Why We Narrate: Metabolic Stakes, Interrecursive Complexity, and the Consequential Past

Abstract:

Narrative is not primarily for information or entertainment, but a solution to a specific evolutionary problem: coordinating with recursively responsive beings whose past actions remain consequentially present. Its central original claim is that interrecursive complexity, the expanding history generated when recursive beings become conditions for one another’s later responses, resists contextual stripping. Portable summaries lose the very relational context that makes later action intelligible, so narrative functions as symbolic contextual retention through time. The essay introduces metabolic buffering to explain why audiences can engage dense interrecursive stakes without direct exposure, and it thereby derives social learning, reason explanation, memory transmission, and narrative’s attraction to being-with from a single architecture: metabolic stakes, recursive counterparts, interrecursive complexity, and mixed pastness and futurity. Narrative exists where consequential pasts remain recursively open for later response. It thus makes absent, dead, unborn, and fictional histories available for present coordination.

I. Introduction: Why Narrative at All?

Human beings devote extraordinary quantities of time and attention to recounting events that are no longer happening. They tell of people who are absent, dead, ancestral, heroic, disgraced or remote, and of people who never existed at all. Children demand stories long before they could say what a story is, and they object at once when a familiar one is told out of order. Epic traditions have carried, across centuries of oral and then written transmission, many thousands of lines concerning quarrels that had ended generations before the first singer we know of was born. Societies of every scale invest in narrative a quantity of memory, training, performance time and material support that would be hard to justify if its only return were entertainment, and equally hard to justify if its only return were information, since most of what stories carry could in principle be stated far more economically.

The usual question about this phenomenon is framed too late. It asks what narrative does: whether it entertains, transmits information, simulates social situations, teaches cooperation, preserves identity, transmits norms or exercises the capacity to attribute minds to others. Each answer captures something, and each takes narrative as already given, a finished capacity whose uses can be listed. The prior question is what kind of problem would make something like narrative necessary in the first place. A capacity so widespread, so costly and so early to appear in each human life invites the hypothesis that it answers a recurring problem rather than having accumulated as a collection of happy side effects. The argument that follows offers a functional-evolutionary explanation of that kind. It proposes what problem narrative is fitted to solve and why the solution takes the form it does, without claiming to reconstruct the prehistoric sequence by which narration actually emerged, for which direct evidence is scarcely to be had.

The problem can be approached through a simple contrast. A mountain has a history. A person has a history. A relationship has a history. These are histories of different ontological kinds. A mountain's present shape may require an elaborate reconstruction of uplift, erosion, glaciation and collapse, and the geologist who reconstructs it tells something with a beginning, a middle and a present state. Yet nothing that happened to the mountain mattered to the mountain. By contrast, understanding why a person now behaves toward another in a particular way may require recovering what one did to the other, what the other took to have happened, how the first responded to that taking, and what was promised, remembered, misunderstood or left unresolved along the way. The history of a relationship is made of responses to responses, and each participant carries a different version of it into whatever happens next.

Narrative, wherever it appears, is overwhelmingly attracted to the third kind of history. The argument of this article is that the attraction is neither accidental nor a mere expression of human interest in humans. Narrative is especially adapted to a particular kind of past: the consequential past generated among recursively responsive beings, a past whose relevance to the present cannot be extracted from the context in which it was produced without destroying its intelligibility. Narrative, on this account, is less a representation of temporality or a container for information than a means of retaining context under conditions where context is causally necessary.

To avoid building this conclusion into the premises, narrative can be given a deliberately thin working definition at the outset: a multisymbolic presentation of temporally ordered events in which at least some earlier events are presented as consequential for what follows. The definition is minimal on purpose. A geological reconstruction satisfies it, and so do a medical case history and a sequence of weather reports. The task of the argument is to show why forms satisfying so thin a definition intensify, overwhelmingly, around living stakes, and above all around the stakes that recursive beings generate for one another.

The Odyssey will serve as a running case. "Poseidon is angry with Odysseus" is an accurate proposition and a radically inadequate explanation of anything that happens in the poem. To understand the anger, one has to narrate: the cave, the eaten companions, the wine, the false name, the blinding, the escape, the boast, the prayer. The anger can of course be summarized, and a short summary is far from useless: "Poseidon is angry because Odysseus blinded his son" already explains a good deal. Yet each compression discards part of the interrecursive architecture that makes the anger what it is, and beyond a certain point the anger remains classifiable, as divine wrath or paternal revenge, while ceasing to be adequately intelligible.

The argument proceeds in steps. Section II derives narrative relevance from living stakes rather than from eventfulness. Section III distinguishes three kinds of past according to the type of recursive counterpart involved. Sections IV and V show why interrecursive pasts resist the removal of context, and why explanation of such pasts sometimes has to take the form of story. Section VI introduces the consequential past, and Section VII locates narrative between mixed pastness and mixed futurity. Section VIII shows how narrative extends being-with across absence, and Section IX reads the Homeric appeal to the Muse as an explicit staging of remembering as an interrecursive activity. Section X explains, through the concept of metabolic buffering, why stories need not matter to their audiences in order to be loved by them. Sections XI and XII draw out empirical implications and situate the argument among existing theories, and Section XIII concludes.

II. Start with Stakes, Not Stories

A theory of narrative that begins with the familiar features of stories, their plots, characters, temporal sequences, causal chains, points of view and representational conventions, begins with products. It can describe with great refinement what narratives contain and how they are built, and still be unable to say why anyone should have started building them. The alternative is to ask why the production of narrative became necessary for the beings who produce it. The starting point is a claim about culture in general. Human life depends metabolically on forms of coordination that reach beyond immediate embodied bonding. Shared touch, rhythm, laughter and ritual settle relations among the co-present, but they cannot by themselves settle relations with strangers, with the absent, or with the dead and the unborn, and human groups of any size cannot avoid depending on all three. Culture is the cumulative cultivation of embodiment, dwelling and being-with through multimateriality and multisymbolization, undertaken because coordination beyond co-presence is a recurring and unavoidable metabolic stake. The hypothesis of this article is that narrative is one particularly important solution to a more specific version of this problem.

The first question, then, is what makes anything worth retaining and recounting at all. The answer is stakes. For something to be worth telling, something has to have mattered to a living being. Events as such are indefinitely many. The world is continuously eventful at every scale, and nearly all of it goes unnarrated because nearly all of it bears on no one. What selects an event for narrative is that it altered, threatened, secured or opened the continuation of some living being's life.

This places the deepest sources of narrative relevance in the three zoetic mediations. Embodiment generates stakes of survival, hunger, pain, injury, illness, fatigue, pleasure, reproduction and ageing. Being-with generates stakes of attachment, dependence, trust, hostility, care, jealousy, rivalry, recognition, obligation and abandonment. Dwelling generates stakes wherever a living being's continuing viability depends upon somewhere: on ground that bears crops or refuses them, on a coast that shelters or exposes, on a territory held or lost, on the difference between the place one belongs and the places one is driven to, and on the possibility of return.

