How to Diagnose a Suspected Civilisation Failure | From Delta to Threshold Breach

The Reason for Existence

This page began with a useful instinct: collapse should be diagnosable rather than treated as a dramatic story.

The earlier version tried to accomplish that through a fixed EducationOS kernel and an immutable set of axes: Depth, Load, Transfer, Outcome, Cohesion, Alignment, Environment and Time. That architecture preserved several things worth keeping.

• Define the system boundary before making a claim.
• Use observable indicators rather than rhetoric.
• Look at trajectories, not snapshots.
• Distinguish capability from appearance.
• Find binding constraints instead of prescribing generic effort.
• Ask whether truth can reach the people able to act.
• Define a retest so repair can be verified.

Those ideas survive.

What does not survive is the claim that every civilisation must be forced through one immutable sequence, that failure usually follows one universal order, or that repair must always begin with one of three learning axes.

Civilisations differ in scale, ecology, technology, institutions, culture, dependency structure, historical inheritance and the capabilities on which ordinary life depends. The same observable symptom can mean very different things in different operating fields.

A diagnostic framework becomes stronger when it can say “this lens does not fit” rather than forcing the world through the lens.

The current CivilisationOS therefore treats the older DLT, OHME and related lenses as optional specialist instruments. Depth, Load and Transfer remain useful when capability formation is the problem. They are not universal civilisation coordinates.

The broader diagnostic spine is now:

OBSERVABLE CONDITION
→ REFERENCE / ACTUAL COMPARISON
→ DELTA
→ SCALE × TIME × CAPABILITY × REALITY LAYER
→ RECEIVER
→ DEPENDENCY / FLOOR
→ VIABILITY MARGIN
→ THRESHOLD TEST
→ PROPAGATION TEST
→ REPAIRABILITY TEST
→ CLASSIFICATION + CONFIDENCE
→ WORLD RETURN

The RFE is simple: do not call something civilisational failure until the evidence earns the label—and do not miss a real floor breach because the average still looks comfortable.

Start with Article 1: distance from perfection is not distance to failure

The companion article How Civilisation Changes Direction | The Perfect Flight Path, Actual Trajectory and the Delta Between Them establishes three comparison objects:

• Perfect Reference Flight Path — what the best defensible civilisation trajectory could look like under declared reality and uncertainty.
• Best Feasible or Expected Path — what competent execution should reasonably achieve from the present starting point and constraints.
• Actual Civilisation Trajectory — what really happened across receivers, systems, places and time.

The difference between them creates useful deltas.

A large aspiration gap is not automatically a crisis. A civilisation can be far from its best possibility and still be viable, improving and well inside its safety envelope.

Diagnosis begins when we ask a different question:

Is the actual trajectory merely imperfect—or is a necessary function approaching or crossing the point below which it can no longer operate?

That is the transition from delta analysis to threshold diagnosis.

The diagnostic object must be named before the diagnosis

“The civilisation is collapsing” is not yet a useful diagnostic statement.

It is too large.

A diagnostician should first be able to write something closer to:

In this city, during this period, safe-water delivery to these receivers has become less reliable because these pumping, power and maintenance dependencies are degrading; the current evidence suggests the service is approaching a specified functional threshold, but the wider water system has not yet been shown to have crossed it.

That sentence is longer than “water collapse.” It is also much more useful.

The diagnostic object should declare at least:

Field	Question
Object	What system, service, capability, place or population are we actually examining?
RFE	What function is this object supposed to perform?
Scale	Person, household, institution, city, state, region, civilisation or world?
Time	Minutes, months, years, generations or longer?
Receiver	Who or what must actually receive the output?
Reference	What is the defensible normal, expected, feasible or perfect comparison?
Actual	What is happening in the world now?
Delta	Which important differences exist and how are they changing?

If the object cannot be bounded, the diagnosis should remain provisional.

The evidence ledger: do not mix observation, inference and model

One of the fastest ways to produce a false collapse diagnosis is to treat different kinds of evidence as if they were the same.

CivilisationOS should separate at least seven evidence states.

Evidence state	Meaning	Example
Observed fact	A directly documented event or state.	A substation tripped at a recorded time.
Measurement	A quantified observation produced by a defined method.	Reservoir level, waiting time, fault rate, mortality rate.
Receiver report	Evidence from the person or system experiencing the output.	Residents report no water despite the distribution dashboard showing service restored.
Claim	A statement asserting what happened or why.	“The outage was caused by poor maintenance.”
Inference	A reasoned interpretation connecting evidence.	Repeated failures after maintenance deferral suggest a causal contribution.
Modelled state	A latent condition estimated from observations.	System resilience, probability of threshold crossing, hidden backlog.
Unknown / disputed	The evidence does not yet support a stable conclusion.	Whether a trust decline is temporary or structural.

