Narrative

011 min

Narrative at a glance

• What it is: A time-ordered account where events affect one another and someone's goals are at stake.
• Origin: Bruner's 1991 essay listed its features; Green and Brock measured its persuasive effect in 2000.
• Use it when: The audience must understand why something happened or decide what to do next.
• Watch for: A tidy account can feel true without being true, and a fiction label barely reduces its effect.

021 min

The problem Narrative solves

A team finishes a review with twelve correct findings and no shared account of what happened. The sender knows the order of events and why each one mattered. The receiver gets the list and has to rebuild the order, the causes and the stakes alone. Two receivers often rebuild them differently.

The sender feels this cost last. It usually appears in a later meeting, when people disagree about what the review actually found. By then the original context is gone.

A narrative supplies the links that a list leaves out: what came first, what caused what, who wanted what, and what is still undecided. Part of the gap is the curse of knowledge. The sender knows the links so well that they stop seeing which ones the receiver lacks.

032 min

How Narrative works

A narrative is an account of events that unfold over time. Bruner calls this feature diachronicity, which means the order of events is part of the meaning. Change the order and the account changes. In his 1991 essay he listed ten features of narrative. Five of them matter most to anyone writing one.

Canonicity and breach is the feature that most often separates a narrative from a report. Bruner says that not every ordered sequence of events is a narrative, even when it is about people. A canonical script is the course of events people expect in a situation, such as how a release normally goes. Bruner writes that a tale must be about how an implicit canonical script has been breached, violated, or deviated from. The breach is the part worth telling.

Green and Brock add a requirement from the audience side. They write that a narrative account requires a story that raises unanswered questions, presents unresolved conflicts, or depicts not yet completed activity. The audience keeps reading because something is still open. When the open question is answered, the account ends.

Green and Brock also proposed a mechanism for the persuasive effect. They called it transportation, defined as absorption into a story, with imagery, affect and attentional focus. That mechanism has its own entry, Narrative Transport. The point here is narrower: narrative form supplies what transportation needs, which is a character, events in order and a problem that is not yet resolved.

041 min

Where Narrative comes from

Narrative has long been studied in literary theory. The version used in psychology and communication comes mainly from Jerome Bruner's 1991 essay "The Narrative Construction of Reality", published in the journal Critical Inquiry. Bruner says psychologists became aware of narrative as a form not only of representing but of constituting reality about a decade earlier. He dates that change to a 1981 collection of essays called On Narrative.

Bruner also credits William Labov, who studied oral accounts of personal experience, with treating temporal sequence as essential to narrative. This is a theoretical framework from a psychologist's essay. It is not a tested result. The claims about persuasion come later, from the experiments of Melanie Green and Timothy Brock in 2000.

051 min

Before and after

This scenario is illustrative. A product manager must report a delayed payments migration to executives in a weekly status email.

Before:

• Migration is two weeks behind plan.
• Vendor changed their API.
• Retry logic needs rework.
• Decision needed on launch date.

After:

Our plan assumed the vendor's payment API would stay the same until launch. In week three the vendor changed how it reports failed payments. With the old retry logic, some customers would have been charged twice. We stopped the migration, rewrote the retry step and retested it. That moved launch by two weeks. One decision is still open: launch on the new date, or launch on the original date without refund support and add it later. We need an answer by Thursday.

The "Before" version orders the facts by topic and leaves the reader to work out the cause. The "After" version orders them by time, states the expectation, names what broke it, says who acted, and ends on the open question. Each of Bruner's five features in the table above is present in it.

062 min

What the research on Narrative measured

Green and Brock published four experiments in 2000, with 97, 69, 274 and 258 participants. They built a transportation scale to measure how absorbed a reader was in a story. Their main story, "Murder at the Mall", was adapted from Sherwin Nuland's 1994 book How We Die. It is a true account of a child attacked at a shopping mall.

To test whether the scale could tell stories apart, they wrote a much less compelling version called "Bubbles in the Mall". In a sample of 38 undergraduates, readers reported less transportation into it. The scale means were 62.31 for "Bubbles" and 73.94 for "Murder".

Across the four experiments, readers who were more transported held more story-consistent beliefs and rated the main characters more favourably. In the second experiment, highly transported readers found fewer false notes in a story than less transported readers. In the fourth, the authors lowered transportation with processing instructions, and story-consistent beliefs fell with it.

A later meta-analysis by Braddock and Dillard (2016) combined many studies. It found positive relationships between exposure to a narrative and narrative-consistent beliefs, attitudes, intentions and behaviors. The correlations were small. Here k is the number of studies, N the number of participants and r the correlation.

The behavior row rests on only five studies, so it is the least certain of the four.

