What Are We Explaining?

The set of readings in Enacting Life draws on several sources. Francisco Varela, with Evan Thompson and Eleanor Rosch, set out the enactive approach in The Embodied Mind (Varela et al., 1991). Later writers, including Thompson (2007) and Ezequiel Di Paolo (2005), developed it in new directions. But behind much of this lies earlier work by Humberto Maturana, some of it with Varela, on the organization of living things. This reading concentrates on Maturana. His line of thought differs in important ways from the enactivism that grew out of it (see Villalobos & Ward, 2015), and it is a distinct and influential perspective in its own right.

Maturana began as a biologist studying how the frog’s eye responds to light, and spent much of his later career asking what it means for an observer to describe anything at all. That may seem an odd progression, but it is the thread of this reading. Along the way, Maturana came to draw a distinction between how a living system operates and how an observer describes what it is doing.

I have been drawn to Maturana’s work without being entirely sure why. Part of it, I think, is that he kept pressing on what we use to explain things, and asked us to be clear about what that commits us to. It also connects to an argument I made some years ago about how psychology and neuroscience relate: reducing psychology to neuroscience would not give a satisfactory account of the relation, and neither would simply calling them different levels of analysis (Marshall, 2009).

Psychology and neuroscience may seem to explain the same things in different terms. One kind of account describes mechanisms, such as how cells in the retina respond to changes in light. Another describes conduct: the frog sees a fly and catches it. It is easy to assume that these are two accounts of one thing, and that the second will eventually reduce to the first. Maturana asks whether they answer the same question at all. He keeps the two accounts distinct while anchoring both in the living organism: an account of mechanism explains how the organism operates, and an account of conduct describes what it does in relation to its surroundings. Both kinds of account are found in psychology and in neuroscience, so this is not a division between the two fields.

In the 1950s, Maturana worked with Jerome Lettvin, Warren McCulloch, and Walter Pitts on the frog retina (Lettvin et al., 1959). Recording from single fibers in the optic nerve, they found several classes that responded selectively to different local patterns of change in illumination, including moving edges and small moving dark objects. Rather than relaying a point-by-point image to the brain, the retina’s circuitry was already transforming patterns of light into patterns of neural activity.

That 1959 paper used the conventional language of information, but Maturana later became critical of treating sensory activity as the transmission of information about an independently specified world (Maturana, 2002). Part of the reason was his work on color vision in the 1960s. He expected to relate retinal activity to the wavelengths of light, but that relation proved elusive, and very different patterns of light can be seen as the same color (Maturana et al., 1968). What we call color depends on the organization and activity of the nervous system, not on wavelength alone. That unsettled the picture in which the visual world is encoded at the eye and then reconstructed inside the brain.

The frog study was part of early work in cybernetics, which studied control and communication in animals and machines, described from the outside. Maturana later worked at Heinz von Foerster’s Biological Computer Laboratory at the University of Illinois, where his attention turned to the observer doing the describing. That shift is associated with what became known as second-order cybernetics, the study of observing systems.

Did the Frog See a Fly?

The 1959 paper suggested that one class of optic nerve fibers might be called “bug perceivers.” The name made sense. These fibers responded best to a small dark object moving across the frog’s field of view, which is roughly what a flying insect looks like from where a frog sits. But the same paper noted that a frog will starve surrounded by food that does not move (Lettvin et al., 1959). The authors were careful about what this meant: the frog “does not seem to see or, at any rate, is not concerned with” the stationary detail around it. Either way, a fly that does not move does not draw a strike.

So did the frog see a fly? From where we stand, as observers who know what flies are, it is natural to say that it did, and most of the time the description fits: the tongue goes out and a fly is caught. But the fibers do not respond to flies. They respond to small, dark, moving things, and the frog’s strike follows much the same pattern. “Fly” is our name for what the frog catches, and that category alone does not explain what draws a strike. It is not something carried into the frog’s nervous system as an instruction for what to catch.