The diagnosis becomes stronger when these states remain visible instead of being compressed into one confident story.

Evidence can support a diagnosis without eliminating uncertainty.

The four coordinates: Scale × Time × Capability × Reality Layer

Once the object is resolved, locate the suspected failure on the four whole-picture coordinates used by the Civilisation Hub.

Scale — where does the failure exist?

A household can cross a food threshold while the national food system remains well supplied. A local hospital can fail while the national health system reroutes demand successfully. A national institution can collapse while a wider cultural civilisation continues.

Scale prevents a true local failure from being inflated into a false total diagnosis—and prevents a national average from hiding a real local breach.

See How Civilisation Changes With Scale.

Time — what kind of movement is this?

A one-day outage, five-year maintenance decline and multi-generational loss of technical knowledge are different failure objects.

Time tells us whether the system is experiencing shock, seasonal variation, chronic degradation, path dependence, delayed consequence or continuity loss.

See The Time Structure of Civilisation.

Capability — what can no longer be done?

Do not diagnose the object by prestige. Diagnose the capability.

Can people still survive, cooperate, coordinate, remember, exchange, build, govern, learn, maintain, repair, adapt and hand forward?

A magnificent institution can remain visually intact after its capability has gone. A modest informal network can continue carrying essential capability after formal institutions fail.

See What Must a Civilisation Be Able to Do?.

Reality Layer — where is the constraint carried?

Nature, human survival, infrastructure, institutions, knowledge, culture and human outcomes can each carry or transmit failure.

A school may appear institutionally healthy while the human-survival layer—sleep, food, health, safety—prevents learning. A water utility may have strong institutions but an ecological source constraint. A technically functioning digital service can fail at the human-outcome layer because the receiver cannot use it.

See The Layers of Civilisation.

The receiver test: where does the output finally have to arrive?

Systems can remain busy while their receivers have already fallen through the floor.

A relief warehouse is stocked. The household receives nothing.

A school teaches. The student cannot transfer the learning.

A hospital appointment exists. The patient cannot physically or financially reach it.

A water system produces sufficient water. A building’s internal plumbing makes the final water unsafe.

The receiver test asks:

• Who is the intended receiver?
• Can the output reach them?
• Can they understand or use it?
• Does timing matter?
• Does cost block access?
• Does disability, language, geography or identity create a last-mile barrier?
• Who receives risk rather than benefit?
• Which receiver disappears inside the average?

Delivered is not the same as received. Received is not the same as useful.

From delta to viability margin

Article 1 tells us how far the actual trajectory is from the reference.

Article 2 asks how much margin remains before a required function changes state.

REFERENCE
   ↑
   │ ASPIRATION / PERFORMANCE GAP
   │
ACTUAL STATE
   │
   │ VIABILITY MARGIN
   ↓
FUNCTIONAL THRESHOLD
   ↓
FLOOR BREACH

The aspiration gap can be large while the viability margin is comfortable.

The aspiration gap can also be small while the viability margin is dangerously thin.

A railway may be highly punctual but operating with too little maintenance margin. A company may be profitable but dependent on one irreplaceable person. A city may have abundant electricity but insufficient black-start capability. A civilisation may have excellent universities but a thinning pipeline of maintainers in critical infrastructure.

High performance does not prove high margin.

What is a threshold?

A threshold is not merely a target that has been missed.

A target says what performance is desired.

A threshold marks a boundary below or beyond which the required function can no longer be sustained, changes qualitatively, or becomes unsafe.

Some thresholds are technically explicit: voltage, pressure, temperature, structural capacity, staffing minimums, reserve limits.

Others are distributed and contextual: enough legitimacy to coordinate without unsustainable force, enough knowledge carriers to train successors, enough trust for financial claims to circulate, enough ecological stability for settlement, enough teacher capacity for a school to deliver meaningful learning.

Where a threshold cannot be measured directly, CivilisationOS should use a defensible evidence bundle rather than inventing a false universal number.

Threshold approach is not threshold breach

This distinction prevents premature collapse diagnoses.

Condition	What it means	Diagnostic implication
Comfortable margin	Disturbance can be absorbed with ordinary buffers.	Focus on improvement and maintenance.
Shrinking margin	Capacity, reserves, trust, staffing or other buffers are being consumed.	Investigate degradation and trend.
Near-threshold	Small additional load may produce disproportionate loss.	Stress-test, unload, build alternatives and improve observability.
Threshold breach	The required capability falls below functional viability.	Containment and emergency repair take priority.
Floor failure	The breached capability can no longer support dependent functions.	Map propagation and protect a viable core.