071 min

A second case: Amazon's six-page memos

The first case was one update written by one person. This case is a rule applied to a whole company. In his 2017 letter to Amazon's shareholders, Jeff Bezos wrote: "Instead, we write narratively structured six-page memos." He says Amazon does not use slide presentations. Each meeting begins with everyone reading a memo in silence.

The one variable that differs from the first case is who decides. In the status email, a single writer chose narrative form. At Amazon, the company requires it. A requirement changes the cost. Bezos says leaders often wrongly believe a high-standards, six-page memo can be written in one or two days. He says it might take a week or more, with several rounds of rewriting and review by colleagues.

What the two cases teach together is that narrative form is a choice about the reader's time as well as the writer's. The writer spends more to put the links in. The readers spend less rebuilding them in the meeting. The letter names the form but, in the passage quoted here, does not list its parts, so Bezos's own text cannot tell you which features of narrative Amazon requires. Bruner's five features in the table above are a reasonable checklist for a team that wants to test its own memos.

081 min

A script to use tomorrow

Use this when you send a status update or open a review. Replace each bracket with your own details.

Say the breach in one sentence and name the person or team it affected. Leave the numbers for after the account, so the audience has something to attach them to.

091 min

The overuse trap

The trap is the narrative fallacy: treating a coherent account as proof that the causes it names are the real ones. It has its own entry, Narrative fallacy. A narrative links events in a way that feels like cause, whether or not the links are true.

Bruner explains why this can happen. He argues that narrative acceptability is governed by convention and narrative necessity rather than by empirical verification. A narrative can be well formed and wrong. Green and Brock found that transportation and beliefs were generally unaffected by labeling a story as fact or as fiction. A reader's guard does not go up because the story is said to be invented.

Watch for these tells:

• The account was written after the outcome was known, and every event points toward it.
• One case stands in for many, and no one has checked whether it is typical.
• Events that do not fit have been left out.
• No one can say what evidence would count against the account.

When two or more of these appear, put the numbers next to the account and name at least one alternative explanation.

101 min

Narrative vs. nearby concepts

The nearest neighbour is storytelling. Narrative is the form: the account itself, with its sequence, particular events and breach. Storytelling is the act of telling it, which includes voice, timing and delivery. A written memo can be a narrative without anyone performing it.

The last row comes from Bruner. He contrasts narrative with logical and scientific procedures, whose results can be removed by falsification. Narratives can only achieve verisimilitude, which means they are judged by how believable they are.

112 min

Where the evidence on Narrative is contested

The strongest objection is that narrative does not beat other evidence on every outcome. Zebregs and colleagues compared statistical and narrative evidence in a 2015 meta-analysis of health communication. Statistical evidence was found to have a stronger influence than narrative evidence on beliefs and attitude, whereas narrative evidence had a stronger influence on intention. The authors explain this by matching evidence type to the outcome. Numbers relate more to cognitive responses and stories to affective ones.

A second objection is about how much is known. Braddock and Dillard report mixed results for fiction against non-fiction, and they say that unidentified moderators are probably present. A moderator is a factor that changes the size of an effect. Some of the effects they report are also small, with correlations between .17 and .23.

A third concerns the original experiments. Three of Green and Brock's four experiments used one main story, which is a narrow base for a general claim. The authors did add a second story in the fourth experiment.

Bruner raises a different problem. If narratives are judged by believability, then no test of a narrative can show it is wrong. That makes it a poor tool for settling a factual dispute.

Where does this leave a practitioner? Use narrative when the goal is understanding, action or intention. Use numbers when the goal is a belief that must hold up to challenge. When the stakes are high, use both, and put them side by side so the audience can check one against the other.

121 min

How Narrative changed since

Three changes stand out. First, narrative moved from literature into psychology. Bruner dates that move to the 1981 collection On Narrative, and his 1991 essay set out the features that psychologists could then use.

Second, narrative became something to measure. Green and Brock noted in 2000 that persuasion scholars had studied advocacy messages, not narrative messages, for the past half-century. They also pointed out that a recent graduate-level textbook with 2,800 references had no entry for the persuasiveness of narratives. Their transportation scale gave researchers a way to test it.

Third, results were combined. The 2015 meta-analysis by Zebregs and colleagues and the 2016 meta-analysis by Braddock and Dillard moved the question from "can a story persuade" to "for which outcomes, and by how much".

Current usage has drifted from the original meaning. In workplace writing, "narrative" is often used loosely for any message or brand story. That is not the sense used here, and none of the research above measured it.

132 min

Frequently asked questions about Narrative

?8 questions

Questions people ask

§5 sources

Sources