The Myth of Root Cause: Why Every Investigation Must Eventually Stop Asking “Why?”

By Florin Necula, PhD | Accountable Manager & Licensed Aircraft Engineer

Jul 29^th, 2026

Every occurrence has a history. That does not mean it has one final, fully discoverable cause.

1. There Must Be a Cause

Following an aviation occurrence, one question inevitably arises:

What was the root cause?

The question appears objective and scientific. Yet it already contains an assumption: somewhere beneath the visible sequence of events, there is a fundamental cause waiting to be discovered. If investigators examine the evidence carefully enough and continue asking “why?”, they should eventually reach the bottom.

This expectation is deeply embedded in aviation. A flight crew continues an unstable approach. An air traffic controller issues a conflicting clearance. A technician selects an incorrect configuration. A ground-handling team damages an aircraft during towing. An airport inspection fails to identify a hazardous runway condition.

In each case, we expect the investigation to tell us what caused the event. We are rarely satisfied with the conclusion that several conditions interacted in a way that had not been anticipated. We want a clear explanation—and preferably a single one.

This desire has practical value. Aviation cannot improve without searching for causes. But it also reflects a deeper human need for certainty, order and closure.

The philosopher Gottfried Wilhelm Leibniz expressed this expectation through the Principle of Sufficient Reason: broadly, nothing happens without a sufficient reason why it is so and not otherwise. Every fact or event must have a reason, cause or ground that explains it. Importantly, Leibniz also recognised that such reasons may not always be known to us. The existence of a reason does not guarantee our ability to discover it.

This distinction is central to aviation investigation:

An occurrence may have a causal history without having one singular, final and fully discoverable root cause.

The myth is not that causes exist, nor that root-cause analysis has no value. The myth is that every complex occurrence contains one objectively identifiable root which, once found, completely explains the event.

2. From Causal Chains to Causal Networks

At one level, every aviation occurrence is a sequence of physical events. An action produces a change, the system responds and an outcome follows.

An incorrect input changes a system configuration. A clearance alters an aircraft’s trajectory. A component fails to perform its intended function. A vehicle moves into an area without sufficient clearance.

This causal-mechanistic perspective is necessary. Investigators must establish what happened, in what order and through which physical or operational mechanisms.

However, knowing the sequence does not necessarily explain the occurrence.

Consider a simplified chain:

Incorrect input → incorrect system configuration → unexpected aircraft response.

The sequence may be factually correct. But it does not explain why the input was made, why it appeared correct, why it was accepted, why it was not detected or why the system allowed one action to develop into a hazardous condition.

Aviation is not merely a collection of components acting upon one another. It is a complex sociotechnical system involving aircraft, people, technology, procedures, organisations, commercial relationships, regulators and continuously changing operational environments.

The traditional causal chain: Human error → unsafe condition → occurrence

—may therefore be only the visible surface of a much wider network involving operational conditions, incomplete information, competing goals, organisational expectations, local decisions, ineffective barriers and particular timing.

Many of these conditions are also present during successful operations. Pilots regularly manage changing weather, operational pressure and imperfect information. Controllers handle traffic complexity and uncertainty. Engineers adapt to aircraft availability and unexpected findings. Ground personnel coordinate movements in restricted spaces and changing conditions.

Usually, these adaptations make aviation work. Occasionally, a particular combination of otherwise manageable conditions produces a very different outcome.

If a condition is present hundreds of times during successful operations but contributes to one accident, can it reasonably be called the root cause? If several conditions were each necessary, which one deserves to be called the root?

The closer we look, the less the event resembles a chain and the more it resembles a network.

3. The Limits of What We Can Know

The occurrence is a historical reality. The investigation is a reconstruction of that reality.

Investigators do not return to the event itself. They work with what remains afterward: recorded data, physical evidence, procedures, operational records, witness recollections and traces of the original context.

Some information was never recorded. Some evidence disappears or changes. Informal practices may not appear in official documentation. People describe their decisions after they know how the event ended, and that knowledge inevitably changes how the preceding actions appear.

An investigation report is therefore not the occurrence itself. It is a model of the occurrence.

To build that model, investigators must decide which facts are relevant, which relationships are causal, where the event began, how far the analysis should extend and what level of evidence is sufficient. Different competent investigators may emphasise different factors and produce different, but still defensible, explanations.

David Hume questioned how confidently human beings can speak about causation. We observe that one event follows another, recognise patterns and infer causal relationships. But the necessary connection between cause and effect is not something we directly observe. Our conclusions about causation are interpretations constructed from evidence and experience.

Kurt Gödel’s incompleteness theorems offer another, although more distant, perspective on the limits of proof. Gödel demonstrated that sufficiently powerful and consistent formal mathematical systems contain statements that cannot be proved or disproved using only the rules of those systems.

His theorems do not apply directly to aviation investigation. They do not prove that accidents cannot be fully explained. Nevertheless, they provide a useful warning against the ambition of complete and self-contained explanation. If even rigorous formal systems encounter inherent limits, an investigation of a unique event—based on incomplete evidence and reconstructed human experience—should be particularly cautious about claiming final certainty.