Calling near-threshold conditions “collapse” can create panic and destroy precision. Calling an actual breach “temporary stress” can waste the remaining repair window.

The floor test: is this function load-bearing?

Not every failure is a floor failure.

A floor is load-bearing when other important functions depend on it strongly enough that its sustained loss removes their operating conditions.

The exact floors differ across civilisations.

• Ecology and habitability — climate, soils, water cycles and physical conditions.
• Biological survival — food, safe water, shelter, sanitation, health and care.
• Energy — where dependent systems require it.
• Resource movement — storage, logistics, exchange and last-mile delivery.
• Security — enough protection from violence and arbitrary seizure for planning.
• Trust and cooperation — enough reliable expectation for strangers to work together.
• Coordination — shared signals, timing, roles, standards and communication.
• Institutions and legitimacy — rules, authority, accountability and sufficient acceptance.
• Information and shared reality — evidence can travel, be checked and corrected.
• Knowledge and education — useful learning can accumulate and transfer.
• Maintenance — wear and drift can be detected and corrected.
• Repair — damaged systems can be restored or reconfigured.
• Capability transfer — successors can operate and improve what they inherit.

The companion article The Minimum Conditions for Civilisation looks at these functions from the formation side. Failures of Civilisation | When the Floor Gives Way looks at them after breach.

The dependency test: what fails next if this fails?

A floor is partly defined by what depends on it.

For every suspected critical function, ask:

• Which systems consume its output?
• Which receivers cannot substitute?
• Which systems share the same input?
• Which dependencies are physical?
• Which are informational?
• Which are geographic?
• Which are institutional or legal?
• Which depend on timing or synchronisation?
• Which backup routes secretly share the same vulnerability?
• Which dependent systems will overload if this one fails?

A power failure may stop pumping. Pumping failure may stop water distribution. Water loss may damage sanitation and health. Health loss may reduce the workforce. Workforce loss may slow repair.

That last loop matters.

The most dangerous dependency is often the one that damages the system’s ability to repair the original failure.

The propagation map

If a threshold is breached, do not jump directly to “collapse.” Map how the loss can travel.

Propagation mode	What happens	Diagnostic clue
Serial	A loses output, so B loses input, then C loses B.	Clear dependency chain.
Branching	One failure disables many dependent functions.	One node feeds several systems.
Convergent overload	Demand shifts onto remaining systems until they fail too.	Surviving services become saturated.
Common-mode	Apparently independent systems share one hidden vulnerability.	Backups fail together.
Geographic	Co-located systems are struck by the same event.	Several infrastructures fail in one corridor.
Institutional	One rule, authority or funding failure affects many services.	Different systems deteriorate after the same policy or budget shock.
Information	Bad or missing signals cause multiple wrong actions.	The physical assets exist but decisions diverge from reality.
Temporal	Current service is maintained by consuming future capability.	Deferred maintenance, depleted reserves, missing succession.
Behavioural	Loss changes behaviour in ways that worsen the loss.	Runs, hoarding, withdrawal, flight, panic buying, non-compliance.
Adversarial	An intelligent actor exploits the damaged system or recovery route.	The response itself changes the opponent’s strategy.

Propagation does not prove civilisation collapse. It tells us whether a local breach can become a wider one.

The repairability test

A large failure can remain bounded when the repair system survives.

A smaller failure can become civilisational when the machinery of recovery is missing.

Repairability should be decomposed rather than treated as one reassuring word.

Repair component	Diagnostic question
Observability	Can we see the real state quickly enough?
Interpretation	Can competent people distinguish symptom from mechanism?
Authority	Is someone legitimately able to decide and act?
Resources	Are money, parts, materials, energy, people and time available?
Access	Can repair crews and resources physically reach the failed function?
Coordination	Can multiple owners align timing and priorities?
Knowledge	Do maintainers know how the system actually works?
Fallback	Can the function operate safely in a degraded mode?
Trust	Will receivers and operators cooperate with emergency measures?
Optionality	Are alternative routes genuinely independent?
Retest	Can we tell whether the repair restored the function?
Succession	Can the repair capability survive current operators?

The old article’s insight that some systems “know what is wrong but cannot act on it” therefore becomes more precise:

OBSERVE ≠ UNDERSTAND
UNDERSTAND ≠ AUTHORISE
AUTHORISE ≠ RESOURCE
RESOURCE ≠ EXECUTE
EXECUTE ≠ DELIVER
DELIVER ≠ RECEIVER BENEFIT
RECEIVER BENEFIT ≠ DURABLE RECOVERY

Every broken handoff can become the binding repair constraint.

Damage speed versus repair speed

A useful orientation from the older CivilisationOS work should be preserved: compare how quickly damage compounds with how quickly the system can detect, decide, mobilise, repair and adapt.