Dwelling becomes a stake because somebody dwells. A mountain has no stake in being mountainous, and the sea has no stake in the ships it swallows. Ithaca has narrative force in the Odyssey because Odysseus belongs there: because a wife, a son, an ageing father and a whole body of dependants are held in suspension on that island by his absence, and because twenty years of his life are organised around getting back to it. The island's rocks and harbours matter as the ground of a return. Take away the one who belongs there and the same island becomes an entry in a gazetteer.

Multimateriality and multisymbolization profoundly shape narrative worlds. Ships, weapons, looms, bows, gifts, scars and disguises carry much of the action of the great narratives, and names, oaths, genealogies, prophecies and songs carry much of the rest. Yet in narrative these poietic mediations ordinarily figure as the means by which more basic stakes are stabilized, transformed, communicated or contested. The bow of Odysseus matters because only its owner can string it, which turns a test of the body into a proof of identity and a proof of identity into a claim on a household. Penelope's weaving matters because each night's unravelling postpones a decision about remarriage, and therefore about who will take Odysseus's place. The object is where the stake becomes graspable, and the stake lies in the living beings for whom the object makes a difference.

A first thesis follows: narrative relevance ultimately derives from living stakes rather than from eventfulness as such. A landslide can be physically spectacular and narratively unimportant. A quiet promise between two people, made without witnesses in a single sentence, can generate consequences lasting generations. The difference between the two has nothing to do with the magnitude of the physical event and everything to do with whether something became consequential for beings who could be affected by it and could respond.

III. Three Kinds of Past: The Ontology of Recursive Counterparts

The distinction on which the rest of the argument depends is the distinction among three types of recursive counterparts. Every living being coordinates with what surrounds it and with itself, and it does so recursively: its present discernment alters the conditions under which it will discern next. What varies is the kind of counterpart with which the coordination takes place, and each kind produces a different kind of past.

Nonrecursive pasts. A nonrecursive counterpart does not itself take what happens as a stake, and it does not reorganize itself in response to an interpretation of the being that engages it. Mountains, rivers, weather systems and rocks are counterparts of this kind. They can be engaged with great recursive sophistication by the living beings who depend on them, and they can have extremely complex histories. Their current state contains the consequences of their past, sometimes in legible layers, sometimes in forms that only patient scientific reconstruction can recover. A river's course records floods, a coastline records sea levels, a rock face records pressures and temperatures acting across immense spans of time. None of this is trivial, and much of it is very hard to know.

The crucial point is that causal complexity should not be confused with recursive complexity. A storm can be chaotic, nonlinear and highly sensitive to initial conditions, and therefore very difficult to predict, without being interrecursive in the slightest. The storm does not anticipate the sailor's response to it, does not take the sailor's earlier defiance as an affront, and does not carry the memory of last season's encounter into this one. The distinction is essential, because without it the argument would collapse into the claim that narrative arises wherever things are complicated, and that claim is false. Some of the most complicated processes known have attracted no narrative whatever, while some of the simplest human exchanges have attracted a great deal.

Selfrecursive pasts. Living beings can become recursively consequential to themselves. Something that happened yesterday changes what I can do today, and it does so not only as a cause acting on me but through my own continuing engagement with it. I remember an injury and move differently; I reconsider a decision and act differently next time; I feel shame and avoid a place; I acquire a skill and the world offers itself differently to my hands; I anticipate the recurrence of pain and arrange my days around it. The past here is relevant to the being whose past it was, and it is relevant because that being carries it forward and takes it up. This already creates a type of past unavailable to mountains: a past that matters to its bearer. It gives narrative one of its major domains, the transformation of a self across time, which runs from initiation stories to the Bildungsroman and from conversion narratives to accounts of illness and recovery.

Interrecursive pasts. The decisive escalation comes when the counterpart is itself recursive. Another recursive being is more than another source of causal pressure, as a storm or a slope would be, because the other can respond to me as a being responding to them. A does something to B. B registers what A did, discerns what it means for B, and responds. A encounters B's response and changes accordingly, perhaps by taking B's response as evidence of what B took A's original action to mean. B may in turn register and respond to that change. Each carries a different trace of the encounter, and each trace alters how the next encounter is met. Previous interaction becomes part of the conditions for subsequent interaction, so that what either party does later responds not only to the present situation but to the accumulated sequence of earlier responses.

A definition can now be given. Interrecursive complexity is the complexity generated when recursively organised beings become consequential conditions for one another's subsequent recursive activity.

Two qualifications keep this definition honest. First, interrecursivity does not equal complexity. Many interrecursive relations are settled, predictable and cheap: the long marriage in which each knows what the other will say, the greeting exchanged between neighbours for thirty years, the handshake that closes a routine sale. Their settledness is itself an achievement of accumulated interrecursive history, but it is settledness nonetheless, and at their ease such relations produce little that anyone would think to narrate. Second, some nonrecursive systems are vastly more complex, by any measure one might choose, than most human relationships. What distinguishes interrecursivity is less a quantity of complexity than a distinctive capacity: the capacity to generate an expanding history of mutually consequential responses, in which every new response enlarges the set of earlier responses that later ones may be responding to. That capacity is what narrative takes hold of.

IV. Why "Context" Becomes Causally Necessary

The distinction between the three kinds of past can be tested against a sequence that has been retold for nearly three thousand years. In the first book of the Odyssey, the gods assemble in Poseidon's absence, and Zeus explains why Odysseus is still kept from home: Poseidon is angry with him. That is the proposition. Ask what it explains and why it holds, and the story begins to unfold backwards.

Poseidon is angry because Odysseus blinded his son, the Cyclops Polyphemus. Why did Odysseus blind him? Because Polyphemus had shut Odysseus and twelve of his companions in his cave, blocked the entrance with a stone that no human crew could shift, and begun eating them two at a time. Why did Odysseus not simply kill him as he slept? Because, as Odysseus himself reasons, the survivors would then have been sealed in the cave for ever; the monster had to be disabled rather than killed, so that he would roll the stone away himself. How did they escape? Through a sequence of recursive manoeuvres: the strong wine offered as a guest-gift and accepted; the false name "Nobody" given in answer to the Cyclops's question; the olive stake hardened in the fire; the blinding; the cry to the neighbouring Cyclopes, who ask who is harming him, are told that Nobody is, and go home; the men bound beneath the rams; and Polyphemus feeling the backs of his flock as they pass out to pasture. How does Polyphemus learn whom to curse? Because Odysseus, safely aboard ship and against the pleading of his crew, shouts back his real name, his father's name and his home. Polyphemus then recalls an old prophecy that he would lose his sight at the hands of one Odysseus, which he had expected to be fulfilled by someone large and mighty rather than small and cunning. And why should the curse of a blinded shepherd on a remote island follow Odysseus across the whole sea? Because Polyphemus prays to his father, Poseidon, asking that Odysseus never reach home, or, if he is fated to, that he arrive late and in misery, having lost all his companions, on another's ship, to find troubles in his house. Nearly everything the poem goes on to narrate fulfils that prayer.

