The Piñata Protocol
Fascomachy · Document Zero
Three of the four instruments below can be run alone. That is what is wrong with three of them.
Evidence tier: Operational. Four tests the reader runs against their own argument. Passing them settles nothing about truth; it records what happened under the tests. Evidence Protocol · Theme Explorer
Bottom line up front
An argument cannot fully inspect the limits of the categories that produced it. The Piñata Protocol gives it four tests: wait, paraphrase, independent attack, and inventory. The third test matters most because the author controls the other three.
- Wait. Separate the claim from the momentum that made it feel complete.
- Paraphrase. Restate it without the vocabulary it invented for itself.
- Independent attack. Give it to someone who did not help build it and does not need it to survive.
- Inventory. Record what failed, changed, survived, and remains uncertain.
Passing the protocol does not make a claim true. It shows what happened under four stated tests. That narrower result is less impressive and more useful.
Why the method comes first
The rest of this series describes how judgment gets captured. It was written by someone holding an argument he wanted to be true. The protocol therefore comes before the claims it tests. A method invented after the conclusion can always be adjusted until the conclusion passes.
The four instruments were already present in the theory. Latency supplies delay. Fixed Terms requires stable language. Separation moves control outside the interested party. Records preserves what happened. The account described all four before applying them to itself. The equipment was available. The office had not completed the requisition.
The independent test produced the hardest result. Outside readers recovered Fixed Terms and Prohibition, while Symmetry failed to reproduce. Internal attack could improve the existing account. It could not reliably identify what the account had made unavailable to itself.
The four instruments
1 · Wait
Rule: Do not publish a claim on the day you write it. Date the draft. Before rereading it, reconstruct the reasons from memory. Then compare those reasons with the draft.
The test does not ask whether you still believe the claim. It asks whether the reasons remain the same after the feeling of arrival has cooled. If Wednesday needs a new argument to support Tuesday’s certainty, Tuesday supplied momentum rather than proof.
Limit: Delay can weaken excitement. It cannot expose a category the author never formed.
2 · Paraphrase
Rule: Rewrite the claim without terms coined by the argument. Keep the technical version beside the plain version and mark every difference in scope, certainty, and causation.
“Multiplicative across families,” “the reactive subject,” “eliminative,” and “single-covered” may be useful terms. They are not evidence. If the plain version becomes weaker, identify whether the technical language added precision or merely supplied authority by sounding established.
Limit: Paraphrase tests what the language is carrying. The author still chooses the paraphrase and can quietly return anything inconvenient to the terminology department.
3 · Independent attack
Rule: Give the claim, evidence, and definitions to a reader who did not help construct them. Do not tell the reader where you expect the weakness to be. Ask them to classify, contradict, and identify missing cases. Record disagreement before responding.
This is the only instrument that removes the key from the author. Seven attacks were run against Document Three; four landed. The two most productive came from outside the construction process. They removed one confident family, recovered two missing families, and did not require the author’s permission to notice any of it.
Limit: Independence is not magic. An outside reader brings other assumptions, incentives, and blind spots. Use more than one reader when the claim matters, and preserve disagreement instead of averaging it into calm.
4 · Inventory
Rule: Record the tested claim, date, tester, method, result, change made, and unresolved question. Put the record beside the argument rather than in a ceremonial appendix.
A failed claim and a surviving claim both produce information. The protocol did not create the press-release quotation, the missing family, or the disagreement over Symmetry. It made them available for correction.
Limit: The inventory contains failures the tests found. It does not contain failures the tests were unable to describe. A complete ledger of detected errors remains a ledger of detected errors.
Disposal rule: Keep a claim that survives the stated tests, but keep the test record with it. When a claim fails, keep the pieces long enough to identify what was wrong. Refusing the test produces no evidence at all.
Ash-Thumb
Fiction. The council, the vessel and its examiners are invented.
Inventory appended to the examination report.
Ash-Thumb broke the vessel because the council had asked what it contained. He said an intact vessel concealed its evidence. This was persuasive, particularly to members who had already purchased hammers.
