Frame Rotator

Frame Rotator — Alternative Formulations

A problem statement is one of many possible formulations. The framing chooses what counts as relevant and constrains the solution space. Frame-rotator rotates through alternatives that would lead to different solution spaces. The goal isn't to find "the right framing" — that's downstream — but to expose that the original framing is not unique.

The skill applies three rotation operations: How Might We (IDEO), level-shifting (Polya specialize/generalize), and framing-assumption swap (via assumption-excavator). Most useful alternatives come from a mix of all three.

The Three Rotation Operations

1. HMW (How Might We) — IDEO

Reformulate the statement using the canonical HMW formula:

How might we [intended action verb] for [potential user/subject]
so that [desired outcome]?

The verbs are deliberately exploratory: "explore", "enable", "remove the need for", "redesign", "reframe", "delight". Not "build", "implement", "fix" — those are mechanism-shaped, not goal-shaped.

The "might" matters: it preserves multiple possible solutions. "How do we" forecloses; "how might we" opens.

2. Level-shifting — Polya specialize/generalize

Two opposite moves Polya named in How to Solve It:

For each statement, both directions are valid candidates. The skill produces at least one of each on deep-mode runs.

3. Framing-assumption swap

Call assumption-excavator (philosophy/logic) with depth=deep, category=framing. Receive:

• The framing assumption(s) the statement makes
• Counterfactuals: "if this framing assumption is false, then [consequence]"

For each contestable framing assumption, produce an alternative formulation that swaps that assumption. Example: a statement framed as "engagement optimization" might be rotated to a "wellbeing optimization" framing — the metric (and so the solution space) changes.

Process

1. Verify input is an audited statement (from statement-grader). Unaudited input gets routed back.
2. Run HMW reformulation — at least one alternative.
3. Run level-shifting — at least one specialize and one generalize alternative.
4. Call assumption-excavator with depth=deep, category=framing. Receive the framing assumptions.
5. For each load-bearing framing assumption, produce an alternative that swaps it.
6. Deduplicate — collapse near-identical alternatives. Aim for 2–3 meaningfully different formulations in standard mode; ≥3 in deep mode.
7. Diff-check — confirm each alternative is structurally different from the original, not just lexically rephrased.

The Diff-Check Discipline

A frame rotation must change what counts as a solution, not just what words describe the problem. The diff-check question: would a downstream solver propose a different solution given the alternative formulation? If no, the rotation is cosmetic, not structural. Discard.

Example (cosmetic, fails check):

• Original: "Reduce customer churn"
• Alternative: "Decrease the rate at which customers leave"
• → Same solution space. Reject.

Example (structural, passes check):

• Original: "Reduce customer churn"
• Alternative (specialize): "Reduce churn specifically in the first 30 days post-signup"
• Alternative (generalize): "Improve the customer's value-realization curve — churn is one symptom"
• Alternative (framing swap): "Stop optimizing for retention; optimize for the customers we actually want to retain"
• → All three open different solution spaces. Pass.

Output Format

FRAME ROTATION — [first 60 chars of original statement...]
─────────────────────────────────────────────
Original framing:        [the statement as input]
Framing assumption(s):   [from assumption-excavator — what's being assumed]

Alternative 1 (HMW):
  Formulation: [rewritten as HMW]
  Operation:   How Might We
  What changes: [what assumption is relaxed; what solution space opens]
  Diff-check: pass

Alternative 2 (Specialize):
  Formulation: [narrower version]
  Operation:   Specialize
  What changes: [the narrowing dimension; why the smaller case is more tractable]
  Diff-check: pass

Alternative 3 (Generalize):
  Formulation: [broader version]
  Operation:   Generalize
  What changes: [the abstraction; what becomes solvable at the higher level]
  Diff-check: pass

[Optional, deep mode] Alternative 4 (Framing swap):
  Formulation: [swapped framing]
  Operation:   Framing assumption swap
  Original assumption: [what's being swapped]
  What changes: [new solution space]
  Diff-check: pass

Diff-check failures: [list of generated alternatives that were cosmetic-only and were rejected]

Mode Behavior

Failure Modes

Output Contract for six-eyes

Called from Phase 4 (Reframe). Returns 2–3 (standard) or ≥3 (deep) alternative formulations. Each is a candidate for six-eyes to grade via statement-grader. If any alternative scores better than the original on the audit axes, that alternative becomes the new working statement for downstream phases.

Connections

• assumption-excavator (philosophy/logic) — primary cross-domain call for framing-assumption excavation
• polya-method (binding-vow) — foundational reference for specialize/generalize operations
• inversion-tool (binding-vow) — sibling reframing skill (Munger inversion is complementary to HMW)
• stakeholder-rotator (binding-vow) — sibling for stakeholder-perspective rotation
• statement-grader (binding-vow) — downstream consumer; grades each alternative
• claim-decomposer (research) — escalation path when rotation fails due to narrow scope

Sources

• IDEO / d.school — How Might We methodology. (designkit.org)
• Pólya, G. (1945). How to Solve It. — specialize/generalize heuristic family.
• See [[polya-heuristics-catalog]] in skill-lab for the catalog of Polya heuristics this skill draws from.