The point of this reconstruction is methodological rather than interpretive: it shows that the successive removal of "context" progressively destroys causal intelligibility. Remove the prayer, and there is no link between the island and the sea. Remove the boast, and Polyphemus has no name to pray against, which is exactly what the "Nobody" device had been designed to secure and exactly what Odysseus's need to be known as the one who did it undoes. Remove the cave, the eaten men and the stone, and the blinding becomes gratuitous cruelty, which would give Poseidon's anger a quite different character and Odysseus a quite different claim on the audience's sympathy. At each stage, what is removed is something one participant did in response to what another had done, and at each stage the explanation of what follows loses its hold. The short explanation that Poseidon is angry because Odysseus blinded his son survives some of this removal, and that is instructive rather than embarrassing, since the explanation already compresses an interrecursive architecture of its own: it names an act, attributes it to an agent, and invokes the kinship relation through which the act became an injury to someone else. What it loses is how the act came to be done, how its author came to be known, and why the injury was taken as it was. Compressed further, to "Poseidon was angry", the explanation deletes what the anger is anger at, and therefore what could ever appease it. Every compression discards some of the interrecursive architecture, and beyond a certain point the action remains classifiable while ceasing to be adequately intelligible.

A term is needed for the operation being resisted here. Contextual stripping is the removal of particulars in order to produce a portable description: a description that can be carried from the case in which it was formed to other cases without the case having to be carried along with it. In many domains contextual stripping works extraordinarily well, and much of what counts as scientific progress consists in learning how to do it. "The harbour wall collapsed because wave forces exceeded its structural resistance" is useful precisely because it discards almost everything that happened meteorologically in the hours and days before the collapse: the particular sequence of gusts, the shape of each wave, the order in which the stones worked loose. Contextual stripping has limits even here, and the limits matter. Nonrecursive processes can be irreducibly path-dependent. A fuller engineering account of the same collapse might need the particular sequence of storms, since fatigue failure depends on the order and history of loading and not merely on its total, and hysteresis, erosion, ecological succession, weather and geology all require, in their different ways, reconstruction of the particular path by which a present state arose. Resistance to historical compression cannot, therefore, by itself distinguish interrecursive from nonrecursive pasts. What persists in the wall persists as physical state: cracks, deformations, displaced stones. The wall does not take any wave as something, attribute it to anyone, or respond to what it takes the wave to have meant.

An interrecursive event is path-dependent in a different way, because earlier events persist as discerned actions to which the participants are presently responding. Polyphemus's prayer responds to a boast, which responds to a deception, which responds to a cannibal imprisonment, which itself responds to the arrival of uninvited strangers who have helped themselves to his cheeses and who, when he asks who they are, appeal to Zeus and the protection of guests, an appeal the Cyclops scorns. Nothing in this chain can be replaced by a sum of forces, however carefully sequenced. Each link carries forward, in the form of the next participant's discernment of what had been done to them, the particular significance of the link before. Hence a second thesis: interrecursive complexity resists contextual stripping.

The thesis needs its qualification stated at once. What resists stripping is history of a particular kind, since nonrecursive processes carry history too: it is interpretive, or stake-sensitive, path dependence. The relevant past includes what earlier events meant to recursive participants, how they attributed them, and what they did because of that discernment. Where a wall merely bears the effects of waves, Polyphemus takes what was done to him as an injury by a named enemy, appeals to a father who can act on it, and Odysseus's later sufferings respond to that taking and that appeal. This layer of the past, the past as discerned and answered, is what portable descriptions characteristically discard, and it is the irreducibly interrecursive feature of the case.

This requires a more precise understanding of context than the ordinary one. Context is often taken to mean background: the setting against which an event occurs, relevant to its interpretation but separable from its causes. In interrecursive situations this picture fails, because the relevant background is frequently composed of earlier interactions, and earlier interactions are exactly what present action is responding to. Context can therefore be defined as the selectively retained past required to make a present event intelligible. On this definition, context in interrecursive cases belongs to the causal architecture itself rather than supplementing it, and whoever strips it away replaces the explanation with another, usually worse, one.

A brief contrast with classification clarifies what is at stake. A portable category, of the kind that becomes institutionally stabilized at L4, seeks what remains invariant once particulars have been removed. "Blood feud", "revenge", "divine wrath" and "violation of hospitality" are all categories under which the Polyphemus episode could be filed, and each would allow it to be compared with thousands of other cases. Such categories are indispensable to any comparative enterprise. Yet the particulars they remove are frequently the ones whose removal would make subsequent action unintelligible, and narrative often protects exactly those particulars. The tension between the portability of the category and the intelligibility of the case runs through the whole history of comparative scholarship, and it is one reason why ethnography, which must preserve cases, has never settled comfortably into a science of categories.

V. Why Explanation Sometimes Has to Become Story

The Mahabharata offers a second and far larger case. Asked to explain the war at Kurukshetra in a sentence, one might say that it resulted from a dynastic conflict over rule of the Kuru kingdom between two sets of cousins, the Pandavas and the Kauravas. That is true, and almost useless as an explanation of anything anyone does in the war. It does not explain why particular warriors fight as they do, why some of them fight on the side they do, why certain duels matter more than the outcome of the battle, or why the killings that end the war take the specific and often shocking forms they take.

To explain those, one has to go back. There is the dice game, in which Yudhishthira, the eldest Pandava, loses to Shakuni, who plays on Duryodhana's behalf, first his wealth, then his kingdom, then his brothers, then himself, and finally Draupadi, the wife of all five brothers. There is Draupadi's humiliation in the assembly hall: dragged in by Duhshasana, taunted as a slave, subjected to the attempt to strip her before the assembled elders, and posing the question none of them can answer, whether a man who had already lost himself had any right to stake her. There are the vows the humiliation produces. Bhima swears that he will break the thigh Duryodhana bared to her and that he will drink Duhshasana's blood, and in many tellings Draupadi vows to leave her hair unbound until it can be washed in that blood. Behind the dice game lie earlier injuries: Duryodhana's humiliation in the Pandavas' marvellous hall at Indraprastha, where he takes a floor for a pool and a pool for a floor and is laughed at, and the long rivalry of cousins raised together. Behind those lie questions of inheritance reaching back a generation and more, to Bhishma's vow of celibacy, which removed the rightful heir from the succession, and to the blindness that kept Dhritarashtra from the throne. Around them are threaded alliances, debts and promises: Karna's loyalty to Duryodhana, who raised him from a charioteer's son to a king; Karna's hidden kinship with the very brothers he fights, and his promise to their mother that he will kill none of them but Arjuna; the Pandavas' thirteen years of exile, the last of them in disguise, completed on terms that are then disputed. Through all of this runs a series of divergent understandings of obligation, dharma in its many and conflicting senses, which the epic does not resolve but stages again and again as a problem that each participant must confront in the particular form that his or her own history has given it.