None of this is a criticism of the 1959 paper, or of naming things. Science needs words for what it studies, and the people who use those words have to agree on them. That agreement is among observers, though. Whether it matches what the organism is doing is a further question, and the organism’s conduct is where the answer lies. Here the paper’s own observation, frogs starving among motionless food, already set “bug,” the observers’ category, beside what actually draws the frog’s strike. The tension is within the 1959 account itself. Asking that question of our descriptions makes explanations more exact.

A small dark spot moving across the frog’s eye can certainly affect the frog, and a psychologist can call it a stimulus. The trouble starts when the stimulus is treated as though it already contains its meaning, or specifies the response that should follow. In Maturana’s terms, the moving spot perturbs the frog. What happens next depends on the frog’s structure: the particular cells, connections, and muscles it has at that moment. The spot can trigger a strike only because the frog is built so that such a spot can trigger one. This is the sense in which Maturana describes living systems as structure-determined (Maturana, 1978).

This cuts across the stimulus-response language common in much twentieth-century psychology. An experimenter can identify and manipulate a stimulus, but for Maturana its effect is not specified by the stimulus alone.

A History of Fly Catching

What would it take to understand how a frog comes to catch flies at all? A tadpole does not catch flies. It feeds in the water, and the tongue strike becomes possible only as the body is rebuilt during metamorphosis. A historical account of fly catching would have to follow that development, and further back the evolutionary history of frogs and of the insects they eat, each of which has changed the conditions for the other. Maturana’s term for this kind of history is structural coupling: through recurrent interaction, an organism and its surroundings, including other organisms, change in ways that fit them to one another (Maturana, 1978).

None of that history is needed to describe how the retina and tongue work during a single strike. As Maturana put it, “History is necessary to explain how a given system or phenomenon came to be, but it does not participate in the explanation of the operation of the system or phenomenon in the present” (Maturana, 1978, p. 39). This does not make the frog ahistorical. Its history is carried in the structure it has now, the structure through which the next strike happens. History explains why the frog is built as it is; the present structure explains what happens when a spot moves across its eye.

The Cognitive Domain

Watch a frog for an afternoon and you could list what it does. It sits, shifts position, strikes at small moving things, swallows, and leaps away when a large shadow passes overhead. Maturana called the range of interactions in which a living system can act in ways relevant to keeping itself alive its cognitive domain (Maturana, 1970/1980a). For the frog, that domain includes striking at moving specks. It does not seem to include striking at a motionless fly. In his early formulation, living and cognizing are inseparable: a living system continues to exist through effective activity in its surroundings, and what it can do depends on how it is built.

We do not usually think of cognition as behaving, but Maturana and Varela compressed the point into the phrase “all doing is knowing, and all knowing is doing” (Maturana & Varela, 1987, p. 26). On this view, knowing is not first the possession of an internal representation that is then used to guide action. We say the frog knows how to catch flies because its strikes work.

When we say that the frog is hunting, sees its prey, or misses, we are describing its conduct. Those are useful descriptions, and the frog really does act: whether it eats or starves does not depend on anyone’s description. But “hunting” and “missing” relate the frog’s movements to something we have in mind, such as a fly that should have been caught. For Maturana, saying that an action succeeded or failed is a description of this kind. It should not be confused with the processes in the retina, brain, and muscles through which the strike is produced.

A neuroscientific account can tell us about those processes. A psychological description tells us what the organism is doing: perceiving something, learning a task, remembering a person, acting intentionally. One does not become the other simply by changing vocabulary. If we want to connect them, we need to show how the organization and activity of the nervous system generate the conduct that the psychological description picks out.

None of this limits what neuroscience can do. It asks that the categories we use to describe conduct be tested against what the organism does, not built into the mechanism in advance. Krakauer and Ramsey (2026) made a related point: a neuron’s reliable response to a stimulus shows that its activity carries information about the stimulus, but not that it functions to represent it, any more than a car’s spark plugs represent the position of the accelerator. They trace this way of reading neural responses back to the 1959 frog paper, while keeping representation as part of higher-level cognitive explanation. But Maturana, one of that 1959 paper’s authors, drew a different conclusion: that talk of representation belongs to an observer’s description of the animal’s conduct, not to the operation of its nervous system.