IF DAMAGE + DRIFT GROWS FASTER THAN
DETECTION + CONTAINMENT + REPAIR + ADAPTATION,
THE RECOVERY CORRIDOR NARROWS.

This is not a universal physical equation. It is a diagnostic comparison.

Repair can lose the race because:

• the damage accelerates;
• signals arrive late;
• bad news is suppressed;
• authority is fragmented;
• parts are unavailable;
• maintainers are exhausted;
• the public no longer trusts instructions;
• the repair consumes a resource needed elsewhere;
• the original system is also required to repair itself;
• the environment has changed enough that restoration is no longer viable.

The diagnostic question is not merely “can it be fixed?” It is:

Can sufficient function be restored before propagation removes the capabilities required for restoration?

The CivilisationOS diagnostic states

A diagnosis should classify the best-supported current state rather than reaching immediately for “collapse.”

State	Meaning	Primary question
Normal variation	The system is moving within expected variation and margins remain healthy.	Is anything actually wrong?
Stress	Load has increased or inputs weakened, but buffers preserve function.	How much margin is being consumed?
Degradation	Quality, reach, reliability or margin is deteriorating over time.	Is repair keeping pace?
Regression	The actual trajectory is moving away from the reference across important deltas.	Which gap is widening and why?
Threshold approach	A load-bearing function is near the boundary of viability.	How much recoverable margin remains?
Threshold breach	A required capability has fallen below the level needed for its function.	Can it be restored before propagation?
Floor failure	The breached capability can no longer carry important dependent functions.	Which systems are losing support?
Cascade	Failure is propagating through dependencies, overload, feedback or common vulnerability.	Can the cascade be isolated?
Recovery failure	The route back is being lost because maintainers, knowledge, authority, trust, materials or access are gone.	What repair machinery remains?
Continuity failure	The society cannot reproduce enough of the operating configuration to remain substantially itself.	What can still be inherited or reconstituted?
Insufficient evidence	The current information does not justify a stronger classification.	What evidence would discriminate between the plausible states?

“Insufficient evidence” is a valid CivilisationOS output.

A runtime that cannot say “we cannot yet tell” will eventually manufacture certainty.

Diagnosis needs confidence, not only a label

Two analysts can both say “threshold approach” while possessing very different evidence.

A useful diagnosis should state confidence explicitly.

Confidence	Typical condition
Low	Sparse, stale, indirect or contradictory evidence; threshold poorly defined.
Moderate	Several relevant indicators agree, but mechanism or receiver coverage remains incomplete.
High	Independent evidence converges, mechanism is plausible and tested, receiver effects are observed, and competing explanations have been considered.

The confidence should be local to the claim.

We may have high confidence that electricity service failed, moderate confidence about the initiating fault and low confidence about whether the event reveals a structural civilisational vulnerability.

What would change the diagnosis?

Every serious diagnosis should include a discriminator.

If we think a system is merely stressed rather than degraded, what observation would prove that wrong?

If we think a threshold has been breached, what evidence would show that the function remains viable?

If we think repair capacity is intact, what stress test would demonstrate that claim?

A diagnosis that cannot name evidence capable of changing it is drifting towards ideology.

The hidden state problem

Civilisations contain important states that cannot be read directly from one metric.

Trust is not one observable. Institutional competence is not one observable. Resilience is not one observable. Maintenance debt is often distributed across assets. Knowledge loss can remain hidden until the experienced person leaves.

This means diagnostic humility must increase when the target variable is latent.

CivilisationOS should triangulate hidden states through:

• multiple indicators;
• different measurement methods;
• receiver reports;
• stress performance;
• historical comparison;
• counterfactual tests;
• independent evidence channels;
• changes after intervention.

A stable dashboard can coexist with a deteriorating hidden state if the chosen metrics are insensitive, delayed or gamed.

The centre can be wrong because the edge sees first

Many failures become visible first at the edge.

The person with the least buffer crosses the threshold first.

The remote community loses service before the national average changes.

The disabled commuter finds the broken route before the able-bodied traveller notices.

The junior worker sees a training gap before the organisation’s competence indicator falls.

The maintainer sees the workaround before management sees the reliability problem.

The diagnostic must therefore ask not only “what does the centre measure?” but:

Who would detect this floor first if it were giving way—and can their evidence travel?

The pre-breach deception: systems can look healthy by consuming themselves

A civilisation can preserve visible output by consuming hidden reserves.

• Staff work unpaid overtime.
• Maintenance is deferred.
• Inventory buffers are reduced.
• Debt replaces current revenue.
• Experienced workers carry undocumented knowledge.
• Families absorb institutional care failures.
• Teachers compensate for weak systems through personal labour.
• Ecological resources are consumed faster than regeneration.
• Bad cases are excluded from performance measures.
• Future investment is cancelled to preserve current output.