The fulfilment of Bhima's vow shows what summary loses. In the final mace combat, Bhima strikes Duryodhana on the thigh, below the navel and against the rules of the fight, and Balarama, who taught both men, is enraged by the foul blow. To say that Bhima defeated Duryodhana misses everything that makes the episode matter: that the illegitimate blow fulfils a vow made thirteen years earlier in response to a gesture of sexual contempt, that its fulfilment and its illegitimacy are one and the same act, and that the victory of the righteous side is thereby compromised at the very moment of its completion. The blow is intelligible as a response to the dice hall, and only as that.

The extraordinary length of the Mahabharata, traditionally reckoned at a hundred thousand verses, is therefore theoretically interesting. Its compositional history is a matter for textual scholarship, and accretion over centuries explains much. Whatever that history, the text's refusal of compression answers to something in its subject. An interrecursive history of this density cannot be radically shortened without progressively eliminating the relationships that make the final conflict intelligible. Every summary of the epic meets the problem: summaries either swell back into genealogy or shrink into moral types, and in shrinking they lose the very particulars that made each participant's action a response to someone else's.

The central distinction can now be stated. A causal summary tells us that something produced something else. Narrative can preserve how stakes became consequential across successive recursive encounters: how one being's response became the condition for another's, how that response was taken, and how the taking entered what followed. Hence a third thesis: narrative is a form of contextual retention. Or, more strongly: narrative is a symbolic technique for preserving recursively relevant context through time.

This also clarifies why temporal sequence alone is insufficient for narrativity in the strong sense. A geological account has sequence, often measured with great precision, and a weather simulation has causal succession step by step. Both satisfy the thin definition given at the outset, since earlier events are presented as consequential for later ones, and both are narratives in that minimal sense. Yet neither draws on the resources that make narrative indispensable, and when either is told as a story in the fuller sense, the telling almost always introduces a stake: the settlers who farmed the slope, the fishing fleet caught by the storm. Aristotle had already distinguished, in the Poetics, between events that merely follow one another and events that happen because of one another, and ranked plots by the second. The present argument adds a specification of which "because" matters most. Narrativity intensifies when sequences are organised around what mattered to recursively organised beings, and it intensifies further when those beings were responding to one another.

VI. The Consequential Past

What happens after an event ends? The physical event is gone. The insult has stopped being spoken; the battle has ended; the promise was uttered years ago and the breath that carried it long since dispersed. Yet each of these can go on organising subsequent life, sometimes more powerfully than anything happening in the present. A consequential past is a past event whose effects remain recursively relevant in the present: relevant, that is, to living beings who continue to take those effects up, respond to them and reorganize themselves in light of them.

The event need not remain continuously represented by anyone. Its consequences may persist through bodies, relations, places, objects and symbols: through a scar or a limp, through an obligation that shapes whom one may marry, through a field left unfarmed or a grave that is visited, through a weapon handed down, through a name that commemorates an ancestor or a curse that names an enemy. How remembering is distributed across all five mediations is a larger subject than can be pursued here. The present argument needs only the principle: narrative depends upon the possibility that something absent can remain consequential.

The examples themselves force a further distinction. Odysseus, Poseidon, Polyphemus and Draupadi belong to narrated worlds, whatever historical memory may lie behind the epics, and the argument has no need to establish that any of them existed or that anything told of them occurred. Narrative can retain an actual consequential past, and it can construct a possible one. Fiction exploits the same architecture that testimony, chronicle and family history use. Within the narrated world, the blinding of Polyphemus remains consequential for Poseidon whether or not there ever was a Polyphemus, and the audience follows it as consequential on exactly those terms. What narrative preserves for its audience is therefore not necessarily an actual historical event but a structured sequence in which stakes and responses remain intelligible across time. Nor does fiction thereby become a later distortion of an originally factual mnemonic technique, since retaining actual pasts and constructing possible ones draw on a single capacity, best described as making absent interrecursive experience, actual or imagined, available to those who did not undergo it.

Draupadi's humiliation is the clearest instance. It does not simply "remain in memory" as a record occasionally consulted. It remains consequential through transformed relations, since the cousins can no longer meet as cousins; through articulated vows, which turn a moment of powerlessness into a binding claim on the future; through obligations that fall on her husbands and their allies; through the bodily sign of her unbound hair, in the tellings that give her that vow, visible to anyone who sees her in the years that follow; and finally through bodily action on the battlefield, where Bhima's hands and mouth carry out what his words committed them to. The humiliation leaves the assembly hall in all these forms at once. Polyphemus's injury likewise leaves the island, and it does so through attribution, since he learns whom to blame; through kinship, since the blinded son has a divine father; and through divine response, since that father has power over the whole sea.

This points to what may be called recursive non-closure. Stories continue because past events remain unfinished, in the sense that their consequences have not ceased to propagate through the recursive activity of those they affect. A promise is not over when it has been uttered, since its whole point is to remain in force. An insult is not necessarily over when the words end, because the one insulted may carry it for decades. A disappearance remains open for as long as someone waits. A killing can remain consequential across generations, which is what the institution of the feud, from the Icelandic sagas to the customary law of highland Albania, both acknowledges and attempts to regulate. The Odyssey itself does not close its own consequential past. In the underworld, Tiresias tells Odysseus that even after he has killed the suitors he must set out again, carrying an oar inland until he meets people who know nothing of the sea and take it for a winnowing fan, and there make sacrifice to Poseidon. The poem ends before that journey begins. Poseidon's anger, the engine of the whole narrative, is left not resolved but scheduled for resolution beyond the text.

A fourth thesis can therefore be stated: narrative is organised around events that outlive their occurrence. Within a narrated world, an event ceases to function as an active element of the plot once nothing later responds to it: where nothing remains to be taken up, answered, honoured or avenged, it drops out of the story's recursive economy and survives, if at all, as circumstance. That does not prevent the whole narrated world from being reopened for an audience thousands of years later, as every new reading of the Odyssey shows. Consequence within a narrated world and the recursive opening of that world for later audiences are distinct matters, and Section X returns to the difference.

