Counterfactual Assumption Test¶
Reasoning prompt — instantiates Implicit Assumption Surfacing
Asks what would change if a key assumption were false, reversed, or only locally true.
Counterfactual Assumption Test takes a single, already-named premise and deliberately breaks it, to see how much of the plan is leaning on it. It is not a hunt for hidden assumptions and it gathers no new data; it is a thought experiment applied to one premise you have already put on the table. You negate it ("suppose this is simply false"), reverse it ("suppose the opposite holds"), or localize it ("suppose it is true here but not there"), and then trace the consequences downstream. The point is to convert a premise that felt like solid ground into a dependency — something the outcome would visibly move if it gave way. Where an elicitation ritual asks what are we assuming?, this prompt starts from a stated assumption and asks what if we are wrong?
Example¶
A wildfire incident commander's afternoon plan rests on a quiet premise: the wind will keep blowing from the west, pushing the fire toward a natural rock break. Nobody has flagged it because it matches the morning's readings. The counterfactual test flips it: suppose the wind swings to the south by 15:00. Tracing that single change, the plan unravels in a specific way — crews staged on the south flank, currently in "safe" black, would be directly below the head of the fire; the containment line drawn for a westerly push becomes useless. That downstream trace is the influence path made visible. The reversed premise — a southerly wind — is the concrete alternative the plan should have a contingency for. And working it through surfaces the boundary condition nobody stated: the whole line holds only while the wind stays westerly and under about 20 mph. That is not a forecast; it is a stated dependency, and it changes where crews are allowed to stand.
How it works¶
- Pick one load-bearing premise. The test is only worth running on assumptions the outcome actually hinges on; pick a premise, not a detail.
- Apply a negation move. Full negation (false), directional reversal (opposite), or localization (true only within a boundary) — each exposes a different kind of fragility.
- Trace the consequences. Follow the flipped premise forward: which decisions change, which "safe" choices become exposed, which options reopen.
- Read off the boundary. The localization move names the conditions under which the original premise actually holds — its validity envelope.
The distinguishing discipline is that it operates on one given premise at a time, by imagination — it neither discovers the premise nor tests it against evidence. It is a fast generator of dependencies and boundary conditions, not a verdict.
Tuning parameters¶
- Negation mode — full-false vs. directional-reverse vs. boundary/local. Full negation stress-tests existence; reversal stress-tests direction; localization finds the edge. Choose by what kind of wrongness would hurt most.
- Trace depth — how many steps downstream you follow the flip. Shallow tracing is fast but misses second-order effects; deep tracing catches cascades but can spiral into speculation.
- Target selection — which premise you flip. Aim at the load-bearing few; flipping trivia manufactures busywork and false reassurance.
- Chaining — single flips vs. flipping two premises together. Joint flips catch interaction effects a one-at-a-time test hides, at the cost of combinatorial blow-up.
When it helps, and when it misleads¶
Its strength is speed and cheapness: with no meeting and no data it turns a comfortable premise into an explicit dependency and a boundary condition, and it does so in a way that naturally generates the contingencies worth pre-building. It is the reasoning move that answers "how much of this rests on that?"
Its central failure mode is an imagination ceiling: you can only flip the premises you can think to name, so a genuinely unexamined assumption stays invisible — the test sharpens known dependencies but cannot discover unknown ones. Plausibility bias compounds this: a flip that feels absurd ("the wind won't really reverse") gets dismissed exactly when it is most worth carrying, so the test can be run backwards — flip a premise, declare the flip unrealistic, and thereby bless the original. The guarding discipline is to treat every counterfactual as a hypothesis to watch or test, never as a verdict that the original is safe, and to borrow the consider-the-opposite habit deliberately rather than only when a flip feels comfortable.[n1]
How it implements the components¶
influence_path— tracing "what would change" is mapping how the assumption governs the downstream choices; the flip renders the dependency visible.alternative_assumption— the negated or reversed premise is the concrete rival the plan is held up against.validity_condition— the localization move ("true only under these conditions") states the boundary within which the original premise holds.
It does not name previously-unstated premises or route them to owners and monitoring (implicit_assumption, assumption_owner, update_or_documentation_rule) — that discovery-and-disposition work is premortem_assumption_probe; and it runs on imagination, not disconfirming data (evidence_test) — that is red_team_assumption_review.
Related¶
- Instantiates: Implicit Assumption Surfacing — this prompt supplies the influence-and-boundary analysis for a premise the surrounding archetype has already surfaced.
- Consumes: a named premise from reflective_interview or premortem_assumption_probe — it flips what those mechanisms surface.
- Sibling mechanisms: design_critique · model_assumptions_log · premortem_assumption_probe · red_team_assumption_review · reflective_interview · stakeholder_assumption_elicitation · assumption_audit
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: The method negates, reverses, or localizes one load-bearing premise and traces the resulting decisions, exposures, options, and validity boundary, so its operative form is counterfactual reasoning analysis.
Nearest alternative: Interface, Display & Cue — A prompt initiates the exercise, but inferential consequence tracing rather than the question's presentation is the defining mechanism.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Philosophy
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Logic and philosophical analysis cohered counterfactual variation of a premise to expose which conclusions depend on an implicit assumption.
Related originating lineages:
- Futurism & Strategic Foresight — Scenario practice supplies disciplined assumption variation and downstream consequence tracing.
- Psychology — Consider-the-opposite debiasing research operationalizes reversal of a premise before judgment.
Review resolution: Negating, reversing, or localizing a premise is fundamentally a philosophical counterfactual operation. Psychology and foresight independently turned that operation into practical debiasing and scenario methods; the entry synthesizes those uses into a single assumption-focused prompt.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- Stanford Encyclopedia of Philosophy: Counterfactuals
- PubMed: Considering the opposite—a corrective strategy for social judgment
- OECD Strategic Foresight Toolkit for Resilient Public Policy
Notes¶
The test's output is a dependency and a boundary, not a decision. Its natural consumer is a register that will hold the boundary condition and watch it; the flip tells you what to watch, model_assumptions_log is where the watch actually lives.
[n1] The "consider-the-opposite" strategy — deliberately asking why a favored judgment might be wrong before accepting it — is a documented debiasing technique in the judgment-and-decision literature; the counterfactual prompt is its assumption-focused form, which is why the guard against plausibility bias is built into the method rather than left to good intentions. ↩