The system may appear stable precisely because it is eating the capacity that would be needed later.

Current output can be financed by future fragility.

That is why maintenance, staffing, reserves, succession and ecological condition belong inside diagnosis even when the headline service is still functioning.

False positives: when civilisation is diagnosed as failing but is not

Overdiagnosis can be harmful too.

Shock mistaken for structural decline

An earthquake, storm, war or pandemic produces severe loss, but the wider capability, repair system and continuity remain intact.

Political change mistaken for civilisation collapse

A dynasty, government or state structure falls while culture, institutions, people and capabilities continue or reorganise.

Reference change mistaken for regression

Standards improve, making the measured gap larger even though actual capability has not worsened.

Transition cost mistaken for failure

A system temporarily performs worse while replacing unsafe infrastructure, retraining people or changing technology.

Local failure inflated to total failure

One region, institution or population experiences serious failure while the larger system contains it.

Narrative selection

Only negative evidence is collected because the diagnosis was selected before the investigation began.

False negatives: when civilisation is failing but the dashboard says it is fine

Underdiagnosis can be more dangerous because it consumes repair time.

Average masks the tail

Most receivers remain above threshold while a vulnerable group has already fallen below it.

Stock masks flow failure

The asset still exists, but maintenance, parts, staffing or funding required to keep it functional has collapsed.

Formal institution masks capability loss

The office, law or organisation remains while real competence and enforcement disappear.

Temporary workaround masks structural failure

Heroic staff, emergency funding or informal networks keep output alive while the formal system loses viability.

Suppressed bad news

The data remain positive because reporting failure is punished or inconvenient.

Delayed consequence

The cost appears years later in infrastructure, ecology, health, finance or succession.

The diagnostic sequence

1. NAME THE OBSERVABLE CONDITION
2. RESOLVE THE OBJECT AND RFE
3. SET SCALE AND TIME WINDOW
4. NAME THE RECEIVERS
5. SELECT THE REFERENCE AND ACTUAL MEASURES
6. CALCULATE / DESCRIBE THE DELTA VECTOR
7. SEPARATE FACT, MEASUREMENT, REPORT, INFERENCE, MODEL AND UNKNOWN
8. LOCATE SCALE × TIME × CAPABILITY × REALITY LAYER
9. IDENTIFY THE LOAD-BEARING FUNCTION
10. MAP DEPENDENCIES AND ALTERNATIVES
11. ESTIMATE VIABILITY MARGIN
12. TEST WHETHER THE THRESHOLD IS APPROACHED OR BREACHED
13. MAP PROPAGATION PATHS
14. TEST OBSERVABILITY, AUTHORITY, RESOURCES AND REPAIRABILITY
15. CHECK WHO FALLS THROUGH FIRST
16. CLASSIFY THE CURRENT STATE
17. STATE CONFIDENCE
18. NAME THE DISCRIMINATING EVIDENCE THAT WOULD CHANGE THE CLASSIFICATION
19. DEFINE THE NEXT SAFE ACTION OR INFORMATION REQUEST
20. OBSERVE WORLD RETURN AND UPDATE

The Diagnostic Gate

SERIOUS CONDITION OBSERVED?
        ↓
OBJECT / FUNCTION / RECEIVER DEFINED?
   NO → HOLD: DEFINE THE OBJECT
   YES
        ↓
REFERENCE AND ACTUAL COMPARABLE?
   NO → HOLD: REPAIR COMPARISON
   YES
        ↓
IMPORTANT DELTA PRESENT?
   NO → NORMAL VARIATION / NO DIAGNOSIS
   YES
        ↓
LOAD-BEARING FUNCTION INVOLVED?
   NO → LOCAL / NON-FOUNDATIONAL PROBLEM
   YES
        ↓
VIABILITY MARGIN SHRINKING?
   NO → IMPERFECT BUT VIABLE
   YES
        ↓
THRESHOLD CROSSED?
   NO → STRESS / DEGRADATION / THRESHOLD APPROACH
   YES
        ↓
DEPENDENT FUNCTIONS LOSING SUPPORT?
   NO → BOUNDED THRESHOLD BREACH
   YES
        ↓
PROPAGATION OUTRUNNING CONTAINMENT?
   NO → FLOOR FAILURE WITH RECOVERY PATH
   YES
        ↓
REPAIR CAPACITY PRESENT?
   YES → CASCADE / TRIAGE / REPAIR
   NO  → RECOVERY FAILURE
        ↓
CAN CORE CAPABILITY BE REPRODUCED?
   YES → TRANSFORMED / SUCCESSOR CONTINUITY
   NO  → CONTINUITY FAILURE

The gate does not replace judgement. It prevents judgement from skipping steps.

