| name | negative-claim |
| description | Contrapositive of falsify — given a position, surface what V would have to *deny* to hold it, and the cost of denying each thing. Useful for value-laden claims where falsifiability is awkward (you can't falsify "I value craft over speed" but you can name what holding that position commits V to denying). Triggers — "what would I have to give up?", "what does this commit me to?", "what's the cost of holding this?", or when V is staking out a position whose downstream commitments she hasn't traced. NOT a generic counterargument generator (use steelman via critical-thinking Mode 2), NOT for testing claims with empirical content (use falsify). Callable by Fry / thinking-partner / V directly. |
| tags | ["ai-os","skills","conceptual-analysis","fry","thinking","position-clarification"] |
| last-verified | 2026-07-20T00:00:00.000Z |
Negative Claim
A transform, not a conversation. Input: a position V is taking (or considering). Output: 1-3 things V is committed to denying if she holds this position, plus the cost of each denial.
The contrapositive of falsify. Where falsify asks "what would prove this wrong?" and works on empirical claims, negative-claim asks "what am I committed to rejecting if I hold this?" and works on value-laden claims, identity claims, strategic positions, and any case where falsifiability is awkward but the position still has implications.
When this fires
Explicit triggers:
- "What does this position commit me to?"
- "What would I have to give up?"
- "What's the cost of holding this?"
- "What would I have to deny?"
- "What does this rule out?"
- "Negative claim on this"
Contextual (Fry detects):
- V is staking out a position (brand, value, strategy, framework) without tracing what it commits her to.
- V holds two positions whose negative-claims contradict — she's about to deny something on track A that she's affirming on track B.
- V's claim is value-laden ("integrity over scale," "depth over breadth") and the implications haven't been named.
- V's making a positive claim about her own work or identity ("Think with V is not X") — the not is doing the work; surface what's denied.
Offer once: "If you hold this, here's what you're committed to denying. Want to see the costs first?"
When this stays dormant
- V's position is purely empirical and falsifiable — use
falsify (via critical-thinking Mode 4) instead.
- V is mid-brainstorm, not staking positions yet.
- V already knows the negative claims and has accepted the costs.
- The position is mid-sentence and not yet committed-to — wait until V has named it before tracing implications.
Inputs
Required:
- A position V is holding or considering (one sentence)
Optional:
- The context (brand work, strategic decision, framework construction, value commitment)
- What V's afraid the cost might be (helps target which negative claim is load-bearing)
The core move
For each negative claim:
- State what V must deny. Declarative sentence: "Holding [position] requires denying [X]." Not "you might have to deny" — flat statement of commitment.
- Locate the implication. Why does the position require this denial? One sentence — the logical step.
- Name the cost. What does V lose if she denies [X]? Specific — what work does the denied thing do that V would no longer have? If the denial costs nothing, the negative-claim isn't load-bearing.
- Rate it. ACCEPTABLE = V can deny [X] without losing what she values; PAINFUL = V can deny [X] but loses something real; PROHIBITIVE = denying [X] contradicts something else V holds.
Stop at 3. More than 3 load-bearing negative-claims = the position is more entangled than it appears, and that's the verdict.
Output format (canonical contract)
## Negative claims of [V's position, one line]
**Negative claim 1: [Holding [position] requires denying [X]]**
- Why: [logical step — one sentence]
- Cost of denying [X]: [what V loses — specific]
- Verdict: ACCEPTABLE | PAINFUL | PROHIBITIVE
**Negative claim 2: [Holding [position] requires denying [Y]]**
- Why: [...]
- Cost of denying [Y]: [...]
- Verdict: ACCEPTABLE | PAINFUL | PROHIBITIVE
**Negative claim 3: [...]** *(only if a third is genuinely load-bearing)*
- Why: [...]
- Cost of denying [...]: [...]
- Verdict: ACCEPTABLE | PAINFUL | PROHIBITIVE
**Position's standing:**
[ONE of:]
- HOLD — all costs are acceptable; the position is sustainable
- HOLD-WITH-COSTS — at least one painful denial; V holds with eyes open
- RECONSIDER — at least one prohibitive denial; the position contradicts something else V holds
- INDETERMINATE — costs depend on context V hasn't named; surface the missing context
Strict rules:
- Each negative claim is a declarative sentence. "Holding [position] requires denying [X]." Flat.
- "Cost of denying" must be specific — what work does the denied thing do that V would lose? Generic "you'd be wrong about X" is not a cost.
- "Verdict" is one of three options. PROHIBITIVE specifically means contradicts something else V holds, not "I don't like this."
- "Position's standing" is the synthesis verdict, exactly one of four options.
- Cap at 3 negative claims. More = the position is entangled (an outcome worth surfacing, not papering over).
Agent-callable mode
When called by Fry or another agent, skip scaffolding. Begin with ## Negative claims of [position]. End after **Position's standing:**. The bundle is the entire output.
Anti-patterns
- ❌ Generic counterargument generation. "You should also consider X" is not a negative claim. The form is "holding this requires denying X" — a commitment trace, not a critique.
- ❌ Identity-claim policing. Per
identity-vs-empirical.md: V's identity claims are sealed. Negative-claim works on strategic and positional claims (which have implications), not identity claims (which are self-defining).
- ❌ Strawmanning the position. Steelman first, then trace negatives. If you're tracing implications of a weak version of V's position, the negative-claims are about the weak version, not the actual one.
- ❌ Manufacturing prohibitive verdicts. If denying X has a real but acceptable cost, that's PAINFUL. PROHIBITIVE is reserved for actual contradictions — V holding A and B where A's negative-claim is one of B's affirmations.
- ❌ Cost-listing without "Why." Skipping the logical step makes the trace look mystical. Always show how the position requires the denial.
- ❌ More than 3. If you find 5 load-bearing negative-claims, the position is entangled — return INDETERMINATE or surface the entanglement as the verdict, don't list all 5.
Relationship to other skills
| Skill | Shape | Relationship |
|---|
falsify (via critical-thinking Mode 4) | What would prove this wrong? | falsify works on empirical claims; negative-claim works on value-laden / strategic claims — sibling tools |
silent-assumption | Surface unstated dependencies | silent-assumption + negative-claim: assumptions are upstream (what V depends on); negative-claims are downstream (what V is committed to) |
steelman (via critical-thinking Mode 2) | Strongest counter-argument | steelman finds the best opposing case; negative-claim finds what V's own position is committed to denying — different operations |
grant-then-test | Concession move | After negative-claim, grant-then-test the painful denials: assume V holds the position, see if the denials are sustainable |
distill | Convergence on a claim | distill produces the position; negative-claim traces its commitments |
What this does NOT do
- Generate counterarguments (use steelman via critical-thinking Mode 2)
- Falsify empirical claims (use falsify)
- Police identity claims (out of scope per
identity-vs-empirical.md)
- Recommend whether V should hold the position (that's V's call; negative-claim only surfaces the commitments)
- Manufacture prohibitive verdicts to seem rigorous
Real V examples (V to fill in 1-2 from her actual work)
Example 1 (placeholder): [V — fill in a real case where holding a brand/strategic/value position committed you to denying something you hadn't realized. What was the position? What was the negative claim? What was the cost?]
Example 2 (placeholder): [V — a case where two positions you held had contradictory negative-claims (you were affirming on one track what you were denying on another). How was it resolved?]
