny, coupled with attention to the texture of our experiences of consciousness, agency and meaning can give us a clearer view of the prospects and the limitations of this project. In particular, it can help us see how it differs from an alternative way of understanding the place of mind in nature: in terms of organisational properties of living things and their environments. Informationalism distorts or abstracts away from these properties. According to enactivism this alternative understanding can accommodate essential facts about consciousness, agency and meaning that informationalism cannot. The lesson is not that informationalism is wrong, or to be unequivocally discarded; it is that it is partial, and that failing to see its partiality obscures important aspects of ourselves and our relationship to the world that enactivism helps us keep in clearer view. Although the power of the classical cognitivism with which enactivism is often contrasted waned long ago, informationalism has only become more dominant, threatening to crowd out alternative visions of how minds and meanings arise in the world from popular, scientific and philosophical imagination. The most important payoff of distinguishing TEM’s enactivism from neighbouring variants is that, whether we agree or disagree with its diagnoses, it offers us a unique perspective on this situation, its prospects, and its perils.