Case study: MRT disruption

A major rail disruption affects thousands of commuters.

Is this civilisation failure?

Start smaller.

• Observable condition: train service on defined lines is unavailable or degraded.
• Function: move people safely and reliably to useful destinations.
• Scale: affected lines, corridors, stations and receivers.
• Time: minutes, hours, repeated incidents or chronic degradation.
• Receiver: passengers, including those with fewer alternatives.
• Reference: safe, reliable, maintainable and recoverable urban mobility.
• Delta: travel time, accessibility, reliability, crowding, safety and recovery.

Then ask whether the disruption is bounded. Are buses, alternative lines and transport modes available? Can faults be diagnosed? Are parts and maintainers available? Does communication reach passengers? Can the system restore service?

A severe outage can remain a transport-system failure without becoming civilisational if the city can reroute, repair and continue.

It moves closer to a civilisational floor problem if repeated rail, road, power, communications and logistics failures begin eliminating mobility needed for food, health, work and repair across the city.

The size of the inconvenience is not the same as the depth of the floor breach.

Case study: education decline

Suppose examination results weaken for several years.

That is evidence. It is not yet a diagnosis of education collapse.

Check the reference and the full capability.

• Are learners understanding less?
• Is transfer weaker?
• Are results lower because the assessment became harder or more valid?
• Are particular groups falling while others improve?
• Are teacher capability and retention changing?
• Are schools using workarounds that consume staff sustainability?
• Can the system still train competent successors?
• Can bad instructional outcomes be reported without distortion?
• Are lower-layer factors such as health, attendance, language or family stress driving the signal?

Here the old DLT lens remains useful as a specialist instrument. Depth, Load and Transfer can help diagnose learner or capability formation. It should not be mistaken for the whole civilisation diagnosis.

The civilisational threshold would be much stronger: education and knowledge-transfer capability has degraded far enough that society can no longer reproduce enough competent people for critical roles, and the repair pipeline is itself failing.

Case study: a government falls

A government falls after political crisis.

That is a major event.

It is not automatically civilisation collapse.

Ask what continues.

• Do courts still function?
• Do utilities operate?
• Do records remain trustworthy?
• Do hospitals and schools continue?
• Can taxes or other public resources still be collected legitimately?
• Do security institutions obey a workable chain of authority?
• Can new political authority be constituted?
• Does the population still cooperate across major divisions?
• Are people fleeing because the wider social order can no longer protect ordinary life?

The political system may experience regime failure while the wider civilisation remains continuous. Conversely, a government may remain formally in office while real civilisational floors—truth, infrastructure, security, food, health, maintenance—deteriorate badly.

Case study: water under drought

Drought reduces available water.

The trigger is environmental. The failure diagnosis depends on the field.

• What storage remains?
• How fast is it depleting?
• Which uses are essential?
• Which receivers have alternatives?
• Can demand be reduced?
• Can other sources be activated?
• Does treatment capacity remain?
• Does energy remain available for pumping?
• Can low-income households still access safe water?
• What ecological damage follows additional extraction?

A drought can remain a severe stress while the water floor holds. A threshold breach begins when required safe-water service cannot be sustained for the defined receivers. Floor failure deepens if sanitation, health, food production or firefighting begin losing support.

Case study: information and shared reality

Suppose public disagreement becomes intense and misinformation spreads.

“Truth collapse” is too strong as an opening diagnosis.

Ask what function is impaired.

• Can evidence still be collected?
• Can independent sources compare claims?
• Can corrections travel?
• Can institutions revise decisions?
• Can people distinguish event, claim, inference and opinion?
• Do factions share enough common reference points to coordinate?
• Are false beliefs producing material failures in health, security, finance or governance?
• Are people punished for reporting inconvenient facts?

The shared-reality floor becomes civilisational when enough coordination depends on reliable signals that widespread corruption of those signals materially damages other load-bearing systems.

Case study: a company or institution that looks healthy

Revenue is rising. Projects are shipping. Leadership reports success.

But experienced staff are leaving, maintenance work is deferred, difficult customers are excluded from metrics, and one senior person carries the architecture in their head.

The actual output is strong. The hidden capability state may be degrading.

This is the diagnostic value of separating:

OUTPUT STOCK
from
REGENERATIVE CAPABILITY

The institution has not failed yet. But the succession, maintenance and repair margins may be narrowing.

Historical diagnosis: never ask only “why did it collapse?”

Historical cases require even more caution because the evidence is incomplete and the word civilisation often bundles states, cultures, populations and institutions together.