The Consensual Domain

Go back to the name “bug perceiver.” It came from four scientists talking with one another, and from a wider scientific community that already shared words like “bug,” “perceive,” and “information.” The name could then be checked, within that same community, against what frogs actually do. For Maturana, this is where description lives. When people describe conduct in terms of objects, goals, perceiving, or knowing, they make those distinctions in language, with one another. He called the domain in which this happens a consensual domain (Maturana, 1978).

When we talk with someone about something, we use words as part of a shared way of communicating, but for Maturana a consensual domain does not begin with people explicitly agreeing on the meanings of words. It develops through recurrent interaction in which their conduct becomes coordinated over time. One person’s action changes the circumstances for the other, whose response in turn changes what happens next.

Consider an infant reaching toward a toy and a caregiver handing it over. Across repeated exchanges, reaching, looking, offering, and taking can become an established way of requesting and passing objects between them. Now consider a child saying “not that one, the red one” as the caregiver offers a toy. The words adjust how an activity they are already doing together proceeds. This illustrates what it means to coordinate coordination: participants can change how they coordinate their actions through further activity together. Maturana places this recursive coordination at the basis of a linguistic domain. Language, on this account, is not fundamentally the transfer of information from one mind to another but an ongoing coordination of conduct between participants with a history of interaction (Maturana, 1978).

For Maturana, whatever we distinguish as an object is distinguished by an observer with a particular structure and history, and an observer is not a detached spectator standing outside the world being described. Human observing depends on language, and language depends on histories of coordinated interaction with other people. Maturana’s way of putting this was that everything said is said by an observer to another observer, who may be yourself (Maturana, 1978). “Bug perceiver,” “stimulus,” and “cognition” all have histories in the practices of the people who use them. This holds for neuroscience as well. “Neuron,” “spike,” and “synapse” are also distinctions made by observers in language, and there is no describing a nervous system from outside it. What Maturana asked was that we keep track of which domain a description belongs to: the relations among components through which the nervous system operates, or the conduct of the animal as we see it in its medium. A mechanistic explanation is still said by an observer, but it can be tested by asking whether the proposed mechanism, working on its own terms, would produce the conduct we see.

This does not mean that frogs or flies are invented by observers, or that there are no stable regularities in the world. We investigate a shared physical world together, but we do not do so from outside our own embodiment and language. This does not make science less objective. Objectivity is something observers achieve together, by checking their distinctions against what they find, as the 1959 authors did when they set “bug perceiver” beside frogs that starve among motionless food. For Maturana, then, the observer is not outside the biology. Describing and explaining are themselves the conduct of living beings.

Maturana’s view is sometimes read as solipsism, the idea that each of us is confined to a private world. He rejected that reading. Language, for him, requires at least two organisms whose conduct becomes relevant to each other, so the domain in which we describe and explain is social from the start: “The central feature of human existence is its occurrence in a linguistic cognitive domain. This domain is constitutively social” (Maturana, 1980b, p. xxiv).

His view is also not one that can simply be accepted from outside. If everything said is said by an observer, that includes what a reader says about Maturana. The view asks you to notice the distinctions you are making yourself, about frogs, stimuli, or minds, and to ask where they come from. Trying it out on your own descriptions matters more than agreeing with it.

Maturana’s ideas are hard going, and some of the difficulty lies in the ideas rather than in how he wrote them. The line between describing how a nervous system operates and describing what an animal does is not always clean in practice, since words like “receptor,” “signal,” and “error” carry meaning over from descriptions of conduct. His claim that everything said is said by an observer can also seem to leave nowhere firm to stand.

The main point of this reading does not depend on settling these difficulties. Explaining how an organism operates and describing what it does in its surroundings are different kinds of account, and both are made by observers about a living organism.