The vessel was old, sealed, and painted with figures no living reader could name. Some believed it held seed. Some believed poison. A few believed it was empty and therefore sacred. Ash-Thumb proposed an independent attack: strike it from several directions, compare the fragments, and refuse the authority of the whole.
The first blow produced a crack. The council applauded the method. The second produced dust. The third produced a smell that caused three observers to leave. Their departure was recorded as resistance to scrutiny.
Inside were black grains, a strip of cloth, two small bones, and liquid that disappeared into the floor before anyone assigned to liquids arrived. Ash-Thumb ordered an inventory. The grains went to Agriculture, the cloth to Culture, the bones to Health, and the stain to Facilities. Each office received a fragment and a deadline.
Agriculture reported that the grains were not viable seed. Culture reported that the cloth was too damaged to establish provenance. Health reported that the bones were non-human, with confidence limited by breakage. Facilities removed the stained floorboard and disposed of it as a slip hazard.
The reports agreed on one point: no fragment proved the vessel’s alleged purpose. Ash-Thumb announced that the vessel had been epistemically disarmed.
An apprentice asked whether the contents might have changed when the vessel was broken. Ash-Thumb said the question confused method with object. The apprentice asked whether the arrangement of the objects had mattered. He was assigned to reassemble the vessel so the council could evaluate that possibility.
There were 214 pieces. The apprentice found that seven had been distributed as samples, 12 retained by Security, and one used by Ash-Thumb to prop open a window during the smell. The inventory listed all 20 as available upon request.
Weeks passed. The council needed a conclusion. Ash-Thumb wrote that the vessel had contained a heterogeneous symbolic-material assemblage whose prior claims could not survive independent examination. This was accurate. Nothing about the contents had survived independent examination.
The black grains later sprouted between the floorboards in a room no longer occupied because of the smell. Facilities treated the shoots with solvent. Agriculture was not notified because its report was complete.
At the public meeting, Ash-Thumb displayed the reconstructed vessel. Gaps remained where the samples had been. He said the gaps represented uncertainty and proved the honesty of the method.
Someone asked what the vessel had contained.
Ash-Thumb held up the inventory.
What the protocol cannot claim
Visible self-criticism can make an argument more persuasive. That gain cannot be used as evidence for the argument. Running the four instruments proves only that the instruments were run and produced the recorded results.
A concession also matters where it appears. A limit placed after the argument has done its work may function as decoration. Put the limit beside the claim it changes. State the reduced claim there.
Agreement is not a symptom. Agreement does not prove capture, and disagreement does not prove independence. A method that treats both responses as confirmation cannot fail and therefore cannot test anything.
The protocol tests claims, not people. Record a proposition, definition, inference, source, or procedure. Do not use the protocol to diagnose a person’s motives, character, or mental state.
Field exercises
- Wait. Date one important claim and leave it for a day. Reconstruct the reasons before rereading it. Record the differences.
- Paraphrase. Rewrite the claim without its coined terms. Mark every change in scope, certainty, or causation.
- Use an outside reader. Supply the claim, evidence, and definitions without identifying the expected weakness. Record the reader’s classification before responding.
- Inventory the result. Record the claim, date, tester, method, result, revision, and unresolved question beside the argument.
Nothing in this series becomes settled merely because it survived its author. The fourth step records a result. It does not close the yard.
Sources and further reading
- Feynman, R. (1974). “Cargo Cult Science,” Caltech commencement address. “The first principle is that you must not fool yourself — and you are the easiest person to fool.” The whole page is a footnote to that sentence.
- Pronin, E., Lin, D. & Ross, L. (2002). “The Bias Blind Spot.” Personality and Social Psychology Bulletin. People detect bias in others and exempt themselves, and the exemption grows with expertise.
- Mercier, H. & Sperber, D. (2011; 2017). The Enigma of Reason. Reasoning evolved to produce and evaluate justifications — which is why we are poor auditors of our own and competent auditors of everybody else’s. The argument for instrument three.
- Nosek, B. et al. Pre-registration and the distinction between prediction and postdiction — instrument one, institutionalized.
- Orwell, G. (1946). “Politics and the English Language.” On what happens to a claim that can only be said in its own dialect.