Ask instead:

• What exactly stopped functioning?
• What geographic region is meant?
• Which population experienced the loss?
• Over what interval?
• Which institutions ended?
• Which capabilities declined?
• Which people migrated?
• Which languages, religions, technologies, legal concepts or practices continued?
• Which successor systems inherited earlier capability?
• Which parts of the apparent collapse are archaeological visibility changes rather than human disappearance?

The article How Civilisations Persist, Transform and Inherit owns that continuity question.

A good diagnosis does not automatically tell you the repair

The old page moved directly from diagnosis to a selected recovery mode.

That can be too fast.

Several repairs may be compatible with the same diagnosis. The best move depends on authority, evidence, timing, cost, receiver needs, dependencies and reversibility.

Diagnosis should therefore output an admissible repair problem, not pretend to solve the whole political or technical choice automatically.

For example:

The evidence supports a high-confidence diagnosis that the maintenance floor for this service is degrading and approaching a staffing threshold. The immediate RFE is to preserve safe function and succession. Any proposed repair must restore staffing and maintenance margin without hiding workload in another group, reducing safety, or eliminating future alternatives.

That output can then be sent through Rules of Advancing a Civilisation before action.

The smallest discriminating question

When evidence is incomplete, the next best move may be a question rather than an intervention.

A strong diagnostic asks:

What single piece of information would most strongly separate the leading explanations?

Examples:

• If falling output is caused by demand, do fault rates remain stable?
• If the problem is maintenance debt, does reliability worsen with asset age and backlog?
• If reported access is real, can receiver-level records confirm actual receipt?
• If a backup is independent, can it operate when the primary dependency is removed?
• If capability transfer is healthy, can a successor perform without the expert present?

This reduces the risk of taking a large action when a small information gain would change the diagnosis.

World Return closes the diagnostic loop

A diagnosis is provisional until its predictions and repairs return evidence from the world.

If maintenance was the binding problem and maintenance capacity is restored, reliability should improve in the predicted places and time window.

If the receiver problem was last-mile access, changing the final handoff should change actual receipt.

If trust was the key coordination constraint, a repair should change observable cooperation rather than merely public messaging.

If the expected change does not appear, the diagnosis must be updated.

A failed repair is evidence about the diagnosis.

The Civilisation Failure Diagnostic Report

A reader-safe diagnostic output can fit into twelve fields.

Field	Output
1. Observable condition	What has actually happened?
2. Object and RFE	Which function is being diagnosed?
3. Scale and time	Where and over what interval?
4. Receivers	Who is affected and who is missing?
5. Reference / actual / delta	What should be happening, what is happening and how far apart are they?
6. Evidence ledger	Facts, measurements, reports, inferences, models and unknowns.
7. Capability and layer	What ability and reality layer are implicated?
8. Floor and threshold	Is the function load-bearing, and is viability threatened or breached?
9. Propagation	How can the loss spread?
10. Repairability	Can observability, authority, resources, access and knowledge restore function?
11. Classification + confidence	Stress, degradation, breach, cascade, recovery failure, etc., with confidence.
12. Discriminator + World Return	What evidence changes the diagnosis, and what result proves repair?

The 30-question Civilisation Failure Audit

1. What exactly happened?
2. What is observed and what is inferred?
3. Which object is being diagnosed?
4. What is that object supposed to do?
5. At what scale is the claim true?
6. Over what time window?
7. Who are the receivers?
8. Who is missing from the evidence?
9. What is the Perfect Reference?
10. What is the Best Feasible Path?
11. What is the actual trajectory?
12. Which deltas matter?
13. Are those deltas narrowing or widening?
14. Which capability is impaired?
15. Which reality layer carries the constraint?
16. Is the function load-bearing?
17. Which systems depend on it?
18. What alternatives exist?
19. Are the alternatives genuinely independent?
20. How much viability margin remains?
21. What evidence defines the threshold?
22. Has the threshold actually been crossed?
23. How can loss propagate?
24. Who falls through first?
25. Can bad news reach authority?
26. Can authority mobilise resources in time?
27. Can maintainers, parts, knowledge and access restore the function?
28. What is the best-supported classification and confidence?
29. What evidence would change that classification?
30. What World Return will prove that recovery is real?

What this article refuses to do

• It does not diagnose every civilisation through one immutable set of axes.
• It does not assume all failures follow one universal sequence.
• It does not call every crisis collapse.
• It does not equate government failure with civilisation failure.
• It does not convert latent social states into fake precise numbers merely for a dashboard.
• It does not treat one average as the whole receiver population.
• It does not assume the largest delta is the most dangerous.
• It does not treat external shocks as proof of incompetence.
• It does not treat good outcomes as proof of good decisions.
• It does not treat poor outcomes as proof that the chosen decision was irrational.
• It does not prescribe a recovery before authority, evidence, dependencies and receivers are understood.
• It does not hide uncertainty.
• It does not allow the model to win an argument against World Return.