For research, none of this means giving up experiments or mechanistic explanation. It means checking the categories we bring against what the organism does, as the 1959 authors did with “bug perceiver,” and being clear whether a claim is about how a nervous system operates or about what an animal is doing, so that a word like “represents” does not slide between the two. The connection between the two accounts then becomes something to be shown, by working out how the operation produces the conduct. It is not shown just because neural activity correlates with a task. In a typical study, the task conditions are defined by the researchers, who decide in advance that one condition involves recognizing a face and another does not. A brain region whose activity differs between the conditions is tracking a difference the researchers built into the design. That is a real finding, but it does not show that the region recognizes faces, or that “recognizing a face” is how the organism’s own activity is divided up. The frog’s fibers responded reliably when small flies moved past, yet the same paper’s frogs ignored motionless food, which showed that the category “fly” alone does not explain what draws a strike. Showing the connection means asking how the activity contributes to what the organism does, including when it does something the design did not anticipate.

The approach can also help in choosing what to study. The most revealing observation in the 1959 paper was the frog starving among motionless food, where the organism’s conduct and the observers’ category came apart. Places like that, and the histories that made an organism what it is, are often where the worthwhile questions are.

When we say that a frog’s strike succeeded or failed, is that only an observer’s description, or can the success or failure be the frog’s own? Maturana’s framework, which keeps the frog’s operation apart from our descriptions of it, leaves that question open. What Makes Living Things Different? approaches it through what it takes for a living thing to keep itself going.

References

Lettvin, J. Y., Maturana, H. R., McCulloch, W. S., & Pitts, W. H. (1959). What the frog’s eye tells the frog’s brain. Proceedings of the IRE, 47(11), 1940–1951. https://doi.org/10.1109/JRPROC.1959.287207

Maturana, H. R., Uribe, G., & Frenk, S. (1968). A biological theory of relativistic colour coding in the primate retina. Archivos de Biología y Medicina Experimentales, 1, 1–30.

Maturana, H. R. (1980a). Biology of cognition. In H. R. Maturana & F. J. Varela, Autopoiesis and cognition: The realization of the living (pp. 5–58). D. Reidel. (Original work published 1970)

Maturana, H. R. (1980b). Introduction. In H. R. Maturana & F. J. Varela, Autopoiesis and cognition: The realization of the living (pp. xi–xxx). D. Reidel.

Maturana, H. R. (2002). Autopoiesis, structural coupling and cognition: A history of these and other notions in the biology of cognition. Cybernetics & Human Knowing, 9(3–4), 5–34.

Maturana, H. R. (1978). Biology of language: The epistemology of reality. In G. A. Miller & E. Lenneberg (Eds.), Psychology and biology of language and thought: Essays in honor of Eric Lenneberg (pp. 27–63). Academic Press.

Maturana, H. R., & Varela, F. J. (1987). The tree of knowledge: The biological roots of human understanding (R. Paolucci, Trans.). Shambhala.

Marshall, P. J. (2009). Relating psychology and neuroscience: Taking up the challenges. Perspectives on Psychological Science, 4(2), 113–125. https://doi.org/10.1111/j.1745-6924.2009.01111.x

Krakauer, J. W., & Ramsey, W. (2026). Mental representation without neural representation: Understanding the evidence. Philosophy and the Mind Sciences, 7. https://doi.org/10.33735/phimisci.2026.12204

Varela, F. J., Thompson, E., & Rosch, E. (1991). The embodied mind: Cognitive science and human experience. MIT Press.

Thompson, E. (2007). Mind in life: Biology, phenomenology, and the sciences of mind. Harvard University Press.

Di Paolo, E. A. (2005). Autopoiesis, adaptivity, teleology, agency. Phenomenology and the Cognitive Sciences, 4(4), 429–452. https://doi.org/10.1007/s11097-005-9002-y

Villalobos, M., & Ward, D. (2015). Living systems: Autonomy, autopoiesis and enaction. Philosophy & Technology, 28(2), 225–239. https://doi.org/10.1007/s13347-014-0154-y