VII. Mixed Pastness and Mixed Futurity

Different recursive counterparts generate differently knowable futures. Nonrecursive processes can sometimes be predicted with great reliability, as with tides and eclipses, and sometimes hardly at all, as with earthquakes or the weather more than a few days ahead. Whatever its degree, their predictability does not depend on anyone's response to the prediction: the tide does not come in later because it has been forecast. Selfrecursive futures depend on the changing trajectory of the being whose future they are; my future skill depends on how I practise, and how I practise depends in part on what I anticipate. Interrecursive futures acquire a distinctive openness, because another recursive centre can respond to my own response, including my response to what I expect it to do. The future of a relationship is jointly generated. Neither party holds it, and neither can fully compute it, because each is part of the other's conditions. Here too, interrecursivity must not be equated with unpredictability. Settled relations are often highly predictable, and much of culture consists in making them so. What differs is the kind of openness: an interrecursive future is open in a way that better observation alone cannot close, since the participants' anticipations and interpretations of one another can enter the very process they anticipate. This is mixed futurity.

The same structure can be mirrored backwards. Different kinds of past are differently reconstructable, a condition that may be called mixed pastness. A nonrecursive past may be extremely difficult to reconstruct, as any palaeoclimatologist knows, but it contains no other participant's stake in what happened; reconstruction is a matter of recovering causes and conditions. A selfrecursive past contains a recursively consequential trajectory, and to reconstruct it is to recover not only what happened to a being but what that being made of what happened, and what that making did to what followed. An interrecursive past may require reconstructing multiple interacting stakes, interpretations and responses, and it has a further peculiarity. It is historically closed, in that what was done cannot be undone, and recursively open, in that what was done can still be taken up, reinterpreted, forgiven, avenged, disputed or honoured by those who carry it forward. Each participant retains a different version, and the differences are themselves among the facts to be reconstructed. The quarrel over whether Yudhishthira had the right to stake Draupadi is a quarrel over what happened in the dice hall. It is never settled in the hall; it is fought out on the battlefield instead.

This gives narrative a precise temporal location: narrative mediates between mixed pastness and mixed futurity. It takes a past whose components are differently reconstructable and a future whose components are differently open, and holds them together in a form that allows both to be apprehended at once.

An overly instrumental version of this claim should be resisted. It would be easy to say that narrative exists to tell us what to do next, as a guide to action compiled from the past. That is sometimes true, and it is far too narrow. What narrative does, more fundamentally, is make enough of a consequential past available for a temporally extended field of recursive possibilities to remain intelligible. The listener to the Odyssey is not told what to do. The listener learns what the situation is, what hangs on what, and what could happen from here. The intelligibility of possibility is the more basic achievement, and guidance for action is one of the things that can be built upon it.

The basic narrative tension follows directly. If the past had no consequences, events would not hang together; they would follow one another like items in a list, and nothing would be at stake in their order. If the past completely determined the future, the question "what happens next?" would cease to be open, and suspense would become mere delay. Narrativity flourishes between these extremes, where the past matters and the future is not exhausted by the past. Prophecy in the ancient epics confirms rather than contradicts this. The prophecies of Polyphemus's seer and of Tiresias say what will happen and leave open how, when and through whose actions, and it is the how, the when and the who that the audience waits for. Even the fate of Oedipus, the paradigm of foreclosure, is told as a sequence of recursive responses to the prophecy itself, each designed to escape it and each bringing it closer.

Viable recursive life has the same structure. It requires that the past be relevant without totally foreclosing the future: that what has been learned, promised and suffered carry forward, while enough remains undecided for response still to be possible. A being for whom nothing carried forward could coordinate with nothing; a being for whom everything was already fixed could no longer respond. Recursive-mediational fluidity, the capacity to be shaped by the past while remaining able to meet what comes, is a condition of living, and narrative gives that condition symbolic form.

VIII. Narrative as Extension of Being-With Across Absence

The argument can now return to the evolutionary problem with which it began. Embodied bonding has a fundamental limit: shared touch, rhythm, laughter and ritual require co-presence. They settle relations among those who are there, and they cannot by themselves coordinate relations with strangers, with absent persons, or with the dead and the unborn. Narrative radically extends this range.

Consider the simplest case. Someone undergoes an event. They tell someone who did not undergo it. That second person can now orient toward a world whose relevant past exceeds their own experience: they know whom to trust in the neighbouring camp, which way the herd has moved, what passed between two people who are about to arrive. A generation later, someone else can orient toward events that occurred before they were born, among people they never met, in places they have never been. At each step, experience is detached from the circumstances of its occurrence without being detached from its consequence. Hence: narrative makes the metabolic usefulness of experience detachable from the lifetime and location of the experiencer.

This must be phrased carefully, because it is easily mistaken for a claim about information transmission, and it is a stronger claim than that. A warning can be passed on without narrative: there is danger at the river. What narrative can carry, and a warning cannot, is the structure of relations among events, stakes and responses: this person trusted that person, who did this, whereupon that followed, and the one who was betrayed has not forgotten. Rather than receiving a fact about the world, the listener is placed, however partially, inside an interrecursive history, and is thereby equipped to coordinate with its surviving participants as someone who knows what they have been through. Ethnographic work on storytelling among foragers is consistent with this. Polly Wiessner's recordings of Ju/'hoansi conversation found daytime talk dominated by economic matters, complaint and criticism, and night-time talk by the fire dominated by stories, many of them concerning people not present; she argued that such firelight talk extended the networks within which people could coordinate beyond those they met directly (Wiessner 2014).

The dead are the limiting case. A dead person cannot participate in embodied being-with. There is no touch, rhythm or laughter to share with them, and no response they can give to anything the living do. Yet their actions, obligations, promises, warnings and relationships can remain consequential through narration. The ancestor who cleared the land, the grandfather who quarrelled with a neighbouring lineage, the saint who founded the shrine, the traitor whose name may not be spoken: each continues to condition the recursive activity of the living, and the living coordinate with them in the sense that they take them into account, honour or repudiate them, and fulfil or refuse what they began. Likewise the unborn enter worlds already organised by narratives of events they never witnessed. A child is born into feuds, alliances, debts and loyalties that are explained to her, when they are explained at all, by being told.

Hence a strong formulation: narrative converts temporally impossible being-with into mediated being-with. The formulation needs precision if it is not to blur the ontology of recursive counterparts at exactly the point where that ontology matters most. Narration does not make the dead currently interrecursive, since they no longer respond and nothing the living do can become a condition of their further activity; what it achieves is that present being-with can remain organised by absent or dead recursive others whose earlier actions remain consequential. Their past recursive activity stays available, and the living can make their own actions responses to it, fulfilling, refusing, honouring or avenging what was begun. "Mediated being-with" is shorthand for this asymmetric relation: the dead cannot answer, but they can still be answered.

