Counterfactual Minimal-Change Checklist¶
Checklist — instantiates Necessity-Possibility-Contingency Framing
A checklist for preserving relevant facts when using nearby alternatives to test necessity or contingency.
Counterfactual Minimal-Change Checklist governs the construction of a single nearby alternative used to test whether some factor was necessary: change only the factor under test, hold everything else relevant fixed, and check — item by item — that no confound rode along. Its defining move is discipline over one counterfactual rather than comparison across many. When you say "if X hadn't happened, the outcome would have been the same," the whole inference rests on the counterfactual world being minimally different; if your imagined world quietly rewrote three other things, the test is worthless. The checklist is the short list of questions that keeps a counterfactual honest before anyone leans on it.
Example¶
An economic historian wants to test the claim that a 1994 tariff cut caused a regional manufacturing recovery. The tempting counterfactual is "imagine the tariff had stayed in place." But a lazily imagined no-cut world tends to differ in a dozen ways at once. The historian runs the checklist against the constructed counterfactual: hold the exchange rate path fixed (don't let the imagined world also assume a stronger currency); hold the commodity-price cycle fixed; hold the neighboring region's policy fixed; change only the tariff.
Item four catches the error. The intuitive "no tariff cut" scenario had also silently assumed the incumbent government lost the next election — a second change smuggled in because it felt narratively natural. Stripping that back out, the historian finds the recovery still partly appears in the minimally-changed world, which means the tariff was contingent-contributory, not necessary. The checklist did not settle the causal question; it stopped a sloppy counterfactual from settling it the wrong way.
How it works¶
The checklist runs against one already-drafted counterfactual and asks, in order: what is the single factor being varied; what facts are relevant and must be held; did the drafted world change anything beyond the target; and does any held fact now sit in tension with the change. The distinguishing constraint is minimality — the counterfactual must be the closest world in which the target factor differs, not a comfortable narrative that differs in whatever ways make the conclusion easy. It disciplines one nearby case; it does not enumerate or grade a whole field of cases, and it does not hunt adversarial defeaters — its confound check is an informal self-check, not a defeater register.
Tuning parameters¶
- Relevance boundary — what counts as a fact that must be held versus an irrelevant detail. Draw it too wide and no counterfactual is ever "close enough"; too narrow and confounds slip through.
- Minimality strictness — how aggressively to reject a counterfactual that changes more than the target. Strict minimality is cleaner but sometimes unbuildable; loose minimality risks smuggled confounds.
- Single-versus-few varied factors — whether the checklist permits varying a tightly-coupled pair. Two coupled changes are sometimes unavoidable but make attribution harder.
- Backtracking tolerance — whether the counterfactual may revise the past to accommodate the change, or must keep history fixed up to the intervention point.
When it helps, and when it misleads¶
Its strength is protecting counterfactual reasoning from its most common corruption: a counterfactual that drifts into an unrelated scenario and then "proves" whatever the drift was chosen to prove. Keeping the alternative minimal and nearby is exactly the closest-possible-world discipline that makes counterfactual claims evaluable at all.[n1]
Its failure mode is that "minimal" is itself a judgment, and a motivated reasoner can define the relevant-facts set to hold fixed precisely the thing that would flip the result. It also cannot rescue a counterfactual that is ill-posed — some interventions cannot be made minimal because the factor is entangled with everything downstream. The classic misuse is the self-serving hold list: freezing every fact that helps and varying only what hurts. The guarding discipline is to fix the relevant-facts set before seeing which counterfactual the argument wants, and to state the minimality claim explicitly so a reader can contest it.
How it implements the components¶
minimal_change_guardrail— its central purpose: the item-by-item rule that only the target factor may change and every relevant fact must be held.possible_world_or_case_set— it constructs the one nearby counterfactual case the test runs on, checked for closeness.assumption_context_frame— the relevant-facts hold list is a scoped assumption frame recording what stays fixed across the change.
It does not read dependencies across a field of cases — contingency_dependency_map is Possible-World Case Matrix's, which compares many worlds where this checklist perfects one; and it does not systematically hunt defeaters — defeater_and_boundary_case_register belongs to Necessity-Possibility Red Team.
Related¶
- Instantiates: Necessity-Possibility-Contingency Framing — the checklist keeps a single counterfactual minimal so necessity and contingency tests stay valid.
- Sibling mechanisms: Modal Claim Table · Possible-World Case Matrix · Assumption Lock-and-Relax Workshop · Necessity-Possibility Red Team · Legal Obligation-Permission Matrix · Design Constraint-Relaxation Table · Modal Language Review Protocol
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: The checklist evaluates an already-drafted counterfactual for a single changed factor, preserved relevant facts, unintended changes, and internal tension, producing a bounded minimality finding.
Nearest alternative: Representation, Specification & Plan — The checklist externalizes criteria, but the mechanism depends on applying them to assess a specific counterfactual rather than maintaining a reference list.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Philosophy
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Counterfactual logic cohered the minimal-change discipline of varying the antecedent while preserving all independent facts when testing necessity or contingency.
Review resolution: Both reviewers identify closest-world and ceteris-paribus semantics as the origin. The checklist expresses an established philosophical discipline across many uses; broad reach does not warrant a universal claim or encyclopedia-synthesis label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] On David Lewis's semantics, a counterfactual "if A had been the case, C would have been" is true when C holds in the closest possible worlds where A holds — closeness measured by overall similarity. The whole apparatus fails if the imagined A-world is needlessly far from actuality, which is precisely the drift this checklist is built to catch. ↩