3. Why Causality Carries the Weight

The causal chain is usually asserted and left there. It can be argued, and from several independent directions — classical poetics, psychoanalytic narratology, structuralist analysis, discourse psychology and institutional history. This section assembles the case, and then, in §3.7, takes most of it back — or rather, states the precise sense in which the principle is true, which is narrower and more interesting than the assertion.

3.1 Aristotle: because of this, not after this

The distinction on which this chapter rests is the oldest in the field, and it is worth recovering in its original form rather than in the version craft manuals inherit.

In the Poetics, Aristotle ranks plots. The worst, he says in chapter 9, are episodic — those in which the episodes follow one another with neither probability nor necessity. The best are those in which events occur contrary to expectation and yet on account of one another: the Greek is δι’ ἄλληλα (di’ allēla), “through one another”, and the phrase does the work this chapter’s “had to happen next” does, two and a half millennia earlier.

Two refinements matter for practice.

• Aristotle’s criterion is probability or necessity, not certainty. The events must seem to arise from one another as this kind of person, in this situation, would act. This is a plausibility standard, not a logical one — which is why the model is a craft instrument rather than a proof procedure.
• His two great structural devices are the ancestors of the disaster beat. Peripeteia, the reversal, is a change to the opposite state of affairs arising out of the action itself; anagnorisis, recognition, is a shift from ignorance to knowledge. Both are ways of ending a unit worse than it began. The “yes, but” disaster of §4.3 is a peripeteia in miniature, and the “got what they wanted and discovered it was not what they needed” variant is anagnorisis domesticated.

The Latin tag usually attached to this — post hoc against propter hoc, “after this” against “because of this” — comes from logic rather than from poetics, where it names a fallacy: the error of inferring causation from succession. Its migration into narrative theory involves a small and instructive irony. In logic, treating succession as causation is a mistake. In narrative it is a competence. Readers do it constantly, automatically and with very little evidence, and the entire art of the causal chain consists in supplying enough of that evidence to make the inference feel earned rather than imposed. §3.7 returns to this.

E. M. Forster’s Aspects of the Novel (1927) supplies the formulation every student already half-knows:

“The king died and then the queen died” is a story. “The king died and then the queen died of grief” is a plot.

Three words — of grief — convert sequence into consequence. Forster’s example is worth pressing further than he presses it, because it demonstrates how cheap the conversion is. He does not dramatise the queen’s grief, motivate it, or show it. He asserts a causal relation in a prepositional phrase, and the reader accepts it. This is the mechanism the “and so / but” test (§7.1, test 6) exploits.

3.2 Brooks: narrative desire, delay and discharge

The most sustained modern theorisation of what the goal–disaster engine actually does to a reader is Peter Brooks’s Reading for the Plot (1984), and it is the scholarly counterpart to the whole of §4.

Brooks’s argument begins by refusing the structuralist habit of describing plot as a static configuration. Plot, for Brooks, is dynamic: it is the “logic of the discourse” as it unfolds in time, and its motor is narrative desire — the reader’s want, produced by the text and sustained across it, for the end that will confer meaning on everything before it. Reading is the experience of being pulled towards a conclusion whose arrival will retrospectively organise the middle.

Two consequences follow that bear directly on the six beats.

First, the middle exists to resist the end it desires. Brooks, reading Freud’s Beyond the Pleasure Principle, describes the narrative middle as a dilatory space — an “arabesque” of detour, delay and repetition whose function is to prevent the short circuit, the “improper end” reached too quickly. A story that discharged its desire immediately would be no story. The disaster beat is the structural device that prevents the short circuit: every time a goal is nearly reached and then converted into a worse problem, the discharge is deferred and the desire renewed. This is the theoretical reason the “clean win” is a structural fault (§8.2) rather than merely a missed opportunity. A clean win is a small death of desire.

