The following is an illustrative production case, not a claim about a real studio.
A small fantasy team has a writer, a research editor, a concept artist, and a narrative designer. Everyone is competent. Everyone has read the same underworld brief. After six weeks, the project still contains three incompatible versions of death.
The problem is not creativity.
The problem is that each discipline is editing a different layer of the canon without a shared change protocol.
Here is how the team fixes it.
Background: one underworld, four mental models
The original brief contains these rules:
- the dead lose ordinary physical hunger;
- memory decays unless anchored;
- only licensed guides can cross certain boundaries safely;
- judgment is not a single moral score;
- communication with the living is rare and regulated.
The writer treats these as plot rules.
The researcher treats them as claims that need historical boundaries.
The artist treats them as visual prompts.
The narrative designer treats them as interaction constraints.
All four interpretations are reasonable. The trouble begins when each person silently fills gaps.
The writer invents a feast scene where the dead eat normally.
The artist designs coin-operated gates after seeing popular imagery of underworld payment.
The researcher adds a note about Greek and Egyptian death traditions but does not mark which details are source evidence and which are merely comparative prompts.
The narrative designer creates a mechanic where players summon any deceased character from any checkpoint.
Nothing is individually absurd.
Together, the canon has stopped meaning anything.
Error 1: the research note looks like canon
The research editor shares a page containing:
- Homer, Odyssey Book 11;
- Egyptian funerary objects;
- a museum explanation of the Book of the Dead;
- modern visual references.
The page has no field identifying evidence, interpretation, or invention.
The artist sees “journey to the dead + funerary object + judgment” and assumes all three are approved components of the fictional afterlife.
Correction
The team changes every research card to four fields:
Source evidence: what the source actually shows or says.
Context: tradition, period, genre, object type, uncertainty.
Design question: what structural question this raises.
Canon answer: the team's original answer, if approved.
Example:
- Evidence: Book 11 of the Odyssey contains a literary encounter between Odysseus and the dead.
- Context: Greek epic; one text is not a universal map of Greek afterlife belief.
- Design question: what procedure makes contact with the dead possible?
- Canon answer: the project's dead can be contacted only through licensed memory anchors.
Now research expands design without becoming accidental canon.
Error 2: visual language creates mechanics nobody approved
The concept artist draws a gate with seven seals because it creates a strong silhouette.
The writer later assumes the seven seals represent seven legal tests.
The narrative designer turns them into seven gameplay keys.
Three weeks later, everyone remembers “the Seven Seals” as if they were always part of the mythology.
Correction
Every concept sheet gets a small legend:
- CANON FUNCTION — approved world rule;
- VISUAL EXPLORATION — image language only;
- OPEN QUESTION — attractive idea not yet canon;
- NON-CANON VARIANT — deliberate alternative.
The gate can keep seven visual plates if the composition works, while the legend says they have no approved metaphysical function.
This tiny label prevents aesthetics from silently writing law.
Error 3: the writer changes a rule locally
A scene needs an emotional reunion. The writer lets a dead parent appear physically for ten minutes.
The scene works beautifully.
It also destroys three prior constraints: return is rare, contact requires a licensed anchor, and the dead do not simply regain ordinary bodies.
The writer planned to “fix continuity later.”
Correction
The team creates a canon change ticket.
It contains:
- old rule;
- proposed change;
- reason;
- affected scenes/assets/systems;
- exceptions created;
- decision;
- migration work.
The reunion can still happen, but it must become either:
- an explicit exception with a cost;
- a misinterpretation by the viewpoint character;
- a memory construct rather than physical return;
- or a genuine rule change whose consequences are updated everywhere.
Creative freedom remains. Invisible rule drift does not.
Error 4: historical sources are flattened into a shared aesthetic
The team's mood board mixes Egyptian funerary imagery, Greek literary underworld scenes, East Asian visual motifs, and original cyber-fantasy architecture under one heading: “Ancient Death Culture.”
That label erases the differences that research was supposed to preserve.
The British Museum's Egyptian funerary materials and Homeric epic do not describe one common historical system. Museum objects, literary texts, ritual practices, and later retellings have different evidentiary roles.
Correction
The mood board is split into:
- source tradition;
- period;
- source type;
- what is visually observed;
- what must not be generalized;
- original transformation.