This is where the argument about narrative grows most directly out of the argument about culture. Culture is necessary because human coordination exceeds embodied co-presence, and its multimaterial and multisymbolic means extend coordination to those whom the body's own bonding cannot reach. Narrative is among the most powerful of these means, and it is distinctive in what it extends. Money extends being-with to strangers by making personal trust unnecessary for a certain class of exchange. Law extends it to the absent by holding them to terms agreed elsewhere. Narrative extends it across time in a form that keeps the relational content of the past available, so that the absent and the dead are carried forward as participants in histories rather than as parties to transactions. Its core operation is multisymbolic, and it has always leaned on multimaterial anchors: the singer's lyre, the carved genealogical post, the shrine at the place where the story happened, and the scar above Odysseus's knee by which his old nurse knows him, and which calls forth, in the middle of the recognition scene, the whole story of the boar hunt in which he received it.

IX. The Muse: Remembering as an Interrecursive Activity

The Homeric poems open with an appeal. The Iliad begins by bidding a goddess sing the wrath of Achilles; the Odyssey by asking the Muse to tell of the man of many turns. Later, before the long catalogue of the ships that sailed to Troy, the Iliad makes the appeal explicit and gives its reason. The Muses are goddesses, the singer says; they were present and know all things, whereas we mortals hear only report, kleos, and know nothing. Even with ten tongues and ten mouths, an unbreakable voice and a heart of bronze, he could not name the multitude unless the Muses recalled it for him.

It would be a mistake to literalize this into a thesis about Homeric psychology, as though the singers experienced recollection as the arrival of a goddess. The invocation is better read as a formal enactment of what the present argument predicts. Instead of announcing the retrieval of a private archive, the singer stages the act of remembering relationally, appealing to another being, one who was present when he was not, to make an absent world available. The ground of the appeal is precisely presence. The Muses can tell because they were there; the singer needs them because he was not. Remembering, in this staging, is a form of being-with across absence, carried out by asking someone else.

Remembering can plainly be selfrecursive, as when one recalls one's own past alone, and it can reactivate a nonrecursive past, as when one recalls a flood. What is extraordinary about the invocation is that epic remembering is explicitly staged as interrecursive. "I remember" becomes, in effect, "help me make this past available." The same structure appears in ordinary remembering whenever more than one person takes part. "Do you remember when...?" One person begins. Another corrects a detail. A third adds something neither had mentioned, or disputes the order in which things happened, or recalls what someone said. The remembered event becomes available through a relational reconstruction in which each contributor responds to the others' contributions, and the version that emerges may belong to none of them singly. Research on conversational remembering has documented how such joint recall reshapes what each participant subsequently remembers, sometimes bringing memories into alignment and sometimes silencing what went unsaid (Hirst and Echterhoff 2012), and Frederic Bartlett's much older experiments on serial reproduction had already shown how stories change shape as they pass from teller to teller (Bartlett 1932). Work on distributed remembering has generalized the point, describing joint remembering as the coordination of cognitive, bodily, social and material resources across several timescales at once (Bietti and Sutton 2015), and treating remembering more broadly as a process spread across bodily and worldly resources as well as the brain (Sutton 2026).

This offers a compact bridge into oral tradition. Narrative is not only what gets remembered; narrative can be an activity of remembering together. The technical means by which oral traditions stabilize such joint remembering, meter, formula, type-scene and the embodied rhythm of performance, and the later redistribution of memory into writing, raise questions well beyond what the present argument requires (Lord 1960). What the argument requires is the observation that the oldest surviving epic tradition of Europe places an interrecursive appeal at the threshold of narration itself. Before the first event of the story is told, the teller turns to someone else and asks for help in making the past present.

X. Metabolic Buffering: Why Stories Need Not Matter to Us

An objection now presses. If narrative developed because consequential interrecursive pasts matter for coordination, why do human beings spend so much time enjoying stories that make no obvious difference to their survival, their relationships or their future action? The reader of the Odyssey will never be shut in a cave by a Cyclops, will never need to string a bow to reclaim a household, and has no relationship with Poseidon to manage. The most widely enjoyed narratives of any period are often those most remote from the practical circumstances of their audiences. An account that derived narrative from coordination and stopped there would have to treat most actual narrative enjoyment as a malfunction.

The answer lies in distinguishing two sets of stakes that are easily run together: the stakes within the narrated world, and the stakes generated for the audience by the act of narration. These can differ radically. The Odyssey is full of shipwreck, drowning, cannibalism, sexual danger, enchantment, bereavement, homecoming, slaughter and divine persecution. Its audience experiences none of these as immediate metabolic problems. The audience is fed, warm, among others and in no danger from anything in the poem. Narrative makes high-stakes worlds recursively available while buffering the audience from having to remediate those stakes directly. This condition may be called metabolic buffering.

Buffering must be defined with some care, as reduced direct recursive exposure rather than reduced bodily activation. A terrifying story can quicken the pulse, bring tears, provoke disgust and cost a night's sleep, and much of the power of narrative lies in exactly such effects. What is reduced is exposure: the listener does not have to remediate Odysseus's situation, and Poseidon cannot respond to the listener's response. The narrated stakes can be felt without being incurred, and the felt response, however intense, opens no interrecursive relation with anyone in the story.

Metabolic buffering rests on a distinction between narrative availability and recursive opening. A narrative makes a narrated world available: its participants, their stakes, their histories and their possible futures can be followed, felt and anticipated. It does not necessarily open anything recursively in the audience's own present world. Nothing in the listener's own coordination with body, place and others is thrown into question by following Odysseus through the cave. One can follow him intensely, with a quickened pulse and real dread, and then make a cup of tea.

Buffering comes in degrees. At one extreme lies strongly buffered narrative engagement: pleasure, curiosity, fascination, absorption, the whole experience of being lost in a story. In the middle, stories may alter expectations, sensitivities or attachments without producing any urgent action; a novel read in youth may change what one looks for in other people without anyone, including the reader, being able to say how. At the other extreme, narratives become themselves metabolically consequential in the present. Testimony in court, an accusation, a family story about who wronged whom, a religious narrative that obliges, a political narrative that mobilizes, a nationalist narrative of origin, a narrative of historical grievance: in each of these the telling recursively opens the audience's own world, and the listener may leave with obligations, enemies, allegiances or grounds for action that were not there before.

The Aeneid shows the transition. Its world was as remote from its first audience as Homer's was from his, and much of it can be followed under full buffering. Yet the poem ties narrated antiquity to claims about the collective identity and political order of its audience's present: Aeneas's son Iulus stands at the head of the Julian line, and the whole poem leans toward the Rome of Augustus. When Dido, abandoned, curses Aeneas and his descendants, praying for unending enmity between her people and his and for an avenger to rise from her bones, the curse reaches out of the narrated world into historical memory, since every Roman listener knew what Carthage and Hannibal had meant. A narrated interrecursive past is here reactivated as the origin of a present political relation, and the buffering thins accordingly.

