Narrative Intelligence
Conversation is not public opinion. Attention is not persuasion. Sentiment is not vote preference.
Narrative intelligence is the systematic interpretation of how accounts of events emerge, spread, acquire emotional meaning and become reinforced—and whether those patterns correspond to broader opinion or behavior.
This definition brings together recurring ideas in my writing. It organizes a field of observation; it does not establish that every observed story causes an opinion to change.
Distinguishing the observations
Attention and conversation volume
Attention concerns how much observable activity a subject attracts. Volume can reveal that a discussion has become more visible. It does not establish how many independent people hold a view, whether participants represent a wider population, or whether anyone has changed their mind. Repeated activity by a small group can look substantial in an aggregate count.
Sentiment
Sentiment describes the expressed emotional or evaluative tone within that activity. It concerns what is expressed in a particular setting, not direct access to private belief. Sarcasm, context and the object being discussed can complicate interpretation. A positive-sounding phrase may express criticism rather than approval.
Sentiment analysis is an input to narrative intelligence, not the whole discipline. Classifying tone does not by itself explain the account of events through which people are interpreting an issue, or whether that account persists beyond the moment.
Narrative formation
Narrative formation is the process by which separate events begin to be interpreted through a shared account. The research question concerns connections and meaning, rather than a collection of keywords. An analyst’s ability to construct a coherent story is not evidence that participants have adopted that story themselves.
Amplification and velocity
Amplification concerns how widely an account circulates; velocity concerns how quickly that circulation changes. Those observations need a defined environment and interval. Rapid repetition may reflect concentrated or coordinated activity. Narrative velocity therefore should not be described as opinion velocity without separate evidence about the population’s views.
Adoption and reinforcement
Adoption concerns whether independent participants begin using a frame themselves. Reinforcement concerns whether they sustain and reproduce it over time. These are more demanding observations than a large number of mentions. The distinction matters because quoting, criticizing and adopting a narrative are different actions that may all contribute to its visibility.
Narrative lock-in
Narrative lock-in describes alternative interpretations becoming increasingly difficult to sustain within the observed environment. It is a conditional interpretation of persistence and narrowing, not proof that an entire electorate has settled its views. A coherent account in one community may be absent or contested in another.
Persuasion, opinion and behavior
Persuasion concerns change; opinion concerns attitudes or preferences; behavior concerns what people do. Each requires evidence appropriate to the question. A shift in expressed tone does not automatically establish changed preference, and a changed preference does not automatically establish changed behavior. Temporal order alone does not show that one caused the other.
What the discipline cannot replace
Narrative intelligence should not be treated as a substitute for representative survey research. Accessible conversation supplies observations about an information environment. A survey supplies structured observations from a defined sample. Both have limitations, and neither becomes representative merely because it contains many observations.
This distinction also limits retrospective stories. An account of an episode can illustrate a possible mechanism without demonstrating that it explains the outcome.
Signal Scarcity
AI and inexpensive content production have made information abundant. The harder problem is determining which signals are authentic, representative and relevant to the question we are trying to answer.
Authenticity asks whether activity is genuine human expression or artificially generated or coordinated. Coordination and automation are not identical, but either can complicate an interpretation based on apparently independent expression.
Representativeness asks whether even genuine activity tells us something meaningful about the population we care about. Authentic speech can still come from a small, unusually active or otherwise unrepresentative group.
Relevance asks whether activity bears on the outcome or question being studied. A genuine, widely shared reaction may tell us about attention without answering a question about preference or behavior. These three tests address different problems; satisfying one does not settle the others.
Signal Scarcity is currently treated here as an evidence-quality problem within Narrative Intelligence, not a separately validated framework.
My earlier writing expressed greater confidence in the resistance of emotional patterns to fabrication. Later work recognizes more clearly that synthetic and coordinated activity can obscure genuine signal. That development matters: emotional signals should not be treated as immune to manipulation. A stronger account of the discipline acknowledges this tension rather than pretending the language has always been identical.
A useful test: the convention experiment
The paired essays Can a Convention Still Create Momentum? and The Balloons Fell. Here’s What Moved. are useful because of their methodological structure. The first defined what would be observed; the second reported changes in conversation and narrative indicators without treating them as proof of vote movement.
Their value is the connection between a stated question and a bounded conclusion. The pair does not establish a universal mechanism by which conventions change opinion. It demonstrates why the observation being reported and the larger outcome people may care about need to remain distinct. Read the two together in the further reading below.
Further reading
The Slopaganda Election — Signal scarcity and the problem of manufactured activity.
How Races Collapse Before Votes Move — Consolidation, compression and the limits of apparent lock-in.
Narrative Velocity: Why Speed, Not Persuasion, Decides Modern Campaigns — The speed of narrative circulation, distinct from measured opinion change.
Outrage vs. Impact: How to Tell What Voters Actually Care About — Why visible controversy is not automatically consequential impact.
What the Most Obnoxious College Football Fans Can Teach Us About Momentum — An accessible account of momentum and the limits of attention.
Can a Convention Still Create Momentum? — The convention pair, part one: the question and planned observations.
The Balloons Fell. Here’s What Moved. — The convention pair, part two: reported changes and restrained conclusions.
The Narrative Machine — What Happens When AI Starts Writing the Story Too? — The earlier confidence about emotional patterns that later writing complicates.
Watch
EyesOver/Stratus Intel CEO Chris Wilson discusses the Vice Presidential debate on Fox News
Earlier discussion of a related distinction, around 2:21: a memorable debate moment is not the same as a broader trajectory. This is an analogy, not evidence that conversation caused opinion change. Uploaded August 6, 2025; the clip concerns the 2012 debate, despite the current-role wording in its upload title.
Christopher S. Wilson 