How Canon Accumulated for Forty Years
By Nelson McKeeby
Negentropy did not begin with a canon policy. It began with writing. In 1979 I was producing stories, game material, maps, rules, and research notes as quickly as I needed them. An adventure required a settlement, so I created one. A story raised a technological question, so I worked out an answer. A campaign needed geography, weapons, institutions, or history, and those things entered the project wherever there happened to be room for them. I was building a world because I wanted to use it, not because I expected every sentence written along the way to become a permanent declaration about what that world must always contain.
That distinction became important surprisingly early. A handwritten page could preserve an idea that remained central to Negentropy beside another idea abandoned after a few months. A typed description might contain a place name that survived for decades even though most of the surrounding explanation changed. Research could expose an assumption as wrong. Actual play could reveal that a rule or political arrangement did not work the way I expected. The archive therefore became something different from canon. The archive tells me what I wrote at a particular point in the development of the project. Canon tells me what the setting says now, after later work has tested, expanded, corrected, or replaced those earlier answers.
I value the archive because of that difference, not despite it. A pencil page from 1980 can establish that a name, institution, or idea already existed during Campaign 80 even when the modern version is much more elaborate. An old adventure can show that a settlement had already entered the world years before another sourcebook finally described its economy, politics, and history. The early material provides evidence about the development of Negentropy, but evidence about development is not automatically authority over the present setting. Some pages preserve experiments. Some record ideas that later research contradicted. Some contain drafting mistakes. They remain worth keeping because they show how the world grew.
This separation has become particularly important now that I am writing about the history of the project itself. I often need to say both what I believed in 1980 and what Negentropy says in 2026. Pretending those answers were always identical would make the development history less accurate. On the other hand, allowing every obsolete answer to remain active because it can still be found in an old binder would make the modern setting almost impossible to use. Negentropy therefore needs two kinds of memory: a historical archive that preserves where ideas came from and a current canon that tells the next writer, Game Master, or player which answer presently governs the world.
Canon Grew Through Use
The ideas that became most stable usually did so because they continued to be useful. A rule that survived several campaigns acquired weight. A settlement that players returned to accumulated relationships, incidents, and local history. A technological assumption that continued to survive research became increasingly difficult to replace casually because other parts of the setting had grown around it. Repeated use acted as a pressure test. Good ideas generated more material. Weak ideas eventually produced contradictions or awkward questions that forced another look.
That is why some elements created at the beginning remain recognizable after more than forty years while the explanations surrounding them have changed enormously. The Negentropy Cylinder still works as an important constraint because finite capacity kept producing interesting decisions. The larger explanation of the world beyond the Cylinder became much more sophisticated because campaigns kept asking questions that the original notes had never needed to answer. Continuity came from preserving the parts that continued to work, not from declaring the first draft untouchable.
Research became another source of pressure. I have never wanted an early sentence to survive merely because I happened to write it first. If I discover that the original explanation was poor, I would rather improve the world than spend years defending the mistake. XNA is one of the clearest examples. I originally used the term because I wanted some BEMs to possess alien heredity and lacked a serious scientific model for what that meant. Much later I encountered actual work on xeno nucleic acids. That research gave me a far stronger framework, so the modern Project now uses XNA as a developed biological classification rather than the vague concept that existed in the earliest material. The old use remains important to the history of the project, but the better explanation governs the current books.
This kind of revision does not necessarily destroy the identity of an old idea. Often it does the opposite. The creative problem that originally produced XNA remained the same: I needed a biological distinction between terrestrial heredity and genuinely alien systems. Better research improved the mechanism without removing the reason the concept existed. Negentropy has repeatedly developed this way. My knowledge changed, the world became more detailed, and an idea that once rested on a rough explanation acquired a much stronger foundation.
Campaigns played much the same role as research. Published and typed material were never the only things shaping the world. Players altered places. Settlements survived disasters or lost resources. Characters established relationships that later adventures returned to. Replaying the same material also exposed assumptions that had survived simply because the first group never asked the question that broke them. A second or third group often found that question immediately.
Not every event at one table became universal canon. A memorable incident remained part of that campaign unless later writing adopted it into the broader setting. That distinction allowed play to affect Negentropy without making every playthrough binding upon every future group. What mattered was whether a result continued to generate useful material and eventually entered the larger structure.
Reconciliation Became Necessary
For most of the project's history, information could develop in several places at once. Early material lived on paper and in the binders used during play. Computer storage came later and did not instantly absorb everything already written. One subject might therefore appear in an old adventure, a regional source, a military document, and a later city book before anyone sat down to compare all four. That distributed structure helped Negentropy grow because I could work on whatever part of the world the next story required. The price was deferred editorial work.
Eventually simply keeping every version was no longer enough. I began building master sources whose purpose was reconciliation. A master source does not place several contradictory answers beside one another and ask the reader to decide which one is correct. It preserves material that remains compatible with the current setting and allows later established work to replace earlier material where the two conflict. The world remains cumulative, but accumulation does not mean that every obsolete sentence must remain alive forever.
Some corrections are simple. South Fort is a good example. An alternate name appeared in older drafting material for the Imperial city at former Fort Madison. Later work established South Fort as the correct name. Once I recognized the other name as a drafting error, keeping both would not have added historical depth to the setting. It would merely have preserved the mistake. The correction therefore has to propagate through current canon. Finding the older form in an old document should tell us something about the archive, not reopen the naming question.
