Break the Premise on Purpose: A Five-Pass Consistency Audit Before Readers Do

A fictional world usually does not collapse because the creator lacks ideas. It collapses because several individually good ideas cannot all be true at the same time.

The cure is not a larger lore bible. It is a deliberate consistency audit that asks the premise to survive contact with consequences. If a city is isolated by mountains, trade should react. If magic is cheap, medicine, policing, labor, war, and status should react. If one institution controls resurrection, inheritance and family law should react. If a rule matters only when the plot remembers it, the problem is not missing detail; it is broken causality.

This audit is designed for a core premise before a major draft, game build, licensing bible, or team handoff. It does not tell you what your world must contain. It helps separate canon, evidence, inference, and invention, then tests whether decisions still agree with one another.

Pass 1: state the premise in one paragraph with verbs

Do not begin with a glossary. Write one paragraph that describes what the world does.

Weak premise notes often look like this:

The empire is ancient. Magic is rare. The north is dangerous. The priesthood is powerful.

Those statements create atmosphere but few testable relationships.

A stronger working paragraph uses verbs:

The empire taxes legal spellcasting, licenses healers through temple courts, and keeps the northern pass open because three inland cities depend on imported salt. Unlicensed magic is rare in public but common in household repair.

Now the premise can be challenged. Who collects the tax? What happens when the pass closes? Why do households tolerate illegal repair magic? Which institution benefits from healer licensing? A premise with verbs produces consequences.

Audit question: Can every major premise statement produce at least one observable behavior, cost, restriction, or conflict? If not, it may be mood rather than a rule.

Pass 2: build a consequence chain before adding more lore

Choose the three rules that would be most expensive for the world to ignore.

For each one, draw a five-link chain:

Rule → immediate behavior → institution → secondary effect → story pressure

Example:

Teleportation requires a licensed gate
→ merchants consolidate freight near gate cities
→ gate inspectors and bonded warehouses become important
→ land near gates becomes expensive
→ smugglers develop alternatives and political factions argue over new gates.

The purpose is not to predict every consequence. It is to prove that the rule reaches beyond the paragraph where it was invented.

If a supposedly transformative power leaves transport, labor, law, military planning, and ordinary habits unchanged, either the power is less important than the text claims or the world has not yet responded to it.

This is where creators often add explanation when they actually need to change a rule. Do not write three pages of exceptions to rescue a premise that produces outcomes you do not want. Adjust the premise until its natural consequences point toward the stories you want to tell.

Pass 3: run a contradiction table, not a memory test

Memory is a poor database for canon. Use a table with one row per decision and at least these columns:

The crucial column is conflicts with. Most lore systems are designed to store facts, not disagreement. An audit becomes useful when it makes unresolved friction visible.

Zotero collections, tags, and notes can help separate research evidence from creative decisions. Git or another version-control system can mark release points and preserve history. The specific tool matters less than the discipline: a reader should be able to answer, “Which version made this true, and what else depends on it?”

Do not use a tag such as canon as a substitute for a decision state. A useful state model is more explicit: idea, working, locked, deprecated, conflict.

Pass 4: test the premise from four ordinary viewpoints

Worldbuilding often looks coherent from the throne room and breaks in a kitchen.

Take the same premise and ask what it means to four people who do not control the system:

1. a child;
2. a worker;
3. a small business owner or craftsperson;
4. a person who dislikes or cannot access the dominant institution.

If legal magic requires expensive licensing, what does a poor household do when a tool breaks? If travel between realms is possible, what paperwork, superstition, insurance, quarantine, fraud, or family separation appears? If dragons are real and occasionally cross farmland, what do ordinary farmers change before the hero arrives?

This pass catches “elite-only causality,” where the world responds only through kings, armies, councils, and named heroes.

It also catches scale problems. A rule may work for one dramatic scene but become absurd when multiplied across a city. Ask not only “Can this happen?” but “What happens when it happens a thousand times?”

Pass 5: attack the premise with three kinds of edge case

A strong rule survives unusual cases without needing the author to secretly turn it off.

Use three categories.

Boundary case: What happens at the edge of the rule? If healing magic closes wounds but cannot replace blood, where exactly does its usefulness end?

Adversarial case: What would a clever person do to exploit the rule? If contracts are magically enforced, can someone manipulate definitions, jurisdiction, consent, or identity?

Collision case: What happens when two rules apply at once? If divine sanctuary protects fugitives but imperial law requires extradition, which authority wins and what is the cost?