Second, plot is bound up with mortality. Brooks’s most quoted formulation describes plot as the “internal logic of the discourse of mortality” — narrative as the form in which a life, which is merely sequential while it is lived, becomes shapely by being read from its end. The relevance to the sequel half is exact. The Reaction beat is where a character registers that something is irreversible; the Dilemma beat is where the narrowing of remaining options is felt; the Decision beat is where a finite quantity of remaining life is spent on one course rather than another. A sequel is a small rehearsal of the fact that choices cannot be unmade — which is why the reversibility test (§7.1, test 9) catches so much weak material.

Brooks also supplies the vocabulary for §6. The techniques of interruption and deferral are not tricks appended to a plot; they are the plot’s own resistance to its ending, made visible.

3.3 The structuralist description: Barthes, Todorov, Propp

Three structuralist accounts describe, from three directions, what the six beats formalise. None was written with craft pedagogy in view, and their independence from it is what makes the convergence worth noticing.

Barthes’s proairetic code. In S/Z (1970), Roland Barthes distinguishes five codes by which a reader processes a narrative text. The proairetic — from proairesis, the rational determination of the result of an action — is the code of actions: the reader’s running capacity to recognise a sequence in progress and give it a name. Barthes’s own examples are gestures like Stroll, Murder, Rendezvous. The point is that the sequence is constituted by the naming: an action becomes legible as a unit because the reader can title it, and titling it involves projecting its completion.

This is what a stated goal does. A goal is a licence to name the sequence. A character who wants the key allows the reader to open a file called getting the key, into which subsequent sentences are filed and against which they are measured. The reason an unstated goal produces flatness is not that the reader is confused about events but that the reader has no title under which to file them, and so files them separately, as a list.

Barthes’s hermeneutic code — the posing and delaying of enigmas — is the proairetic code’s partner and is treated in Chapter 9 (The Opening Pages). The distinction is useful here: the hermeneutic code makes the reader want to know; the proairetic code makes the reader want to see what happens. Suspense in Meir Sternberg’s sense is proairetic; curiosity is hermeneutic. Most strong scenes run both.

Todorov’s equilibrium schema. Tzvetan Todorov, in The Poetics of Prose (1971; Eng. 1977), proposes that the minimal complete plot moves from an initial equilibrium, through a force that disrupts it, through a recognition of the disruption and an attempt to repair it, to a new equilibrium which is not the old one restored but a second state of the same kind. The schema operates at the scale of the whole narrative; the six beats operate at the scale of the chapter. But they have the same shape, and a scene–sequel pair is legible as a Todorovian cycle in miniature: the goal presupposes an equilibrium worth defending or improving, the disaster disrupts, the sequel recognises and repairs, and the decision establishes a new state from which the next disruption can be launched.

Two things follow. Todorov’s model explains why the model nests: the same structure describes a paragraph, a chapter and a novel, which is why a well-made book feels self-similar at several scales. And his insistence that the final equilibrium differs from the first is the theoretical form of Chapter 1’s value shift.

Propp’s functions. Vladimir Propp’s Morphology of the Folktale (1928; Eng. 1958) analysed a hundred Russian wonder tales into thirty-one functions — units defined not by who performs them or how, but by their significance for the course of the action. This is the definitional move that makes a “beat” a beat: a function is a job, and the same job may be carried by any content.

Three of Propp’s functions map onto the model with uncomfortable precision:

Propp’s function X is a decision beat named as such, thirty-seven years before Techniques of the Selling Writer. The convergence is not evidence that Swain read Propp — there is no reason to think he did, and the first English translation appeared only seven years before his book. It is evidence that both were describing the same thing: a strongly conventionalised narrative grammar, one from the outside as an analyst, one from the inside as a working professional. That two such different projects arrive at the same units is the best available argument that the units are real. It is also, as §3.6 argues, an argument that they are real for a particular family of narratives.

3.4 Burke: form as appetite

One further theoretical strand is worth adding because it names the mechanism the others describe.

Kenneth Burke, in Counter-Statement (1931), offers the most economical definition of narrative form in the literature:

Form is the creation of an appetite in the mind of the auditor, and the adequate satisfying of that appetite.

Two words carry the load. “Appetite” makes form a condition of the reader rather than a property of the artefact — the same move Brooks makes fifty years later. “Adequate” concedes that satisfaction admits of degree, and that the relation between the appetite created and the satisfaction supplied is the whole art.

