Tending the Canon

Creative culture has a rich vocabulary for vision. The auteur. The showrunner. The original creator. We celebrate the person who imagines the world into existence. We have an equally rich vocabulary for contribution. The writer. The worldbuilder. The player who brings something new to the table. And we have a grudging vocabulary for judgment: the editor who cuts, the critic who evaluates, the arbiter who settles disputes.

What we don't have is a vocabulary for maintenance.

There is no Oscar for continuity. No Hugo for "kept the wiki accurate." No credit that reads "maintained the reference document so that four other writers could work confidently in the same world for three years." And yet this is the work that determines whether a shared fictional world stays coherent over time or quietly degrades into a collection of private worlds that happen to share a name.

The governance framework I've been developing for the Transmission Zero universe identifies four functional roles that every shared world needs: the founder who establishes foundational commitments, the contributor who creates, the arbiter who resolves disputes, and the steward who maintains the record. Of the four, stewardship is the most labor-intensive, the least visible, and the most consequential when it's absent.

What stewardship actually involves

The steward's work is continuous and unglamorous. It includes:

Keeping the reference document current. As new creative work enters the world, the bible or wiki needs to reflect it. A reference document that's three months behind the actual state of the world is worse than having no reference at all, because contributors trust it and make decisions based on stale information.

Tracking dependencies. When a new story builds on an established species biology, that dependency should be noted somewhere. Not in a formal database necessarily, but at least as an annotation: "this story assumes the Species B joining biology from the earlier work." These notes are what make impact assessments possible when someone proposes a change.

Watching for hardening. Working canon that has been built upon repeatedly has effectively become established canon whether anyone designated it or not. The steward notices when a detail's dependency weight has shifted, and updates its designation accordingly.

Checking consistency. When new contributions arrive, someone needs to verify they don't contradict what's already established. This isn't about gatekeeping quality; it's about catching the soft contradictions and implication conflicts that individual contributors don't notice because they're focused on their own creative work. A story that casually mentions a technology might be implying something about the world's physics that conflicts with established canon. The steward is the person thinking about those second-order effects.

Auditing periodically. In a long-running project, contradictions accumulate despite everyone's best efforts. A quarterly review of the canon, focusing on recently added content and areas of active development, catches problems before they become entangled with months of dependent work.

None of this is creative in the way that writing a story or inventing a species is creative. It's infrastructure work: invisible when done well, catastrophic when absent.

The absent steward

Projects that fail at collaborative fiction governance almost always fail at stewardship. Not because nobody cared, but because the work is easy to defer. The creative work has energy and urgency: there's a story to write, a session to run, a deadline to meet. Stewardship tasks have no urgency until the problems they would have prevented become visible, by which point they're much harder to fix.

The pattern is recognizable. Creative work continues. The reference document falls behind. Contributors aren't sure whether the bible reflects the current state of the world or last quarter's. They start keeping their own notes, which diverge. Contradictions build up unnoticed. When they finally surface, they're entangled with months of dependent content and much harder to resolve than they would have been if caught early.

This is a familiar pattern in software as well. Documentation falls behind. Tests aren't updated. The shared understanding of the system diverges between team members. Nobody notices until someone makes a change that breaks something nobody knew was load-bearing. The cost of absent maintenance compounds over time in exactly the same way.

Stewardship as infrastructure

The parallel to infrastructure maintenance is worth dwelling on, because it explains both why stewardship is undervalued and why it's critical.

Infrastructure, in the sense that software engineers use the term, refers to the systems and practices that support the work without being the work itself. Testing, documentation, deployment pipelines, monitoring. These things don't ship features. They make it possible to ship features reliably. When they're working, nobody thinks about them. When they fail, everything breaks.

Stewardship in a shared fictional world is the same kind of work. The steward doesn't write stories. The steward makes it possible for stories to be written confidently within a shared world. The reference document, the dependency tracking, the consistency checks, the canonical designations: these are the infrastructure that lets contributors focus on creative work instead of constantly worrying about what they might be contradicting.