Do not solve every edge case. Decide which ambiguity is intentional and which is merely unexamined. Intentional ambiguity can create story pressure; accidental ambiguity creates continuity errors.

The audit should record the result as a decision, not as an essay. “Sanctuary blocks arrest for three nights; temple owes compensation if claimant is fraudulent” is more useful than two pages explaining the history of sanctuary.

Separate source evidence from creative invention

When a fictional premise draws on real history, religion, material culture, law, medicine, or technology, record what the source actually supports.

A research card can have four fields:

• Source fact: what the source directly says or shows.
• Interpretation: your reading of why it matters.
• Creative transformation: what you changed for the fictional world.
• Canon rule: the final rule that now belongs to your setting.

This prevents a common failure: an invented rule later being remembered as “historically accurate,” or a single museum object being treated as universal evidence for an entire culture.

Tools such as Zotero are useful precisely because the same item can be grouped, tagged, annotated, and searched without turning it into canon automatically. Research should remain traceable even after the fictional transformation becomes more dramatic.

Freeze a release, not the entire world

A consistency audit should end with a release point: “Premise v1.0,” “Series Bible 2026-10,” or another explicit marker.

Git’s tagging model is a helpful analogy: a tag marks an important point in history. The same principle works even if the team never uses Git. Freeze the package that downstream work relies on, then allow future changes through a visible change log.

For each change after the freeze, record:

• the old rule;
• the new rule;
• why it changed;
• which characters, maps, scenes, products, or articles are affected;
• who approved it;
• whether the change is retroactive.

This prevents “silent canon drift,” where one person updates a character sheet but the map, game text, website, and marketing deck keep the old version.

The red-team meeting: try to break the premise in 30 minutes

Before a major production milestone, ask one person who did not write the premise to attack it.

Give them only the premise paragraph, rule table, and consequence chains. Their job is to ask:

• What can I exploit?
• What ordinary behavior makes no sense?
• Which two rules collide?
• What should exist in the economy or culture if this rule is real?
• Which rule disappears whenever it becomes inconvenient for the plot?
• Which “fact” has no traceable source or creative decision?

The creator should not defend every answer in real time. Log the questions first. Defensiveness is a poor audit method.

Afterward, classify each issue as fix now, intentional uncertainty, future expansion, or not a problem. That classification is the deliverable.

A premise is ready when consequences are easier to predict

The goal of a consistency audit is not a world with no mystery. It is a world where the team can predict how established rules behave.

A core premise is ready for wider production when a new writer can take an unfamiliar situation, apply the same rules, and reach an answer close to the creator’s answer without asking for a private explanation.

That is the real test. Not the number of pages in the lore bible, but whether the premise produces consistent decisions.

Audit numbers as a system, not as decoration

Numbers create hidden promises. A map scale, travel time, army size, population, spell cost, calendar length, lifespan, or price can connect parts of the world that the writer originally designed separately.

During the audit, collect every number that can constrain another decision. Then ask what each number implies.

If a courier covers 300 kilometers in a day, the transportation system needs an explanation. If a capital has two million residents but food arrives through one mountain pass, storage and supply disruption become major political facts. If a healing ritual costs one month of an artisan’s income, access will not be evenly distributed even if the magic itself is technically available. If a war has continued for 80 years, the soldiers, institutions, memory, and economic structure at year 80 should not behave as if the conflict began last spring.

A simple numeric audit has four columns:

The goal is not realistic simulation for its own sake. Fantasy can use impossible numbers. The audit asks whether the rest of the setting believes those numbers.

Also watch for false precision. A lore document that says a city has exactly 1,843,217 residents creates more implied certainty than “about 1.8 million” unless the exact figure has a reason to exist. Precision should come from the viewpoint or system that could actually measure it.

Finally, mark every high-impact number with an owner and version. If the map scale changes, the travel table must be rechecked. If the magic cost changes, the economy and plot constraints may need review. Numbers are not decorative facts; they are dependencies.

Sources

1. Zotero Documentation — Collections and Tags. https://www.zotero.org/support/collections_and_tags/
2. Zotero Documentation — Notes. https://www.zotero.org/support/notes
3. Git Book — Tagging. https://git-scm.com/book/en/v2/Git-Basics-Tagging.html
4. Zotero Documentation — Searching. https://www.zotero.org/support/searching

Related Reading

• A Repeatable Core-Premise Workflow
• Core-Premise Tools and Templates
• Advanced Core Premise Without Losing Clarity