A negative argument is needed at this point. Buffered narrative should not be reduced to simulation training. Audiences sometimes learn from stories, and stories sometimes rehearse social possibilities in ways that later prove useful; there is good argument and some evidence for such effects (Mar and Oatley 2008). But narrative does not have to justify each instance of itself through an immediately adaptive lesson. Once human beings possess the capacity to engage with recursively consequential worlds under buffered conditions, that engagement can become pleasurable in its own right, much as capacities for movement formed under metabolic pressure can become the pleasure of dancing. The pleasure of a story no more needs to be explained as disguised training than dancing needs to be explained as disguised foraging.

One formulation captures what buffering makes possible: narrative allows metabolically costly forms of interrecursive complexity to become available at metabolically reduced cost. Following an interrecursive history among the living, when one is a participant in it, is among the most demanding things a human being ever has to do. It is exhausting, frightening and often unresolvable. Following one in a story, where every participant's stake can be felt and none of it can come back to oneself, offers comparable density of recursive consequence without the exposure. The reduction lies in exposure rather than in intensity. Fiction carries this to its limit, since it allows whole interrecursive histories to be generated that nobody had to undergo in the first place, so that the density of consequence available to an audience is no longer bounded by what anyone actually suffered. The asymmetry between felt intensity and direct exposure may be one important source of narrative pleasure.

XI. Why Narrative Is So Often About Other People

If the argument is right, narrative should show a strong historical attraction toward embodiment and above all toward being-with. This is what one finds. The great themes of narrative traditions across the world, love, rivalry, betrayal, kinship, loyalty, revenge, death, separation and return, did not become central because literary traditions happened, arbitrarily and independently, to become obsessed with them. They are the domains in which stakes remain recursively consequential over time, where what one person did becomes the condition of what another does, and where the past most resists stripping.

This does not mean that all stories are interrecursive. There are powerful selfrecursive narratives, and there are narratives unusually organised around nonrecursive counterparts. These should, however, be analytically conspicuous, noticeable as departures, precisely because interrecursive being-with constitutes so powerful a baseline attractor. Three limiting cases indicate the range.

The Odyssey is interrecursive throughout, and its interrecursivity crosses kinds of being. Odysseus's history is made of responses to and from humans, monsters, nymphs and gods, each of whom engages him as a counterpart who responds, and it is exactly because a Cyclops and a sea-god can be affronted that their stories become his. Even the sea, in this poem, is largely the domain of a being who remembers.

Madame Bovary runs on an unusually powerful selfrecursive engine. Emma Bovary's story is driven by the gap between what she has taken from novels, sermons and keepsake albums and the life she finds herself living, and by her repeated attempts to close it. Yet even here the selfrecursive trajectory is continually intertwined with other people, with Charles, Rodolphe, Léon and the draper Lheureux whose credit quietly ruins her, and with symbolic imaginaries whose sources are themselves other people's narratives. Part of the novel's force comes from how thin its interrecursive dimension is allowed to become: her lovers respond to her far less than she imagines, and much of the tragedy lies in the discovery of that thinness.

The Martian, Andy Weir's novel of an astronaut stranded on Mars, runs on an unusually strong nonrecursive and multimaterial engine. Its protagonist's problems are atmosphere, soil, water chemistry, power supply and distance, and his solutions are material and calculative. The novel's pleasures are those of problem-solving within a nonresponsive environment, and part of its interest lies in how far a story can travel on such an engine. It is conspicuous, too, that the narrative keeps reintroducing being-with: through the stranded man's logs, addressed to readers he imagines; through the controllers and engineers on Earth; and through his crewmates' decision to turn back for him. A purely nonrecursive narrative seems hard to sustain for long.

These are sketches, and a full comparative analysis of narrative architectures would need a systematic method of its own. The point here is simpler: the theory generates comparative predictions. It predicts that interrecursive narrative will dominate across traditions, that selfrecursive and nonrecursive narratives will be marked and relatively rarer, and that where they do occur they will tend to reintroduce being-with at their margins.

XII. What Existing Theories Have Seen, and What They Have Not Yet Derived

The argument developed here converges with existing work at many points, and the convergence is partial in an instructive way.

Evolutionary theories of storytelling have linked narrative to social learning, reputation, cooperation and the navigation of complex social situations. Brian Boyd treats fiction as an adaptation that sharpens social cognition and attention through play and fosters cooperation (Boyd 2009), and his later synthesis comes closer still to the present argument, holding that stories help humans track reputations, understand reasons through particular histories, and set local events within larger contexts (Boyd 2018). Michelle Scalise Sugiyama has argued very directly that narrative allows transmitted representations to substitute for costly first-hand experience, drawing on the subsistence lore of foraging peoples (Scalise Sugiyama 2001), a claim that comes close to one side of what is here called metabolic buffering. Fieldwork among Agta hunter-gatherers found that skilled storytellers were preferred as campmates and had greater reproductive success, and that the stories commonly carried messages about cooperation and social norms (Smith et al. 2017). Social-brain and gossip accounts connect language itself to the management of increasingly extended social relations. Robin Dunbar's argument that language took over the bonding work of grooming once groups grew too large for grooming to sustain is, among other things, a claim about coordination beyond the reach of embodied bonding (Dunbar 1996). Simulation accounts hold that fiction offers a model of social experience that can be entered and explored at low cost (Mar and Oatley 2008), and cognitive approaches to reading have stressed how intensively fiction exercises the attribution of mental states (Zunshine 2006).

Narrative psychology and philosophy have emphasized that understanding action often requires histories of reasons, beliefs, intentions and circumstances. Daniel Hutto's narrative practice hypothesis holds that making intentional action intelligible essentially involves narratives about particular persons and their reasons, and that the capacity to do so is acquired through engagement with such narratives (Hutto 2008). Alasdair MacIntyre argued that actions become intelligible only when placed within narrative histories (MacIntyre 1981), and Paul Ricoeur made the relation between narrative and lived time the subject of a three-volume inquiry (Ricoeur 1984–1988). Jerome Bruner's distinction between the paradigmatic and the narrative modes of thought is the most important precedent for the contrast drawn here between portable abstraction and narrative particularity (Bruner 1986). Social and distributed approaches to memory, from Maurice Halbwachs onward, have challenged the notion that remembering is simply individual storage and have shown how it is shaped by the groups in which it occurs and the conversations through which it is performed (Halbwachs 1992; Hirst and Echterhoff 2012).

None of this needs to be denied. What these approaches have in common is that each identifies a different function or property of narrative, whether social learning, reason explanation, context dependence, memory transmission, temporal configuration or the preference for social content, without deriving their convergence from a common problem. Each is right about what it sees, and together they amount to a list, the list of things narrative does. The present argument proposes that the list has a single source, and that the source is interrecursive complexity among beings with metabolic stakes.

