Definition #
[Default Gravity] is the pull that installed artifacts and inherited beliefs exert on a surface the operator has already ruled, so the operation executes the default instead of the ruling and produces an outcome the operator never chose but signs his name to.
The condition requires a ruling to exist. This is not a surface the operator never engaged, and it is not a surface he deliberately left loose. He engaged it, decided it, said it out loud, and the checklist in somebody’s hand at 7:40 on a Friday said something else. The artifact executed. The ruling did not.
The pull has two forms. The artifact form is a document, template, procedure, position description, schedule shape, or system configuration that contradicts the ruling and is physically present at the moment of production. The belief form is an installed assumption nobody examined — this is what a kitchen manager does, cast cannot be trusted with comps, labor is a percentage, turnover is the cost of the business — which executes with no document to point at.
Mechanism #
The artifact beats the decision because the artifact is present. The ruling lives in the operator’s head, or in a pre-shift three weeks ago. The artifact is in a cast member’s hand during the rush. When the two disagree, the operation runs the instruction that is there at the moment of execution, every time, and no announcement changes that. The default does not argue with the ruling. It overrides it quietly and produces its own outcome under the operator’s name.
Nothing in the building reports the override. No alarm goes off when a ruling fails to execute. The operator sees the result, not the substitution, so what reaches him is an outcome that disagrees with what he decided and no visible reason for the disagreement.
So he draws the wrong conclusion. This is the expensive part. He decided something, it did not take, and he concludes his decision was wrong, or his people will not execute, or the operation is not ready for it. None of those are true. The decision was never executed. It was contradicted by the checklist, the position description, the schedule template, and a belief nobody examined.
Then reverting manufactures evidence against designing. He reverts, the revert confirms his read that designing does not work in a real restaurant, and the appetite for the next purchased program comes back stronger than before. The loop closes on itself. Default artifacts do not only block rulings, they produce the case against ever ruling anything again, which is what makes this the mechanism that keeps operators buying.
The belief form is worse than the artifact form. A borrowed document can at least be found, read, and rebuilt. A borrowed belief executes with no trace, cannot be audited, and cannot be pointed at. Nobody can name the file that overrode the decision, because there is no file. The belief is carried in the person who trains, so it reinstalls itself with every new cast member.
Gravity, not resistance. Nobody in the building is fighting the ruling. The cast is doing exactly what the operation told them to do, in writing, and the writing came from outside. That is why the operator’s instinct to read this as a people problem is always wrong, and why more insistence, more meetings, and more accountability never fix it. The pull is structural and constant, and it is strongest on the surfaces nobody looks at.
The rule that falls out. An unsupported ruling is an announcement, not a design. Ruling a surface and leaving the artifacts and beliefs that contradict it in place is the same as not ruling it, except that it costs the operator his standing with the cast the third time it happens. So every ruling carries an obligation: name every artifact and every belief that touches the surface it governs, and rebuild each one on contact. Not audit it, not edit it. Rebuild it from the ruling, because an edited default is still a default.
Load-Bearing Distinction #
Not a surface never ruled. A dimension the operator never engaged is plain default, and the correction is to rule it. Here the ruling exists, which is why the operator’s evidence is so misleading.
Not deliberate looseness. A surface ruled to stay flexible on purpose is a ruling. This is a ruling that was overridden.
Not [Default Architecture]. That is the standing condition of a building assembled out of parts nobody in it ruled. This is the specific force that condition exerts against a new ruling, which is why an operator who starts designing feels resistance he cannot locate.
Not [Default Perspective]. That is a closed loop at the Perspective layer, where the operator examines rigorously and the installed perspective interprets every finding. Different object. This is an artifact overriding a decision at the point of production.
Not [The ReRead]. The ReRead is how the effect gets detected, after the fact, by reading produced result against designed standard. It is the instrument, not the condition.
Not [Constraint Inheritance]. That is a step whose original reason is gone and cannot be sourced. This is a live default actively contradicting a current ruling, and its origin is usually easy to find once anybody looks.
