Minimal Causal Diagram¶
Artifact — instantiates Core Model First
Draws only the core variables and causal relations needed to test the central explanation.
A Minimal Causal Diagram is a picture with a rule: nodes are the few variables that carry the central explanation, arrows are the causal relations you are willing to defend, and the edge of the drawing is itself a claim — everything left off the page is being asserted not to matter for the question at hand. Its one idea is legibility-forcing: it takes a tangle that stakeholders can only argue about in words and pins it to a small, shared, testable spine. Where First-Principles Model derives the core relation from invariants, this mechanism draws it — its contribution is making the causal structure visible and bounding its frame so that disagreement can be located on a specific arrow instead of dissolved into "it's complicated."
Example¶
A lake turns green every summer and the fish start dying. A town meeting produces two dozen candidate causes — fertilizer, boats, warm weather, a new housing development, waterfowl, the fish themselves. Rather than litigate all of them, a limnologist draws a minimal causal diagram on the whiteboard: phosphorus runoff → algae growth → algal die-off → oxygen depletion → fish kill, with summer temperature marked as an enabling condition on the algae-growth arrow. Four nodes, one chain. Everything else — the specific algae species, the exact lake-bed shape, the boat traffic — is deliberately drawn outside a boundary box labeled "explains the summer fish-kill, not the whole ecosystem."
The diagram does one thing the word-argument couldn't: it makes the central explanation testable. If phosphorus is really the driver, cutting runoff should break the chain upstream of everything else. The conversation stops being "what matters?" and becomes "does the phosphorus arrow hold?" — a question a monitoring program can actually answer.
How it works¶
- Choose the fewest nodes that can carry the explanation. If removing a variable doesn't change the central story, it doesn't belong on the diagram.
- Draw only arrows you would defend. Each relation is a commitment; speculative links are marked as tentative or left off.
- Fence the frame. State explicitly what the diagram covers and what is excluded — and, for each excluded node, why it is safe to exclude.
- Read off the falsifier. Identify the arrow whose failure would break the explanation; that is what the diagram exists to test.
Tuning parameters¶
- Node budget — how many variables the diagram may carry. Fewer nodes are clearer and more testable but risk omitting a confounder; more nodes creep back toward the tangle the diagram was meant to cut.
- Arrow commitment — whether to draw only defensible relations or include speculative, dashed ones. Speculation preserves candor about uncertainty; too much of it dilutes the spine.
- Boundary tightness — how narrowly the frame is drawn. A tight frame makes a sharp, testable claim but may exclude something that turns out to matter.
- Annotation depth — whether arrows carry signs, strengths, or lags. Annotation adds testable precision but also more to get wrong.
When it helps, and when it misleads¶
Its strength is a shared, testable spine: it converts a diffuse argument into a small object everyone can point at, and it exposes exactly where people disagree. Because the diagram names a falsifier, it also tells you what evidence would settle the question.
Its central failure mode is the omitted variable. A minimal diagram that leaves a genuine confounder off the page will bias every inference drawn from it — the picture looks clean precisely because the thing that would complicate it was excluded. This is the standing hazard of causal DAGs[n1]: a diagram is only as honest as its boundary, and a missing back-door path silently corrupts the conclusion. The classic misuse is drawing the arrows you wish were true and calling the picture a proof. The guarding discipline is to justify each exclusion as deliberately as each inclusion, and to treat the diagram as a hypothesis to be tested, never as evidence in itself.
How it implements the components¶
Minimal Causal Diagram fills the make-the-spine-visible-and-bounded side of the archetype:
core_variable— the nodes are exactly the minimal set of variables the central explanation requires; anything that doesn't change the story is left off.core_relation— the arrows are the explicit, defensible causal relations that give the model its spine.model_scope_statement— the boundary of the drawing states what is in frame and what is excluded, making the model's reach an explicit claim.
It draws the spine but does not test it against data — that validation is Baseline Model's job — and it does not derive the relation from underlying invariants the way First-Principles Model does.
Related¶
- Instantiates: Core Model First — Minimal Causal Diagram is one way to originate the visible, testable core the archetype refines from.
- Sibling mechanisms: First-Principles Model · Baseline Model · Simple Prototype · Stripped-Down Simulation · Minimum Viable Explanation · Core Architecture Sketch · Toy Model
Editorial Notes¶
Form Classification¶
Form family: Representation, Specification & Plan
Rationale: Minimal Causal Diagram operates as a non-executable information artifact that externalizes static or prospective structure because it draws only the core variables and causal relations needed to test the central explanation.
Independent corroboration: The frozen evidence defines Minimal Causal Diagram as 'Draws only the core variables and causal relations needed to test the central explanation', so its operative form is Representation, Specification & Plan.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Minimal causal graphs used for identification and testing belong to modern causal inference.
Related originating lineages:
- Philosophy — Philosophy of causation supplies explanatory adequacy and parsimony concerns.
- Systems Thinking & Cybernetics — Systems diagrams provide the practical directed-relations representation.
Review resolution: Both independent reviews place the primary provenance in statistics_experimental_design. The queued differences (alternate_origin_disagreement, domain_reach_disagreement, encyclopedia_synthesis_disagreement) concern secondary metadata, not primary lineage. The final retains philosophy, systems_cybernetics only where a reviewer supplied a formative-lineage rationale; downstream use or broad applicability by itself is not treated as origin. origin_mode=cross_disciplinary_synthesis because the supplied rationales identify formative contributions that are composed in the mechanism's present form. domain_reach=universal records established application breadth separately from provenance. confidence=medium preserves the more cautious evidence assessment. encyclopedia_synthesis=true records whether either reviewer identified deliberate corpus-level composition.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] In causal inference, a directed acyclic graph encodes which variables cause which; the validity of any conclusion drawn from it depends on no relevant confounder (back-door path) being omitted. A minimal causal diagram inherits this vulnerability exactly — its clarity is bought by exclusions, and a wrong exclusion is invisible on the page. ↩