The novelty of the present argument therefore lies not in the claim that narrative retains social context, transmits experience or makes action intelligible, each of which has been argued before and some of it very well, but in deriving all of these features from a single conjunction: metabolic stakes, the ontology of recursive counterparts, interrecursive complexity, and mixed pastness and futurity. It lies above all in the claim that what narrative preserves is the recursively relevant history by virtue of which one action is intelligible as a response to another. If interrecursive pasts resist contextual stripping in this specific sense, then any being that must coordinate across absence with others whose present action responds to earlier action will need a means of carrying context forward. Social learning follows, because what is learned is how others have responded to one another. Reason explanation follows, because reasons are responses to situations that include the earlier actions of others. Context dependence follows by definition. Memory transmission follows, because context must survive the deaths of those who lived it. Temporal extension follows, because consequences outlive events. And narrative's overwhelming interest in being-with follows, because being-with is where interrecursive complexity is densest. What the literature has recorded as a set of separate observations can be derived, on this account, from a single architecture.

The argument also sharpens one disagreement. Simulation and training accounts tend to explain narrative pleasure as the reward attached to useful practice. The account offered here allows that practice can occur while identifying buffering as one important and independent source of narrative pleasure, alongside whatever curiosity, practice and aesthetic form contribute. The difference is testable in principle. Training accounts predict that pleasure should track usefulness; the buffering account predicts that pleasure may sometimes track the density of recursively consequential engagement available under conditions of low direct exposure, whether or not anything useful is learned. That is a hypothesis on which the two accounts can be made to diverge.

XIII. Conclusion: Why Stories Continue

A mountain has a past. A storm has a past. A person has a past. A relationship has a past. Only living recursive beings have stakes. A single recursive being can make its own earlier responses part of what its later responses respond to, but only among several recursive beings does one centre's response become part of the conditions for another centre's subsequent response. That creates a peculiar explanatory problem. The past remains consequential, yet its consequences cannot always be understood once its relational context has been stripped away, because the context is part of what produced them.

Narrative is one solution to this problem. It preserves selectively rather than exhaustively: it need not reproduce everything that happened, and no narrative ever does. It keeps enough of what mattered available for subsequent events to remain intelligible, and its selectivity is the whole of its art, since what it keeps is fixed by what later events will turn out to be responding to.

At the evolutionary scale, narrative allows experience to exceed the event, so that what happened can go on mattering after it has stopped happening. It allows experience to exceed the experiencer, so that what one being underwent becomes available to others who did not undergo it. It allows experience to exceed co-presence, so that those who were not there can coordinate with those who were. And eventually it allows experience to exceed a lifetime, so that the dead, who can no longer answer, can still be answered, and the unborn enter worlds already organised by what was done before them. In fiction, it allows experience to exceed occurrence itself, so that interrecursive histories nobody underwent become available to anyone who listens. This makes narrative one of the great means by which being-with becomes temporally extensible, and one of the chief ways in which human life has met the recurring problem of coordination beyond the reach of the body.

What began as a way of making consequential absent experience available did not need to remain immediately instrumental. Under metabolic buffering, human beings can engage with narrated stakes, actual or imagined, without being directly exposed to them, and that engagement can become one of the principal pleasures of their lives. At the other end of the spectrum, the same capacity allows narrated pasts to be reactivated as present obligations, grievances and claims, which is why stories can found polities and start wars. Narrative opens worlds without necessarily opening our own, and when it is told in the right place to the right audience, it can open our own as thoroughly as anything that happens in it.

Narrative exists at the intersection of a consequential past and an incompletely determined future. It becomes especially necessary where recursively responsive beings have made themselves part of one another's histories. Their actions cannot always be understood by stripping away what came before, because what came before is what the present action is responding to. Narrative is the human capacity for carrying enough of that past forward that what happens between living beings remains intelligible, even when the event is over, the participants are absent, and the audience itself has nothing immediately at stake.

Selected Readings

Bartlett, Frederic C. 1932. Remembering: A Study in Experimental and Social Psychology. Cambridge: Cambridge University Press.

Bietti, Lucas M., and John Sutton. 2015. "Interacting to Remember at Multiple Timescales: Coordination, Collaboration, Cooperation and Culture in Joint Remembering." Interaction Studies 16 (3): 419–450.

Boyd, Brian. 2009. On the Origin of Stories: Evolution, Cognition, and Fiction. Cambridge, MA: Belknap Press of Harvard University Press.

Boyd, Brian. 2018. "The Evolution of Stories: From Mimesis to Language, from Fact to Fiction." WIREs Cognitive Science 9 (1): e1444.

Bruner, Jerome. 1986. Actual Minds, Possible Worlds. Cambridge, MA: Harvard University Press.

Dunbar, Robin. 1996. Grooming, Gossip and the Evolution of Language. London: Faber and Faber.

Halbwachs, Maurice. 1992. On Collective Memory. Edited and translated by Lewis A. Coser. Chicago: University of Chicago Press.

Hirst, William, and Gerald Echterhoff. 2012. "Remembering in Conversations: The Social Sharing and Reshaping of Memories." Annual Review of Psychology 63: 55–79.

Hutto, Daniel D. 2008. Folk Psychological Narratives: The Sociocultural Basis of Understanding Reasons. Cambridge, MA: MIT Press.

Lord, Albert B. 1960. The Singer of Tales. Cambridge, MA: Harvard University Press.

MacIntyre, Alasdair. 1981. After Virtue: A Study in Moral Theory. London: Duckworth.

Mar, Raymond A., and Keith Oatley. 2008. "The Function of Fiction Is the Abstraction and Simulation of Social Experience." Perspectives on Psychological Science 3 (3): 173–192.

Ricoeur, Paul. 1984–1988. Time and Narrative. 3 vols. Translated by Kathleen McLaughlin and David Pellauer. Chicago: University of Chicago Press.

Scalise Sugiyama, Michelle. 2001. "Food, Foragers, and Folklore: The Role of Narrative in Human Subsistence." Evolution and Human Behavior 22 (4): 221–240.

Smith, Daniel, Philip Schlaepfer, Katie Major, Mark Dyble, Abigail E. Page, James Thompson, Nikhil Chaudhary, et al. 2017. "Cooperation and the Evolution of Hunter-Gatherer Storytelling." Nature Communications 8: 1853.

Sutton, John. 2026. "Distributed Remembering and Cognitive Philosophy." In Cognition, Culture, and Political Momentum: Breaking Down the Silos in Collective Memory Research, edited by Astrid Erll and William Hirst, 27–43. Oxford: Oxford University Press.

Wiessner, Polly W. 2014. "Embers of Society: Firelight Talk among the Ju/'hoansi Bushmen." Proceedings of the National Academy of Sciences 111 (39): 14027– 14035.

Zunshine, Lisa. 2006. Why We Read Fiction: Theory of Mind and the Novel. Columbus: Ohio State University Press.