Applied to the six beats, Burke’s formula produces the clearest statement of what each half is for. The scene creates the appetite — a goal is an appetite installed in the reader, and a disaster is an appetite denied and thereby intensified. The sequel converts an unsatisfied appetite into a new one, which is why a sequel without a decision (§8.8) leaves a reader with a hunger and no dish coming.

Burke’s later Grammar of Motives (1945) supplies a second instrument, less often applied to fiction than it should be: the pentad of act, scene, agent, agency and purpose, and the analysis of motive through the ratios between them. A scene–sequel pair can be described as the working-out of an agent–purpose ratio under pressure from scene and agency: the character wants X (purpose) and the situation (scene) will not permit the means (agency) they have. Where a draft’s motivation feels thin, Burke’s question — which term of the pentad is the narrative pretending is not there? — is frequently more productive than asking what the character wants.

3.5 The cognitive evidence: causal networks and what readers retain

The strongest empirical support for the chapter’s first principle comes from discourse psychology, and it says something narrower and more useful than “causality is good”.

Situation models. Rolf Zwaan and Gabriel Radvansky’s synthesis (1998) holds that readers do not store sentences but construct a situation model — a representation of the state of affairs described — indexed on a small number of dimensions. Two of the five are this chapter’s business:

The model’s central prediction is that readers pay a measurable processing cost whenever an index shifts without warning. A new scene whose goal does not descend from the previous decision is, in these terms, an unsignalled shift on the intentionality and causality dimensions simultaneously — which is a precise cognitive description of the “free goal” failure (§8.1) and explains why it registers as effortful rather than merely as odd.

Causal connection and recall. More directly to the point, a body of work associated with Tom Trabasso and colleagues models a narrative as a causal network — events linked by relations of necessity in the circumstances — and finds two robust effects. First, events that lie on the causal chain connecting the opening to the outcome are recalled better, judged more important, and included more often in summaries than events that lie off it, even when the off-chain events are equally prominent on the page. Second, the number of causal connections an event has predicts how well it is remembered. Events with many links are recalled at rates substantially above events with few, independent of their position in the text.

Alan Graesser, Murray Singer and Trabasso’s constructionist account (1994) adds the mechanism: readers engage in a search after meaning, generating causal antecedent inferences as a matter of course — asking, in effect, why is this happening now? — while generating predictive inferences about what will happen next only weakly and inconsistently.

That asymmetry is the most practically useful finding in this section, and it has two consequences that cut against intuition.

1. Readers will work hard to connect a new event backwards. A chain that is there does not need to be underlined. The reader is already looking for the antecedent.
2. Readers will not reliably project forwards. A goal must therefore be stated, not implied, because the reader will not construct it unprompted. This is the empirical case for the goal test (§7.1, test 1), and it is the reason the model insists that a goal be legible before the conflict begins rather than reconstructible afterwards.

A caution, stated plainly. These findings describe comprehension and memory, not literary value. They demonstrate that causally connected material is easier to retain and more likely to be judged important. They do not demonstrate that easily retained narratives are better ones, and a great deal of ambitious fiction deliberately makes retention difficult. The correct inference is modest: the chain is a device for making a narrative cheap to process, and cheapness of processing is a resource a writer may spend or decline to spend.

3.6 The provenance of the model

A university textbook has an obligation the trade manuals do not: to say where its instruments come from. The six-beat model has a specific and traceable history, and knowing it changes how the model should be used.

Dwight V. Swain (1915–1992) wrote for the American all-fiction magazines and produced category fiction before becoming a professor of professional writing at the University of Oklahoma. Techniques of the Selling Writer (University of Oklahoma Press, 1965) is the codification of a commercial practice, and its title is exact about its purpose. It is not a poetics. It is a manual for producing publishable copy in a market that paid by the word and bought reliably shaped stories in volume.

