Diagnostic Possibility Map¶
Diagnostic map — instantiates Problem Space Mapping
Lays out the plausible causes of a symptom alongside the tests that would confirm or exclude each, so diagnosis proceeds by ruling regions in and out rather than latching onto the first guess.
A Diagnostic Possibility Map represents a troubleshooting problem as the space of plausible causes behind an observed symptom, each paired with the tests, findings, and exclusions that would move it from suspected to confirmed or ruled out. Its defining commitment is that the map is revised by evidence: it is never a static list of hypotheses but a working ledger in which every test result prunes a branch, marks a dead end, or promotes a lead — and where a finding that seemed conclusive but was not is flagged as a trap. This is what makes it a diagnostic mechanism rather than a design or strategy one: it starts from a problem that has already occurred, and its whole motion is convergent — narrowing a field of causes toward the real one by testing, not by generating new possibilities. It is the tool for the question "what is actually wrong here, and how would we know?"
Example¶
A car comes into an independent garage with an intermittent engine stall — it dies at random, then restarts fine, so there is no steady fault to chase. The mechanic builds a Diagnostic Possibility Map instead of swapping parts. The initial state is captured precisely: stalls only when warm, only at idle, no warning light, roughly once a drive. From that symptom the plausible causes fan out — failing idle-air-control valve, intermittent crank-position sensor, fuel-pump dropout, a marginal ground connection — each written with the test that would decide it: monitor sensor voltage during a stall, check fuel pressure under a warm idle, wiggle-test the grounds.
The map's discipline shows in what it prevents. An early code points at the crank sensor, and the reflex is to replace it and close the ticket. But that reading is marked a trap: the same code appears whenever any warm-idle stall interrupts the signal, so it is a symptom of the stall, not a cause. The validation loop does its work when the fuel-pressure test comes back marginal — that branch is promoted, the sensor branch is set aside (not deleted, in case it returns), and the ground-wiggle branch is marked a dead end after it holds steady. Diagnosis converges by exclusion, and the customer does not pay for a sensor that was never broken.
How it works¶
- Fix the symptom as the starting state. Record the fault precisely — conditions, timing, what is and is not present. A vague symptom yields a useless map.
- Fan out the plausible causes. Enumerate the hypotheses that could produce this symptom, broad enough to avoid tunnel vision.
- Attach a discriminating test to each. For every cause, note the observation that would confirm or exclude it — ideally a test that separates several branches at once.
- Run tests and update. Each result rules a branch out (dead end), promotes one, or is flagged as a misleading finding (a trap) that does not decide what it appears to.
- Converge. Repeat until one cause survives with confirming evidence, keeping ruled-out branches visible in case the picture changes.
Tuning parameters¶
- Hypothesis breadth — how wide the initial fan of causes is. Broad guards against missing the real one; narrow gets to a test faster but risks anchoring.
- Test order — sequencing by cheapest-first, or by most-discriminating-first (the test that splits the field fastest). Discriminating-first converges quickest; cheap-first spends less per step.
- Exclusion strictness — how strong a negative result must be to mark a branch dead. Loose leaves zombie hypotheses alive; strict risks discarding an intermittent true cause.
- Trap sensitivity — how readily a finding is flagged as possibly-misleading rather than conclusive. High sensitivity avoids false closure but slows commitment.
- Reopen policy — whether ruled-out branches are deleted or merely greyed, ready to return if new evidence contradicts the current lead.
When it helps, and when it misleads¶
Its strength is convergence without tunnel vision: by holding the whole differential visible and attaching a test to each branch, it fights the single most common diagnostic error — premature closure, latching onto the first plausible cause and stopping the search.[n1] Keeping ruled-out and trap branches on the map is what lets a diagnosis reopen gracefully when the "obvious" answer turns out wrong.
Its failure mode is the mirror image: an over-full map that never converges, every test spawning two new hypotheses until the differential is unmanageable and the car never leaves the bay — analysis paralysis in diagnostic dress. A subtler misuse is confirmation-driven testing, choosing tests that would confirm the favored cause rather than ones that could exclude it. The guarding discipline is to prefer discriminating tests that can kill a branch, to enforce a stopping rule once one cause is confirmed and its rivals excluded, and to distinguish a true dead end from a branch merely resting. Note the map does not maintain a standing ledger of unproven background assumptions across problems — its exclusions are about this symptom.
How it implements the components¶
initial_state— the precisely characterized symptom the diagnosis starts from; the anchor every hypothesis must explain.dead_end_and_trap_marker— excluded causes are marked as dead ends, and findings that mislead (a code that is a symptom, not a cause) are flagged as traps so they cannot drive a false closure.map_validation_loop— the core motion: each test result is fed back to prune, promote, or reopen branches, keeping the map a working model rather than a frozen hypothesis list.
It does not maintain the standing register of unverified background assumptions (evidence_gap_marker) — that cross-cutting ledger is the Unknowns and Assumptions Register's — nor does it hold the feasibility structure of options (constraint_set), which the Constraint Matrix owns.
Related¶
- Instantiates: Problem Space Mapping — externalizes the space of possible causes and converges it by evidence.
- Sibling mechanisms: State / Action Map · Constraint Matrix · Design Space Map · Search Space Diagram · Strategic Option Map · Option Map · Unknowns and Assumptions Register
Editorial Notes¶
Form Classification¶
Form family: Representation, Specification & Plan
Rationale: Diagnostic Possibility Map operates as a non-executable information artifact that externalizes static or prospective structure because it lays out the plausible causes of a symptom alongside the tests that would confirm or exclude each, so diagnosis proceeds by ruling regions in and out rather than latching onto the first guess.
Independent corroboration: The frozen evidence defines Diagnostic Possibility Map as 'Lays out the plausible causes of a symptom alongside the tests that would confirm or exclude each, so diagnosis proceeds by ruling regions in and out rather than latching onto the first guess', so its operative form is Representation, Specification & Plan.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Medicine & Healthcare
Origin pattern: Convergent development
Present-day reach: Multi-domain
Rationale: Clinical differential diagnosis cohered maps of plausible causes paired with tests that confirm or exclude each branch.
Related originating lineages:
- Engineering & Design — Fault-tree troubleshooting independently developed test-linked hypothesis maps for technical failures.
Review resolution: Clinical differential diagnosis cohered maps of plausible causes paired with tests that confirm or exclude each branch. The retained alternate lineages materially shaped the mechanism's form.
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] Premature closure is the well-documented reasoning error of accepting a diagnosis before it has been fully verified, once an early hypothesis feels sufficient. Keeping the full differential and its discriminating tests explicit is the standard corrective, which is why the map refuses to delete rival branches until they are actually excluded. ↩