Library / First Principles Framework (FPF) - Core Conceptual Specification
Jump to passage
In this reading

Link to current text

Published source confirmed at last check

Source changed 2026-10-03 08:25:59 UTC · snapshot created 2026-10-03 08:26:43 UTC · last check 2026-10-03 09:00:05 UTC

C.28.CM:4.5 - Derive a consequence that can distinguish the accounts

Work the same question through each retained model. Start with a direction, a possible or impossible outcome, a conditional independence, or an observation that one account can explain only by adding another premise. Show which relation makes the consequence follow.

Keep three operations distinct. Conditioning restricts attention using an observed value. Intervention replaces the specified mechanism. Omitting a variable from a report or marginalizing it out leaves its possible influence in the subject model. Treating all three as “fix the variable” can reverse the conclusion.

When the mechanisms are specified, C.28.MR derives an intervention consequence while retaining the relevant other mechanisms and underlying inputs. Use C.29 when expressing or computing the model requires further mathematical construction. If the needed function or input law is missing, return that precise gap along with any qualitative result that survives.

Two graphs may entail the same observed independences. Two mechanisms may produce the same measurements in the available range. Retain both when the present material does not distinguish them. A causal-discovery procedure can contribute within its declared model class and assumptions; its returned graph does not remove those conditions.