Not a compliance problem. Compliance means people know the standard and are not doing it. Here they are doing the standard, and the standard in their hand is not the operator’s.
The term is load-bearing because without it the operator has no way to explain a failed ruling except by blaming himself or his people, and both of those conclusions push him back toward buying. Naming the pull relocates the failure to the artifact layer, where it can actually be fixed.
Diagnostic Tests #
Test One — The Hand Test. Take any ruling made in the last two periods and find what is physically in the hand of the person executing it during a shift. Checklist, screen, template, printed procedure. Read it against the ruling. If it does not say what the ruling says, the ruling is not running.
Test Two — The Failed Ruling Inventory. List every decision made in the last year that did not take. For each, ask which artifact or belief was still in place on that surface when the decision was announced. A list where most items have an artifact behind them is the diagnostic.
Test Three — The Authority Read. Pull the position descriptions and find the authority limits. Compare them to what the ruling requires a cast member to be able to do at the table. A limit written by a stranger overrides a ruling made by the operator, every shift.
Test Four — The Sentence Test. Ask a lead why a step is done the way it is. If the answer is a belief stated as a fact about the business rather than a reason about this operation, that is the belief form, and it is executing right now.
Test Five — The Revert History. Ask the operator what he has tried and given up on. Each abandoned attempt is a candidate, and his stated reason for abandoning it is almost always a conclusion about his people rather than a read on the artifact layer.
Test Six — The Training Test. Watch a new cast member get trained on the ruled surface. Whatever the trainer says is what the operation actually runs, and the distance between that and the ruling is the size of the pull.
Family Position #
Child of [Summers Principle], on the default side of the law. Sibling to [Outcomes Formula], which reads outcomes back to thinking, and to [Point Of Experience], where the override resolves in front of the Guest. [Sameness Machine] supplies the defaults doing the pulling. [The ReRead] is the detection discipline. Runs on every Fundamental, because every Fundamental has artifacts and beliefs sitting on it.
Fundamentals Coverage.
Perspective read. The pull produces a false read: my decisions do not take, so designing does not work here. The operator’s own evidence argues against the discipline that would have fixed it, which is the most expensive misread in the business and the one the industry most reliably funds. Detection is the revert history, read for whether the stated reason is ever an artifact. The response is a standing rule that no failed ruling gets interpreted until the artifact layer on that surface has been read.
Product read. The Product the Guest receives is the artifact’s output, not the ruling’s. The operator decides what this operation produces at the moment of contact, and a borrowed procedure, a borrowed trust posture, and a borrowed position description decide it instead, so the GX carries a stranger’s assumptions about a restaurant that is not this one. Detection is walking the Guest’s path against the ruling and finding where the produced result diverges. The response is rebuilding the governing artifact from the ruling rather than editing it.
People read. The cast is caught between an announced standard and a written one, and they run the written one, which is correct behavior. What shows up is people who look unwilling and are actually compliant, and an operator losing standing every time he announces something that does not hold. The belief form lives here permanently, because it reinstalls through whoever trains. Detection is watching a training session and listening for beliefs stated as facts about the business. The response is that authority limits and training content get rebuilt at the moment of the ruling, not after the ruling fails.
Performance read. Execution is where the substitution happens, in the moment, under load, with no time to consult anything. Whatever is present wins, so the shift is the honest record of what the operation actually runs. The recognizable moment is a cast executing cleanly and producing an outcome the operator did not choose. Detection is the hand test during a rush rather than a review afterward. The response is that the ruled surface gets its artifact replaced before the next shift, because every shift run on the default trains it deeper.
Profit read. The money cost is the ruling that never produced its return plus the next program bought to replace it. The operator pays for the decision, gets none of it, then pays again for the purchase his own false conclusion made attractive, which is how the pull funds [Sameness Machine] directly. It never appears as a loss anywhere, because nothing failed on paper. Detection is pricing the abandoned rulings against what was spent trying them and what was spent afterward. The response is that no ruling gets funded without the artifact rebuild funded alongside it, since a ruling without it is a purchase of nothing.