The team also creates a rule: no source-derived motif enters final art merely because it looks “death-coded.”
It must have either a documented canon purpose or be transformed far enough to belong to the fictional culture.
Error 5: nobody owns contradiction resolution
At the end of the week, the writer notices “the dead cannot eat” in one file and a banquet painting in another.
Everyone assumes someone else will decide.
Correction
The team appoints a canon editor for contradictions.
This role does not own all creative decisions. It owns the queue.
Every conflict receives one status:
NO_CONFLICT— two files describe different scopes;CANON_ERROR— one asset violates the approved rule;RULE_CHANGE_PENDING— the team is considering a new rule;EXCEPTION_APPROVED— the deviation is intentional and bounded;NON_CANON— asset remains exploratory.
A contradiction without a status cannot silently ship.
Correction rule: separate “must stay stable” from “safe to explore”
The team also marks every canon field with one of three stability levels.
Hard constraint means changing it would affect many downstream assets. Soft constraint means the current answer is preferred but can move through a normal change ticket. Exploration space means teams are encouraged to generate alternatives without approval.
This prevents an equally damaging mistake: treating every sentence in the lore bible as sacred. Collaboration fails both when nothing is controlled and when everything is frozen. A visible stability level tells people where experimentation is cheap and where a casual edit would create expensive continuity debt.
Result: fewer meetings, more explicit freedom
After the change, the team does not become more rigid.
It actually gets faster.
The concept artist can explore ten gate designs because nobody mistakes exploration for canon.
The writer can propose a resurrection exception because the impact is visible.
The researcher can bring in contradictory historical material without being pressured to turn it into one clean “mythology.”
The narrative designer knows which constraints are hard rules and which are open.
The biggest improvement is not “perfect continuity.”
It is that disagreement becomes inspectable.
A collaboration packet that fits in one folder
For each underworld feature, keep six files or sections:
1. Canon card
- approved rule;
- scope;
- exceptions;
- owner;
- last change.
2. Research card
- source;
- evidence;
- context;
- uncertainty;
- design question.
3. Visual brief
- required function;
- prohibited implication;
- visual freedom;
- exploration labels.
4. Interaction brief
- what players/characters can do;
- what they cannot do;
- cost;
- failure state.
5. Change log
- old rule;
- new rule;
- reason;
- affected assets.
6. Contradiction queue
- conflict;
- files involved;
- status;
- decision owner;
- deadline if needed.
This is enough for a small team. The system does not need enterprise software.
The transferable rule
Historical research and fictional canon should collaborate through questions and evidence, not through silent copying.
Perseus gives access to classical texts such as Odyssey Book 11. Museum resources from the British Museum and the Metropolitan Museum provide object and curatorial context for Egyptian funerary materials. These sources can sharpen a team's questions while remaining distinct traditions and source types.
The collaboration process should preserve that distinction.
A researcher should be able to say, “This source complicates our assumption.”
An artist should be able to say, “This shape is visual exploration, not metaphysical law.”
A writer should be able to say, “I want to break this rule for a scene.”
A canon editor should then be able to answer, “Fine—here is what else changes.”
That is how a world stays coherent without becoming creatively frozen.
Sources
- Perseus Digital Library, Homer, Odyssey, Book 11 — https://beta.perseus.tufts.edu/urn%3Acts%3AgreekLit%3Atlg0012.tlg002.cotterill1911-eng2%3A11/ — accessed 2026-10-03
- Scaife/Perseus Atlas, Homer, Odyssey edition metadata — https://atlas.perseus.tufts.edu/nodes/urn%3Acts%3AgreekLit%3Atlg0012.tlg002.perseus-eng3%3A/ — accessed 2026-10-03
- British Museum, Egyptian death and afterlife: mummies — https://www.britishmuseum.org/collection/galleries/egyptian-death-and-afterlife-mummies — accessed 2026-10-03
- British Museum, What is the Book of the Dead? — https://www.britishmuseum.org/blog/what-book-dead — accessed 2026-10-03
- The Metropolitan Museum of Art, The Weighing of the Heart — https://www.metmuseum.org/de/perspectives/the-weighing-of-the-heart — accessed 2026-10-03