Other cases involve expansion rather than error. Porte-la-Terre existed before City on a Lonely Rock developed it in depth. The earlier version was useful when the city occupied only a small part of the setting, but the later book eventually supplied enough geography, institutions, economics, and local detail that the short sketch could no longer remain the controlling description. The early material still has historical value and may preserve compatible details. Authority simply shifts to the later treatment wherever the two differ.
This ability to replace a sketch with a developed treatment is essential for a world written over decades. An early adventure should be allowed to establish enough material for play without accidentally preventing anyone from ever examining the subject properly again. The first version can be provisional even when nobody bothered to label it provisional at the time.
Reconciliation can also require synthesis. Older descriptions of Imperial religion sometimes reduced it too far toward emperor worship. Later material on the Imperial Church established a Christian institution with a more complicated relationship to the Emperor, the Star, and political sacred symbolism. The useful part of the older observation did not have to disappear. It could instead become one tendency within a larger institution. Later work showed that the old description had noticed something real but applied it too broadly. That is one of the most productive forms of revision because it preserves an observation while placing it inside a better explanation.
Some Questions Should Remain Open
Reconciliation does not mean that every blank has to be filled. One of the harder lessons of maintaining a large setting has been learning that an unanswered question is sometimes more useful than an invented answer. A source can establish an institution without defining every office inside it. A historical event can be securely placed in the past without possessing an exact date. A culture can use a name whose origin nobody has yet explained.
The Crollyites provide a good example. Their political identity and function within the Coastal Empire are established, but the origin of the name has not been fixed. I could easily invent a founder named Crolly, a historical town, or an old institution that supposedly produced the term. Doing so in passing would answer the question, but it would not necessarily improve the material I was actually writing. Worse, it would remove the possibility of answering the question later in a source where the origin could have real historical consequences.
I now treat this as open canon rather than an unfinished mistake. Open canon establishes a boundary around what the project currently knows. It keeps an unrelated paragraph from casually deciding a question that deserves more careful attention. The missing answer remains available until a story, campaign, or sourcebook actually needs it.
Contradictions require the same kind of judgment. Some are simple drafting mistakes. Others occur because a later developed source has replaced an early sketch. Still others belong inside the fictional world because different societies remember the past differently. Those three situations should not be treated as the same problem.
The Chuhopi band count illustrates a source-level contradiction. An older dedicated source described ten bands, while later development established twelve. The current culture material therefore uses twelve. That does not mean every consequence of the change has automatically been solved. The additional identities still have to be reconciled with older naming and history. Canon can be definite about the number while remaining honest about the portions of the problem that still need work.
Historical disagreement inside the setting works differently. If two peoples preserve conflicting accounts of an ancient event, both accounts may belong in canon because the disagreement itself is part of history. The essential question is whether the contradiction exists among fictional witnesses or among my own source files. The first can make the world richer. The second usually means I still have editing to do.
Master Sources and a Distributed Canon
As the library expanded, master sources became maps of authority. They do not all perform the same job. A Coastal Empire master can reconcile political, geographic, and cultural material for that region. A history master needs to organize chronology and relationships among events without turning itself into a complete gazetteer. A military source can control details that a city book only mentions in passing. Authority therefore remains distributed, but it is distributed deliberately.
That is different from trying to build one enormous document containing every fact in Negentropy. A history source should explain how one condition produced another. A culture source should explain how people live. A creature book should describe organisms. A city book should be allowed to know its city in greater detail than an older continental overview. Master references exist partly to record how those specialized sources fit together and which one governs when earlier descriptions conflict.
Modern tools make this work much easier than it was when everything lived in binders, but they cannot replace editorial judgment. Software can find every occurrence of two different names. It cannot always tell me whether one is an error, whether both are legitimate local usage, or whether one belongs to an earlier stage in the project's history. It can reveal the conflict. Someone still has to understand the setting well enough to resolve it.
That is why I do not think of canon as a frozen body of text. Canon is a working model of the world. A good model should become more accurate when research, play, or later development reveals that one of its relationships was wrong. Version numbers on master documents acknowledge that directly. A source marked v0.4 can be authoritative for current work while also admitting that the process of reconciliation continues.
A World With a Controlled Memory
After more than forty years, Negentropy canon resembles accumulated sediment more than a sequence of clean editions. The oldest layers remain important because later ideas grew from them. They contain the origins of names, concepts, places, and assumptions that sometimes remain visible in the modern world. The upper layers control the present surface because they contain the benefit of everything learned since the earlier material was written.
That creates obligations in both directions. I cannot take a modern explanation and pretend that I already understood it during Campaign 80 merely because it now fits an idea created then. I also cannot allow every old sentence to control the present simply because the paper survived. Writing honestly about the development of Negentropy requires preserving the difference between what existed then and what the setting has become now.
The continuity of the project comes from something deeper than preserving every detail unchanged. Certain problems, images, limitations, cultures, places, and questions continued to produce stories, so they survived. Their explanations became more precise. Their relationships accumulated. Some early answers disappeared because later ones worked better. Others were absorbed into larger explanations. Some questions remain deliberately unanswered because the setting has not yet reached the point where they need answers.
That is what canon means to me now. It is not a museum case holding the first version of every idea. The archive performs that job. Canon is the current structure that allows those ideas to continue functioning together. Its purpose is practical: to give the next story, campaign, or sourcebook a reliable world from which to begin.
Negentropy is not finished, and after this much time I am not sure “finished” would even be desirable. What it has instead is a controlled memory. It can preserve where it came from, recognize where it was wrong, retain what continued to work, and use what has been learned since then to decide what comes next.