Three features of that market are legible in the model:

Swain also gives the model a second, smaller unit, which most restatements omit and which deserves rescuing: the motivation-reaction unit. At sentence and paragraph scale, Swain argues, narrative alternates between an external motivating stimulus (something happens, presented objectively) and the character’s reaction to it, rendered in a fixed order — feeling, then reflex, then rational action, then speech. Whatever one thinks of the prescribed order, the MRU makes explicit what the six beats imply: the causal principle is meant to operate all the way down, from the paragraph to the book. §4.0 returns to it.

Transmission. The model reached its present ubiquity chiefly through Jack M. Bickham, Swain’s colleague at Oklahoma, whose Scene & Structure (Writer’s Digest Books, 1993) restated the six beats for a new generation in the widely distributed Elements of Fiction Writing series. From there the model passed into the online craft ecosystem — writing blogs, plotting software, self-publishing pedagogy — where it now circulates in a highly simplified form, frequently without attribution and frequently without the qualifications Swain himself supplies. Much of what a contemporary writer encounters as timeless craft wisdom about “scene and sequel” is a fourth-generation copy of a 1965 trade manual.

The historical coincidence. Swain’s book appears at the height of the American programme era — Mark McGurl’s term, in The Program Era (2009), for the period in which creative writing became an institutionalised university discipline and the dominant patron of American literary fiction. Two pedagogies were therefore being built at once, and they taught nearly opposite lessons:

The two traditions were, for decades, almost hermetically separated, and the separation is the reason the six-beat model can still feel like a discovery to writers trained in literary programmes. It is also the reason the model carries assumptions that are worth surfacing. Its prized virtues — agency, escalation, forward motion, external goals — are the virtues of one tradition, historically situated. They are not neutral descriptions of narrative. §8.7 examines the sharpest case: the prohibition on the passive protagonist, which is very close to a genre preference wearing the clothes of a structural principle.

The honest verdict is double, and students should be able to hold both halves. The model is a commercial pedagogy, and its provenance shows. It is also, demonstrably, a good description of how an enormous range of narrative actually works, including a great deal of fiction its author would not have read and would not have enjoyed. Provenance is a reason for care, not for dismissal.

3.7 The first principle, restated and qualified

We can now state the principle.

The causal chain. A scene’s disaster is the specific event the following sequel reacts to; that sequel’s decision is the specific goal the following scene pursues. Where both links hold, units read as consequences of one another; where either fails, units read as items on a list.

And now the qualification, which is the more valuable half.

Causality in narrative is conferred, not contained. Roland Barthes made the point with characteristic sharpness in “Introduction to the Structural Analysis of Narratives” (1966): everything suggests that the mainspring of narrative activity is precisely the confusion of consequence and chronological succession — what comes after being read as what comes because. Forster’s queen dies of grief only because Forster says so; and, more disturbingly for the model, she would die of grief in most readers’ minds even if he did not, provided the two deaths were adjacent and the marriage had been mentioned.

Three practical consequences follow, and they are the reason §7 insists that a failed test is a question rather than a verdict.

1. Adjacency does much of the work the model attributes to structure. A reader will read a sequel as caused by the disaster that precedes it with very little help. This is a resource: it means the causal signalling can usually be lighter than a writer fears, and that heavy signposting (because of what had happened in the barn, she now had to decide whether…) is nearly always a symptom of insufficient trust.
2. The chain can therefore be simulated, and it can be faked. A draft can pass the chain test at the level of stated motive and still feel arbitrary, because the causality is asserted in the connective tissue rather than dramatised in the beats. The chain test (§7.1, test 5) is designed to be run on events, not on the sentences that link them, for this reason.
3. Refusing the chain is a legitimate and well-populated option. The picaresque runs on succession by design; Don Quixote and Tristram Shandy make the arbitrariness of sequence part of the joke; Catch-22 makes the absence of causal logic its central argument (§9.6); and a large body of modernist and postmodern fiction treats causal plotting as a nineteenth-century convention to be dismantled. A model that classifies these as failures has misunderstood its own status.

The principle, then, is not a law of narrative. It is a description of the dominant mode, an account of the cheapest available route to felt necessity, and — most usefully — a diagnostic instrument that tells a writer what kind of book they are currently writing. That last formulation is the chapter’s real position, and §16 returns to it.