Cross-References To Locked IP #
Parent:
- [Summers Principle] — the law this is the default-side mechanism of
Related:
- [Default Architecture] — the standing condition that supplies the pull
- [Designed Architecture] — the structure that closes the gap by rebuilding the artifacts on contact
- [Point Of Experience] — where the override resolves in front of the Guest
- [Outcomes Formula] — reads the produced outcome back to the thinking that made it
- [The ReRead] — the detection instrument for the gap between designed standard and produced result
- [Sameness Machine] — the supply that installs the defaults doing the pulling
- [Authority To Execute] — the specific limit a borrowed position description overrides
- [Question Dumb Shit] — the instrument that surfaces the contradicting artifact
- [By Design Or By Default] — the principle the override violates every shift it runs
Opposing patterns:
- [Repairman Syndrome] — the posture that reads the failed ruling as a people problem to be fixed
- [Novelty Arbitrage] — what the false conclusion sends the operator back out to buy
- [Transactional Mediocrity] — the produced result when every ruling is running on defaults
- [Static Decline] — the condition the reverted rulings accumulate into
Why This Matters #
The most demoralized operators I work with are not the ones who never decided anything. They are the ones who decided, watched it not take, and concluded they are not capable of running their own operation. They are usually the best readers in the room, and their evidence is real. Their conclusion is wrong.
This is the mechanism that keeps the buying loop alive. The industry does not need to talk an operator out of designing his own operation. It only needs the defaults to stay in place long enough for his own rulings to fail, because the failure does the selling. Every abandoned attempt raises the appetite for the next program, and the next program installs more defaults.
It also names the obligation that comes with every ruling, which is the part operators skip and the part that decides whether any of this work pays. A ruling is not a decision until every artifact and every belief on that surface has been rebuilt from it. Anything short of that is an announcement, and announcements cost standing.
And it is why the rebuild-on-contact rule is not a preference. An edited default is still a default, because the thing that made it wrong was never a line in it. It was where it came from.
Operating Consequence #
No ruling ships without its artifact list. Every ruling names the documents, templates, configurations, and beliefs touching that surface before it is announced, and each one gets rebuilt from the ruling.
Rebuild, do not edit. An edited default carries its origin. The artifact gets built from the ruling or it does not get used.
Stop interpreting failed rulings as people problems. The first read on any decision that did not take is the artifact layer, not the cast.
Read what is in the hand, not what was said. The operating standard is whatever is physically present at execution. Everything else is intent.
Treat beliefs as artifacts. Any assumption stated as a fact about the business gets named out loud and ruled, because an unexamined belief executes with no trace.
Refuse the announcement. The operator stops announcing rulings he has not yet supported, because the third unsupported announcement costs him the cast’s belief that his decisions hold.
Read every revert as a finding. An abandoned attempt is evidence about the artifact layer, not about the operator’s capability or his people.
What Changes Tomorrow #
Name one ruling made in the last year that did not take. Not the biggest one. The one the operator remembers being annoyed about.
Then find every artifact touching that surface and put them on the table physically. The checklist, the position description, the training outline, the schedule template, the system configuration, the printed procedure. Read each one against the ruling and mark every place it says something different. Then ask the person who runs that surface what they were told, and write that down too, because the belief form will not be on the table.
The count that comes back is the answer to why the decision failed. In almost every case the artifacts contradicted the ruling on the specific point that did not hold, and nobody in the building was ever asked to choose between them, because nobody knew there was a conflict.
Rebuild the one artifact that contradicts the ruling most directly, from the ruling, before the next shift. Then run the same ruling again and watch whether it holds. If it holds, the operator has his answer about whether his decisions work, and the next surface gets the same pass. If it still does not hold, there is another artifact or a belief still in place, and the pass runs again until the surface is clear. The principle the operator now runs is that a ruling is not made when it is decided. It is made when every default on that surface has been rebuilt from it.