And like infrastructure in software, stewardship has a deferred-maintenance problem. Skipping a consistency check today saves time today. The cost shows up later, as a contradiction that's harder to fix, a dependency that wasn't tracked, a contributor who makes a decision based on stale information. Each individual skip is rational. The accumulation is corrosive.

Who does the work

In a solo project, the author is the steward. This is both the simplest and the most dangerous configuration, because there's no one else to notice when stewardship falls behind. The solution isn't to stop being a solo author. It's to recognize stewardship as a distinct mode of work and protect time for it. The version of you maintaining the reference document is doing different work than the version of you writing the next story, even though you're the same person. Naming the role makes the work visible.

In a small group, stewardship often falls to whoever naturally gravitates toward organization and detail. The person who maintains the shared document. The person who notices inconsistencies without being asked. This is valuable, and it should be acknowledged explicitly rather than assumed. Assumed roles produce "I thought you were handling that" failures.

In a larger project, stewardship may need to be a team rather than a person, or a rotating responsibility with defined duties. The implementation guide for the framework includes warning signs that the steward is overwhelmed: the reference document falling behind, contributors disagreeing about what's been established, inconsistencies being discovered only after significant dependent work has been built.

AI as stewardship aid

This is where the AI angle enters, and it enters naturally rather than as a pitch. The three documents that make up the constitutional worldbuilding framework (the governance framework, the implementation guide, and the Transmission Zero constitution) were all written in direct collaboration with AI. The framework itself is tool-agnostic; it works just as well with index cards and conversation as with a language model. But in practice, AI turned out to be particularly suited to certain stewardship tasks.

The Transmission Zero constitution describes the AI collaborator's role directly: "a contributor and an advisory input to stewardship (identifying inconsistencies, tracing dependencies, pressure-testing proposals), but does not hold governance authority. Decisions rest with the author."

That framing, advisory input to stewardship, captures something important. The interesting role for AI in shared-world governance isn't "co-author" in the creative sense. It's the role that humans consistently underperform at: tracking what depends on what, noticing when a new detail conflicts with something established three documents ago, maintaining the reference that human memory can't reliably hold.

A language model that has read the entire canon can do a consistency check more thoroughly than a human who last reviewed that section months ago. It can trace dependency chains across documents. It can flag when a new creative decision touches root-level content that constrains other projects. These are exactly the stewardship tasks that human contributors find tedious and that tend to be deferred because they lack urgency.

The constraint that matters is that the AI doesn't hold governance authority. It doesn't decide whether a contradiction should be resolved in favor of one version or another. It doesn't designate canonical levels. It doesn't arbitrate disputes. Those are human decisions that require creative judgment, understanding of the project's goals, and the authority that comes from being a participant in the collaborative agreement. The AI is a tool that makes stewardship feasible at a level of thoroughness that a human steward alone would struggle to sustain.

This is a more modest claim than "AI as creative partner," but it may be a more durable one. Creative AI collaboration raises difficult questions about authorship, originality, and artistic identity. AI-assisted stewardship raises none of those questions. It's the same as using a spell-checker or a version control system: a tool that handles the parts of the work that benefit from machine precision, freeing the human to focus on the parts that require human judgment.

The stewardship disposition

What makes someone a good steward isn't technical skill. It's a disposition: the willingness to maintain things rather than create them, to care about coherence rather than novelty, to do work that nobody will praise because its absence would be noticed far more than its presence.

Not everyone has this disposition, and that's fine. The value of naming stewardship as a distinct role is precisely that it can be assigned to someone who values it, rather than expected from everyone equally. Some contributors are brilliant at creating new content and terrible at maintaining the record. Some are the reverse. A governance framework that names the roles lets people fill the roles that match their strengths.

The deepest contribution of the stewardship concept isn't organizational. It's a reframe. We tend to think of creative worlds as things that are built, with the emphasis on construction: inventing, developing, expanding. But a shared world also needs to be maintained. The biology established in the first story needs to still be accurate in the twentieth. The dependency chains need to remain visible. The reference document needs to track what's settled, what's in flux, and what's been superseded.

Building is glamorous. Maintaining is essential. A governance framework that makes stewardship visible is a framework that takes coherence seriously, not as a pleasant side effect of good intentions, but as a practice that requires sustained, deliberate attention.