Reference
Deep Narrative Capture (DNC)
The governed narrative-reading layer: how clearly, credibly, and consistently an actor is framed in AI interpretation. An overlay, never folded into the score.
This page defines Deep Narrative Capture as a public reference concept. It is descriptive and does not disclose proprietary prompts or methods.
Canonical Reference · Reference Library · v1.2 · Last reviewed 2026-06-03
Applies to: Observatory readings that report the narrative layer — role clarity, trust, narrative fragility
Status: Definitions govern downstream Observatory usage
Quick answer
Deep Narrative Capture is the governed reading of how a model frames actors in a captured response about a market. The response is recorded verbatim; the public reading may select, contextualize and interpret it without silently changing it. Readings are scoped and dated. Deep Narrative Capture remains a narrative overlay, separate from the Market Authority Score and official position readings.
Core definition
Deep Narrative Capture is the governed narrative-reading layer derived from captured model responses about a market. The response is recorded verbatim; selection, context and interpretation belong to the public reading. It describes observed framing within a stated scope and window and remains separate from the Market Authority Score and official position readings.
Interpretation rules
- Deep Narrative Capture reads the narrative layer around the measured signals, observed in long-form AI interpretation — not asserted.
- It is kept as an overlay; it is never folded into the Market Authority Score or the official position readings.
- The public reading may select, contextualize and interpret captured responses without silently changing quoted material. Authorized model, surface and mode provenance may be public; private prompts, orchestration and restricted method details remain private.
- It reads public outputs only — never model weights or internal reasoning.
Purpose & scope
Why this page exists. If a narrative reading silently entered the score, both readings would weaken; kept as an overlay, Deep Narrative Capture adds depth without distorting the position reading.
Where it applies. Every reading that reports role clarity, trust, or narrative stability applies this definition.
What Deep Narrative Capture captures. Standard measurement reads presence, position, and citation. Deep Narrative Capture reads the narrative layer around those measured signals: whether an actor is framed with seriousness, whether its role is understood clearly, whether trust surrounds it, and whether its narrative is stable or fragile. These are observed in long-form AI interpretation, not asserted.
Why Deep Narrative Capture is kept separate. The strength of Deep Narrative Capture depends on not contaminating the measured layer. If a narrative reading silently entered the score, both readings would weaken. Kept as an overlay, Deep Narrative Capture adds depth without distorting the position reading: the official layer answers "what does the governed evidence support," and Deep Narrative Capture answers "how is that reality being narratively processed."
What is published, and what is not. The Observatory publishes selected, scoped and dated narrative readings derived from captured model responses. Quoted response text remains faithful to the capture; the public reading can add context and interpretation. Authorized model, surface and mode provenance may be published with the reading. Private prompts, orchestration and restricted method details remain private. Publication does not imply access to a model’s internal reasoning.
What it is / what it is not
It is
- the governed narrative-reading overlay (role clarity, trust, narrative fragility)
- observed in long-form AI interpretation, scoped and dated
It is not
- part of the Market Authority Score
- a reading of model weights or internal reasoning
- a measure of the business market
Common misinterpretations this page prevents
“The narrative reading is part of the score.”
Deep Narrative Capture is kept as an overlay and is never folded into the Market Authority Score; the official layer answers "what does the governed evidence support," the overlay answers "how is that reality being narratively processed."
“It reveals how the model reasons internally.”
It reads observed framing in public AI interpretation — not model weights, internal ranking systems, or private reasoning.
How this page governs downstream readings
Where a downstream output blends narrative into measurement, this definition prevails:
- A verdict reports narrative observations as an overlay, separate from the score.
- A card that reads role clarity labels it as a narrative reading, scoped and dated.
- A Decision Guide keeps "how it is framed" distinct from "what the evidence supports."
Relationship to other references
- Official vs Narrative Layer — Deep Narrative Capture is the narrative side of that separation.
- Market Authority Score — the overlay is kept out of the score by design.
- AI Behavior Intelligence — the narrative layer sits beside the measured signals.
- Signal Styles — narrative framing builds on the role an actor is assigned.
- Attestation Standard — narrative observations are reported scoped and dated.
- DNC Lens — publishes each capture, with the model’s words recorded verbatim.
- Model Watch — compares captured readings of the same market.
Canonical notice
This is a canonical reference. It defines how Deep Narrative Capture is read and published in Observatory public readings. Authorized model, surface and mode provenance may be published; private prompts, orchestration and restricted method details remain private. That definition governs every downstream use across the site. It does not itself rank, promote, or measure commercial outcomes. Where an editorial page, a card, or a market reading appears to conflict with this definition, this reference prevails.
Intelligence by AAI Kernel. Published by The Observatory.
Deep Narrative Capture reads observed framing in public AI interpretation within a defined scope and window. It does not read model weights, internal ranking systems, or private model reasoning, and it is never folded into the score.
Related references
- Official vs Narrative Layer — because DNC is the narrative side of that separation.
- Market Authority Score (MAS) — because the overlay is kept out of the score by design.
- AI Behavior Intelligence — because the narrative layer sits beside the measured signals.
- Signal Styles — because narrative framing builds on the role an actor is assigned.
- Attestation Standard — because narrative observations are reported scoped and dated.