Counterexample Probe¶
Test probe — instantiates Informal Fallacy Diagnosis and Repair
Attacks a general claim by manufacturing the case that would make it false, then reports whether the claim survives, narrows, or breaks.
A sweeping claim is easy to assert and hard to earn. Counterexample Probe stress-tests a general or universal claim by deliberately hunting for the single instance that would make it false — the customer, the sample, the edge case that the claim says cannot exist. It does not evaluate the claim's plausibility or the goodness of its supporting evidence; it goes straight for the falsifier. Its defining discipline is generativity: rather than waiting for a counterexample to surface, it manufactures candidates on purpose, registers each one, and then routes the claim according to what the register shows — kept as stated, narrowed to a defensible scope, or rejected. This is the mechanism of active falsification, aimed at claims of the form "all," "always," "never," or "every."
Example¶
A founder reviewing quarterly numbers asserts to the team: "Every user who churned this quarter left because of price — so if we cut the plan cost, retention fixes itself." Counterexample Probe takes the universal at its word and goes looking for the exception. It pulls exit interviews and support logs and asks, deliberately, which churned account left for a reason that is not price? Within an hour it has registered three: one enterprise account that left because a required integration was missing, one that cited a data-export limitation, and one that simply went out of business.
Those three are enough. The register shows the universal is false as stated, and the probe routes accordingly: the claim cannot stand at "every," and the price-cut remedy that rode on it is not warranted. The repair is to narrow — "price is the leading, but not the only, churn driver" — which changes the decision from "cut the price" to "cut the price and close the integration gap." The probe's output is that routing recommendation plus the registered cases that forced it, not a claim that price is irrelevant.
How it works¶
- Read the claim as a quantified target. Restate the assertion in the form it commits to — "all X are Y," "this never happens" — so it is clear exactly what a single counterexample would refute.
- Generate candidate falsifiers on purpose. Search systematically for the instance the claim forbids: query the data for the excluded case, construct a hypothetical that satisfies the antecedent but not the consequent, recall a known exception.
- Register the survivors. Record each genuine counterexample that holds up under scrutiny — and note the near-misses that turned out not to qualify, since they mark the claim's real edge.
- Route by the register. If no counterexample survives, the claim is corroborated (not proven); if one does, route to narrowing the scope or rejecting the inference that depended on the universal.
Tuning parameters¶
- Search aggressiveness — how hard you hunt for the falsifier. Turned up, it finds subtle exceptions but can spend effort attacking claims that were never meant as strict universals.
- Counterexample admissibility — how strict the bar is for a candidate to count. Loose admission lets irrelevant or far-fetched cases "refute" a claim; strict admission risks waving away real exceptions as special cases.
- Repair vs reject threshold — how many or how central a counterexample must be before you narrow the claim rather than reject the whole inference. A low threshold preserves salvageable claims; a high one tolerates a leaky generalization.
- Register depth — whether you log one decisive counterexample or the full set. The full set maps where the claim actually holds, at the cost of more work.
When it helps, and when it misleads¶
Its strength is asymmetric leverage: a single well-chosen counterexample can settle a universal that no amount of confirming instances could establish, which is the core insight of falsification as a test of claims.[n1] It is the fastest way to puncture an overreaching generalization and, through its routing, to convert a broken universal into a narrower claim that is actually true.
Its failure mode is mistaking the scope of a claim. A statistical tendency ("most," "usually") is not refuted by one exception, and treating it as if it were is itself a reasoning error — the probe can manufacture a technically-true edge case and use it to dismiss a claim that was only ever meant to hold on average. The classic misuse is the nitpick: producing a contrived or negligible counterexample to reject a claim wholesale rather than to refine it. The guarding discipline is to first fix whether the claim is genuinely universal or merely general, and to route a real-but-marginal counterexample toward narrowing the claim rather than toward triumphantly rejecting it.
How it implements the components¶
counterexample_register— its central artifact: the logged set of manufactured cases that would falsify the claim, each tested for whether it genuinely qualifies.repair_or_rejection_route— it converts the register into an action, routing the claim to survive, narrow, or be rejected depending on what the counterexamples show.
It does not check whether a term keeps one fixed meaning across the argument — term_and_category_stability_check belongs to Category Boundary Stability Check. This probe manufactures cases that would make the conclusion false; the stability check instead watches whether a definition is redrawn to dodge them.
Related¶
- Instantiates: Informal Fallacy Diagnosis and Repair — supplies the active falsification-by-counterexample step within the wider diagnosis.
- Sibling mechanisms: Category Boundary Stability Check · Steelman then Refute Protocol · Premise Relevance & Acceptability Test · Fallacy Triage Matrix
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Counterexample Probe operates as a bounded trial, probe, simulation, or rehearsal that generates evidence from performance because it attacks a general claim by manufacturing the case that would make it false, then reports whether the claim survives, narrows, or breaks.
Independent corroboration: The frozen evidence defines Counterexample Probe as 'Attacks a general claim by manufacturing the case that would make it false, then reports whether the claim survives, narrows, or breaks', so its operative form is Experiment, Test & Rehearsal.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Philosophy
Origin pattern: Single lineage
Present-day reach: Universal
Rationale: Logic and philosophy cohered counterexamples as decisive tests of universal claims and as reasons to reject or narrow overgeneralization.
Related originating lineages:
- Rhetoric — Argumentation practice operationalized constructed counterinstances for diagnosing and repairing informal overgeneralization.
Review resolution: A direct request for a case that defeats a universal claim is a portable philosophical reasoning method; rhetoric is a genuine alternate practice but does not limit the reach.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] In Karl Popper's account, a universal claim can never be conclusively verified by confirming instances but can be decisively refuted by a single genuine counterexample. Counterexample Probe operationalizes that asymmetry, hunting the falsifier rather than piling up confirmations. ↩