Separate Generator and Evaluator Roles¶
Role separation protocol — instantiates Evaluation Criteria Suspension During Divergence
Splits generation and evaluation across different people so no one critiques what they are still trying to produce.
Separate Generator and Evaluator Roles suspends premature evaluation structurally, by division of labour rather than by rule, timing, or workflow. One set of people generates; a different set evaluates; and the two roles are held by different bodies. Because the person producing ideas is not the person who will judge them, there is no internal critic to appease — the generator has literally been relieved of the evaluation job. What makes it this mechanism is that its lever is who does what: the suspension is achieved by making self-censorship impossible, since you cannot pre-emptively fail your own idea against criteria you were told are not yours to apply. Its bet is that the surest way to stop evaluation from contaminating generation is to remove the evaluation from the generator's job description entirely — and, ideally, to strip authorship so the evaluators judge the idea rather than the person.
Example¶
An advertising agency splits an unusual campaign brief into two teams that never share a room during the creative phase. The generator team — copywriters and art directors — is told plainly: your only job is volume and range; you will not defend, cost, or pitch these; someone else decides. They produce ninety concepts, each stripped of the author's name before it leaves the room. The evaluator team — strategists and account leads — receives the anonymized concepts and applies the client's criteria (brand fit, budget, media plan) to them, without knowing which star creative or which intern produced which line.
Two things happen that a single combined room never achieves. The generators, freed from having to answer "how would we ever sell this to the client," put forward concepts they would normally have killed in their own heads. And the evaluators, blind to authorship, rate a junior's odd concept on its merits rather than deferring to the senior creative's reputation — the anonymized odd concept ends up scoring highest and going to the client. The one norm both teams share is that generation and judgment do not happen in the same heads at the same time.
How it works¶
The protocol rests on three structural commitments. First, a distinct evaluating body: a parallel convergence team, separate from the generators, owns the judgment and is the only group that scores. Second, anonymized handoff: ideas cross from generators to evaluators stripped of authorship, so evaluation lands on the idea rather than the idea's owner and status can't tilt the verdict. Third, a shared deferral norm enforced by the split itself: generators are explicitly relieved of the evaluation task, so deferral is not willpower but role definition — there is nothing for a generator to defer to. The mechanism says nothing about which criteria the evaluators use or in what order; it only fixes that different people generate and judge, and that the handoff is blind.
Tuning parameters¶
- Role rigidity — hard, fixed teams versus roles that swap between rounds. Rigid separation maximally protects generation but can breed a generator/evaluator adversarial dynamic; rotation spreads empathy but risks each person re-importing their own critic.
- Anonymity depth — fully blind handoff versus attributed. Blind handoff kills status bias but severs the ability to ask the author "what did you mean?"; attribution preserves context but reopens reputation effects.
- Information handoff richness — bare ideas versus ideas plus rationale and constraints. Richer handoffs help evaluators judge fairly but leak evaluative framing back toward generators.
- Feedback loop — one-way (evaluators decide, done) or iterative (verdicts return to generators for another round). Loops improve ideas but risk generators writing to please the evaluators, collapsing the separation.
- Team composition overlap — zero shared members versus a liaison. No overlap keeps roles clean; a liaison eases translation but can become a back-channel that re-fuses the roles.
When it helps, and when it misleads¶
Its strength is that it makes self-censorship structurally impossible and neutralizes status bias in one move: the generator has no verdict to anticipate, and the evaluator has no author to defer to. It is a role-based cousin of de Bono's Six Thinking Hats, which likewise separates generative thinking from critical thinking so the two never compete in the same moment.[n1]
Its failure mode is the throw-it-over-the-wall problem: generators, disconnected from the criteria, produce polished ideas that are irrelevant or infeasible, and evaluators, disconnected from intent, reject ideas whose point they never grasped — a hand-off gap that wastes both sides' work. A subtler failure is an adversarial drift in which evaluators enjoy rejecting and generators stop taking risks to avoid the sting. The classic misuse is letting the roles quietly re-fuse — the same senior person who "generates" also sits on the evaluation panel — which restores exactly the self-censorship the split was meant to remove. The guarding discipline is to keep at least one genuine handoff of context with the ideas (intent, not verdict) and to police the role boundary so no individual sits on both sides of the same decision.
How it implements the components¶
parallel_convergence_team— a separate evaluating body owns all judgment, running convergence in parallel to and independent of the generators.anonymous_idea_mode— ideas cross the handoff stripped of authorship, so evaluation targets the idea and status cannot sway the verdict.judgment_deferral_norm— deferral is enforced by role definition: generators are relieved of evaluation entirely, so there is no internal critic to suppress.
It does not implement idea_capture_buffer (parking ideas to score later in time — that's Deferred Scoring Queue) or criteria_reintroduction_sequence (staging which criteria return first — that's Two-Pass Evaluation); this mechanism separates generation from evaluation across people, not across time or criteria order.
Related¶
- Instantiates: Evaluation Criteria Suspension During Divergence — supplies the structural, division-of-labour route to suspending evaluation.
- Sibling mechanisms: No-Critique Rule · Yes-And Extension · Wild-Card Quota · Morphological Matrix · Deferred Scoring Queue · Two-Pass Evaluation · Timeboxed Blue-Sky Sprint
Editorial Notes¶
Form Classification¶
Form family: Organization, Role & Governance
Rationale: Separate Generator And Evaluator Roles operates by maintains separate generating and evaluating bodies with exclusive scoring authority and anonymized handoff. That concrete deployed or enacted form is Organization, Role & Governance under the frozen taxonomy.
Nearest alternative: Structure, Architecture & Configuration — Although Structure, Architecture & Configuration can support this mechanism, the frozen evidence makes its operative form the act that maintains separate generating and evaluating bodies with exclusive scoring authority and anonymized handoff; the alternative is therefore secondary rather than defining.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Organizational & Management Science
Origin pattern: Convergent development
Present-day reach: Universal
Rationale: Separating ideation from judgment is a facilitation and organization-design pattern for preventing premature evaluation from suppressing production.
Related originating lineages:
- Art & Aesthetics — Studio practice distinguishes generative making from later critical review.
- Education & Pedagogy — Writing pedagogy separates drafting from critique to sustain fluency and revision quality.
- Psychology — Divergent and convergent thinking rely on different cognitive sets and inhibition levels.
- Systems Thinking & Cybernetics — Systems thinking, feedback control, and cybernetics supplies a parallel or contributing lineage for the mechanism's defining operation: splits generation and evaluation across different people so no one critiques what they are still trying to produce.
Review resolution: The blind reviewers agree that organizational_management is the primary origin and differ only on alternate origin disagreement. I preserve every independently explained alternate from both records rather than imposing a numeric cap. I retain convergent because the combined record shows independent disciplinary development. The broader reach of universal records portability separately from historical provenance, and encyclopedia_synthesis=true preserves the affirmative synthesis judgment where either reviewer identified one.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] Edward de Bono's Six Thinking Hats assigns distinct modes of thinking to distinct "hats" (e.g., green for generative ideas, black for critical caution) so that a group does one kind of thinking at a time rather than letting critique and creation collide. This mechanism achieves the same separation by assigning the modes to different people rather than to different hats worn in turn. ↩