Mechanism Hypothesis Generation Chain

validated scientific chain v1.0.0 cc-by-sa

Block ID: 9f20b5e2-0c20-4d87-a3f6-929f85a5e95a

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.

Template

Generate and rank candidate mechanisms that could explain the observation in {observation}. (1) State the phenomenon precisely, including its boundary conditions — when it occurs, when it doesn't, its magnitude, and its time course, because a good mechanism must account for all of these, not just the headline effect. (2) Generate multiple candidate mechanisms deliberately, spanning levels of explanation (molecular, systemic, environmental, behavioral) — settling on the first plausible mechanism is the trap; force at least three genuinely different ones. (3) For each candidate, derive what else must be true if it operates — the additional observable consequences — since mechanisms are distinguished by their side predictions, not by their fit to the original observation which they all explain by construction. (4) Check each against the boundary conditions from step 1: a mechanism that can't explain when the effect vanishes is incomplete. (5) Rank by prior plausibility (consistency with established science) times explanatory completeness, penalizing mechanisms that require unlikely coincidences. (6) Design the single experiment or observation whose result would most cleanly separate the top two mechanisms, and state what each predicts.

Generate and rank candidate mechanisms that could explain the observation in {observation}. (1) State the phenomenon precisely, including its boundary conditions — when it occurs, when it doesn't, its magnitude, and its time course, because a good mechanism must account for all of these, not just the headline effect. (2) Generate multiple candidate mechanisms deliberately, spanning levels of explanation (molecular, systemic, environmental, behavioral) — settling on the first plausible mechanism is the trap; force at least three genuinely different ones. (3) For each candidate, derive what else must be true if it operates — the additional observable consequences — since mechanisms are distinguished by their side predictions, not by their fit to the original observation which they all explain by construction. (4) Check each against the boundary conditions from step 1: a mechanism that can't explain when the effect vanishes is incomplete. (5) Rank by prior plausibility (consistency with established science) times explanatory completeness, penalizing mechanisms that require unlikely coincidences. (6) Design the single experiment or observation whose result would most cleanly separate the top two mechanisms, and state what each predicts.

Variables

Name	Type	Required	Trust level
observation		yes

mechanismhypothesis-generationabductionchain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18