Skip to content

Research and Verification for Creative Work

Craft guidance for assessing certainty, structuring research findings, and handling uncertain facts in fiction—from corroborated history to plausible invention.


Why Research Posture Matters

Not all facts are equally certain. Fiction writers need to know:

  • What they can state with confidence
  • What requires hedging
  • What sources conflict
  • What has no backing at all

Posture captures this certainty level, enabling informed creative decisions.


The Posture Framework

Posture Levels

Posture Definition Confidence
Corroborated Multiple reliable sources agree High
Plausible Reasonable based on available evidence Medium
Disputed Sources actively conflict Variable
Uncorroborated No sources found Low

Corroborated

Definition: Multiple reliable sources agree on the claim.

Indicators:

  • 2+ independent, reputable sources confirm
  • Wikipedia plus one other source agree
  • Academic or government sources confirm
  • Multiple news sources report consistently

Surface Treatment: State directly without hedging.

Example:

Victorian London began installing electric street lamps in 1878.

Plausible

Definition: Reasonable to assume based on available evidence.

Indicators:

  • One good source (encyclopedia, textbook)
  • Logical extension of corroborated facts
  • Expert opinion without contradiction

Surface Treatment: Soft hedge—"believed to be," "thought to have been."

Example:

The practice was believed to have originated in the northern provinces.

Disputed

Definition: Sources actively conflict; multiple valid interpretations exist.

Indicators:

  • Sources actively disagree
  • Historical debate ongoing
  • Regional or cultural variation exists

Surface Treatment: Present as in-world disagreement rather than stating one version as fact.

Example:

Historians disagree about what sparked the uprising. Some point to the grain shortage; others blame the governor's decree.

Uncorroborated

Definition: No sources found to support the claim.

Risk Levels:

Risk Impact if Wrong Examples
Low Flavor detail, no story impact Color of tavern sign, background character name
Medium Could cause plot inconsistency How organization made decisions, character motivation basis
High Could break the story Central premise, safety-critical claims, medical/legal facts

Surface Treatment: Use neutral phrasing; avoid definitive claims.

Examples:

  • Low risk: "The old lighthouse had stood for as long as anyone could remember."
  • Medium risk: "It was said the council met in secret, though no records survived."
  • High risk: Requires research escalation—don't proceed with central premises unsupported.

Research Memos

Purpose

A research memo provides decision-ready information to creative workers without overwhelming them with raw data.

Memo Structure

Section Content Length
Question Exact question being answered 1 line
Short Answer Decision-ready synthesis 2-5 lines
Posture Certainty classification 1 word + justification
Citations Sources consulted 2-5 sources
Caveats What's uncertain or context-dependent Brief list
Creative Implications Affordances and constraints 2-4 items each
Neutral Phrasing Player-safe wording options 1-2 lines
Handoffs Who else needs this information Agent/role names

Example Research Memo

Question: Did Victorian London have electric street lights?

Short Answer: Yes, limited deployment from 1878. The Holborn Viaduct
installation was first. Most streets remained gas-lit through the 1880s.
Electric lighting spread unevenly—wealthy areas first.

Posture: Corroborated (Wikipedia + 2 historical archives agree)

Citations:
- Museum of London archives
- "London: A Social History" (Roy Porter)
- British History Online primary sources

Caveats:
- Exact dates vary by district
- Private vs public lighting differed
- Quality and reliability varied significantly

Creative Implications:
  Affordances:
  - Can use contrast between lit and dark districts
  - Class divisions visible through lighting
  - Flickering/unreliable power adds atmosphere

  Constraints:
  - Widespread electric lighting anachronistic before 1890s
  - Don't assume consistent illumination
  - Gas remained dominant for residential areas

Neutral Phrasing:
"The occasional hum of electric lamps marked the newer districts,
though most streets still flickered with gaslight."

Handoffs: @scene_smith (prose), @style_lead (register check)

Research Principles

Answer Narrowly

Problem: Researchers dump everything they found.

Fix: Answer the specific question asked. Provide decision-ready synthesis, not Wikipedia summaries.

Target: 2-5 lines for short answer, 2-5 citations maximum.

Preserve Creative Context

Research should support creative work, not derail it.

Good: "Here's what you need to know for this scene."

Bad: "Here's everything about Victorian street lighting, including manufacturing processes, political debates, and economic analysis."

Report Constraints, Not Decisions

Research identifies what the world allows or forbids. Creative decisions belong to other roles.

Good: "Electric lighting existed in wealthy districts by 1880."

Bad: "You should set the scene on the Holborn Viaduct to use electric lighting."

Never Surface Sources

Players should never see research artifacts.

Never on player-facing surfaces:

  • "According to Wikipedia..."
  • "Sources indicate..."
  • "Historians believe..."
  • Citation markers or footnotes

Instead: Integrate facts naturally into narrative.


Handling Research Gaps

When You Can't Find Sources

  1. Assess risk level — How bad if wrong?
  2. Document the gap — What exactly is unknown?
  3. Escalate if high-risk — Central premises need resolution
  4. Provide neutral phrasing — For low-risk gaps

Gap Documentation

Field Content
Question What was asked
Search performed What sources checked
Gap identified What couldn't be found
Risk assessment Low/Medium/High
Recommendation Proceed cautiously / Escalate / Block

Creative Options for Gaps

Risk Approach
Low Proceed with neutral phrasing
Medium Flag as provisional, note in canon
High Escalate—don't proceed without resolution

Surface Treatment by Posture

How to phrase facts based on certainty:

Corroborated (State Directly)

The factory opened in 1847. The guild controlled the waterfront trade. Three bridges crossed the river.

Plausible (Soft Hedge)

The practice was believed to date back centuries. By most accounts, the family had lived there for generations. The traditional explanation held that...

Disputed (Present Disagreement)

Some said the fire was accidental; others suspected arson. Historians disagreed about the colony's founding date. The truth, as with many things, depended on who you asked.

Uncorroborated (Neutral Phrasing)

No one knew for certain how the tradition began. The origins were lost to time. Whether by accident or design, the record was silent.


Common Research Mistakes

Over-Researching

Problem: Spending hours on details that don't affect the story.

Fix: Research what matters for the scene. Background details can be plausible without being verified.

Under-Researching

Problem: Getting central facts wrong, breaking reader trust.

Fix: Verify anything readers might know or check. Historical dates, famous events, technical processes.

Treating Fiction as Source

Problem: Using other novels or movies as factual reference.

Fix: Fiction is inspiration, not evidence. Verify against primary or academic sources.

Assuming Universal Knowledge

Problem: What's common knowledge varies by reader.

Fix: Consider your audience. Technical readers catch technical errors.

Displaying Research

Problem: Showing off research instead of serving the story.

Fix: Use research to write confidently, not to demonstrate expertise.


Quick Reference

Posture Definition Surface Treatment
Corroborated Multiple sources agree State directly
Plausible One good source or logical extension Soft hedge
Disputed Sources conflict Present as disagreement
Uncorroborated No sources Neutral phrasing, assess risk
Risk Level Impact Action
Low Flavor detail Proceed with neutral phrasing
Medium Affects plot Flag as provisional
High Central premise Escalate—don't proceed

See Also