The purpose of diagnosis is not to sound certain. It is to make the next useful distinction before the remaining repair window closes.

Frequently Asked Questions

How do you diagnose a civilisation collapse?

Begin with the specific observable failure rather than the word collapse. Define the function, scale, time and receivers; compare reference and actual trajectories; locate the important delta; test the viability threshold of any load-bearing capability; map propagation; and assess whether repair and continuity remain possible.

What is the earliest sign of civilisation failure?

There is no single universal earliest sign. Useful warnings include shrinking margins, slower recovery, rising maintenance debt, permanent workarounds, thinning skill pipelines, hidden receiver failures, dependency concentration and blocked bad news. The mechanism and relevant floor must still be identified.

Is decline the same as collapse?

No. Decline or regression means important deltas are worsening. Collapse is a stronger claim involving major floor, recovery or continuity failure.

Can a civilisation be failing even when GDP is rising?

Yes, if growth coexists with worsening load-bearing deltas such as ecological viability, maintenance, receiver access, institutional truth, repair capability or succession. GDP is one outcome measure, not the whole civilisation state.

Can civilisation be healthy even if it is far from perfect?

Yes. Distance from the Perfect Reference measures unrealised possibility. Failure risk depends on viability margins, thresholds, propagation and repairability. A civilisation can have large improvement opportunities and still be stable, learning and resilient.

Why not use one civilisation health score?

Because civilisation is multidimensional, distributions matter, some dimensions are not commensurable, and a high score can hide a breached non-compensable floor. Dashboards can be useful, but the underlying vector and hard thresholds should remain visible.

What if there is not enough evidence?

Return “insufficient evidence,” preserve the leading hypotheses, and ask for the most discriminating missing information. Uncertainty is a diagnostic state, not a defect to be hidden.

When does the Failures of Civilisation branch take over?

When a necessary capability has crossed its viability threshold, the floor is no longer carrying dependent functions, propagation has begun, or repair capacity itself is being lost. The Failures branch then studies breach, cascade, recovery loss and continuity failure in depth.

Teaching and discussion guide

1. Take a dramatic headline claiming “collapse” and rewrite it as a bounded diagnostic statement.
2. Separate observed fact, measurement, claim, inference and model in one case.
3. Choose a service and locate it on Scale × Time × Capability × Reality Layer.
4. Identify the receiver that would detect failure first.
5. Draw the reference, actual state, aspiration gap and viability margin separately.
6. Name one target that can be missed without a threshold being breached.
7. Name one threshold that could be crossed while the average still looks healthy.
8. Map three dependencies and test whether the backups are truly independent.
9. Distinguish stress, degradation, regression and threshold approach.
10. Design a discriminator that separates two competing diagnoses.
11. Find one example where current output is being maintained by consuming future capability.
12. Test repairability across observability, authority, resources, access, knowledge and retest.
13. Write a diagnosis with explicit confidence rather than absolute certainty.
14. Define the World Return that would show the diagnosis was wrong.

The shortest useful summary

• Do not diagnose civilisation before diagnosing the observable failure.
• A large delta from perfection is not automatically a failure threshold.
• Define the object, function, scale, time and receiver first.
• Separate observation, measurement, receiver report, claim, inference, model and unknown.
• Locate the problem on Scale × Time × Capability × Reality Layer.
• Ask whether the impaired capability is load-bearing.
• Measure or estimate the viability margin before calling a breach.
• Threshold approach and threshold breach are different states.
• Map propagation before calling a local failure civilisational.
• Repairability depends on observability, authority, resources, access, knowledge, coordination, trust and retest.
• Use state + confidence, not dramatic labels alone.
• “Insufficient evidence” is a valid output.
• A diagnosis must name evidence that could change it.
• World Return can confirm, weaken or overturn the diagnosis.

The real diagnostic discipline

Civilisation does not become more diagnosable by giving every problem a grander name.

It becomes more diagnosable when the object is bounded, the reference is explicit, the actual world is observed, the delta is visible, the receiver is heard, the threshold is defensible, propagation is mapped, repairability is tested and uncertainty is allowed to remain.

The purpose of a civilisation diagnosis is not to win the argument that collapse is happening. It is to know what is failing early enough, accurately enough and modestly enough that civilisation still has a chance to repair it.

OBSERVE.
BOUND THE OBJECT.
COMPARE.
LOCATE.
TEST THE FLOOR.
MAP THE CASCADE.
CHECK THE REPAIR PATH.
STATE WHAT WE KNOW.
STATE WHAT WE DO NOT.
RETURN TO REALITY.