Investigators also rely on counterfactual conclusions:

If this action had not occurred, the accident would have been prevented.

Such a conclusion may be reasonable and well supported, but the alternative history cannot be observed. We cannot reproduce the exact occurrence while changing only one condition. It remains an evidence-based inference, not mathematical proof.

A credible explanation is therefore not necessarily a complete explanation. An actionable conclusion is not necessarily the final truth.

Recognising these limits does not make investigation futile. It makes investigation intellectually honest.

4. The World as It Was—and as It Appeared

There are at least two realities in every occurrence:

1. The situation as it can be reconstructed after the event.
2. The situation as it appeared to the people while it was unfolding.

Aviation occurrences happen in the physical world, but human decisions are made in the perceived world.

After an event, investigators know the outcome. They know which warning was significant, which decision proved incorrect, which information should have been questioned and which risk eventually materialised.

The people involved had none of these advantages.

A pilot does not respond to every fact about the aircraft, weather and surrounding airspace. The pilot responds to a mental representation constructed from the information that is visible, expected and considered relevant at that moment.

The same applies to an air traffic controller, technician, dispatcher, ground operations, cabin crew member, airport employee or operational manager.

Their actions are influenced by the information available, previous experience, expectations, immediate goals, workload, fatigue, time pressure, communications from others and practices considered normal in that environment.

This is the principle of local rationality: people generally act in ways that make sense to them, given their goals, knowledge, attentional focus and understanding of the situation at that particular time.

A flight crew may continue an approach because the deviations still appear recoverable. A controller may maintain an initial interpretation of the traffic situation because subsequent information appears to support it. A technician may select a configuration believed to be applicable because of its labelling and the system default. A ground handler may continue towing because the available signals indicate adequate clearance.

From outside the event, and with knowledge of the outcome, the decision may appear obviously wrong. From inside the situation, it may have appeared normal, reasonable or even necessary.

This is why conclusions such as “incorrect perception,” “loss of situational awareness,” “lack of attention” or “poor judgment” are not satisfactory root causes.

If a person incorrectly perceived that a runway was clear, a component was correctly installed, an aircraft was at the expected level or an approach remained stable, the investigation has not yet explained the event. It has identified another phenomenon requiring explanation:

• What cues shaped that perception?
• Which information was absent, weak or ambiguous?
• What did the person expect to see or hear?
• Did previous experience reinforce that expectation?
• What competing demands directed attention elsewhere?
• Why did no other person or system correct that understanding?

Perception is not a passive recording of the external world. People interpret information through their experience, expectations and current goals. They act not upon reality in its entirety, but upon the reality they have constructed from the information available to them.

The investigative question should therefore not be:

Why did the person make such an obvious mistake?

It should be:

What did the person understand the situation to be and why did that understanding make sense at the time?

5. The Root Is Where the Investigation Stops

The cause identified by an investigation often depends on where the investigator draws the system boundary.

If the boundary surrounds the individual, the conclusion may be that the person made an incorrect decision. If it includes cognitive conditions, expectation, attention or an inaccurate mental model may become relevant. If it includes the task, ambiguity, interruption or competing demands may become visible.

Expand the boundary further and the explanation may involve team communication, interface design, organisational planning, training, resources, commercial arrangements or regulatory assumptions.

None of these explanations is necessarily false. Each describes the event at a different level.

This leads to an uncomfortable conclusion:

The root cause is often not the deepest cause. It is the point in the causal history at which the investigation decides to stop.

That stopping point may be influenced by the investigation mandate, available evidence, legal responsibility, organisational authority, time, resources and the need to recommend actions that someone can implement.

“Human error” becomes an attractive stopping point because it is close to the outcome, visible and relatively easy to document. But “failure to follow procedure,” “poor judgment” and “loss of situational awareness” may describe what happened without explaining why it happened.

The fifth “why” in a Five Whys analysis is not necessarily the root cause. It is simply the fifth answer. The questioning could continue into task design, organisational history, management decisions, regulatory expectations and industry structure almost indefinitely.

Eventually, every investigation must stop asking why. The critical issue is whether it stops because the system has been understood sufficiently to support meaningful action—or simply because a convenient explanation has been found.

This does not mean that systemic explanations should eliminate personal accountability. Understanding why an action made sense does not automatically make it acceptable.

An investigation may still determine that the information was clear, the risk was understood, adequate resources were available and the person possessed the necessary competence. It may establish that a conscious and unjustified deviation occurred or that conduct fell below reasonable professional expectations.

A Just Culture approach should therefore keep two questions separate:

1. Was the individual’s conduct reasonable, considering what they knew and perceived at the time?
2. Why was the aviation system vulnerable to that action and why did its barriers not detect or contain the consequences?

Local rationality is not a declaration that every action was reasonable. It is a way of understanding the action before judging it.

6. From Finding Causes to Learning

Root-cause analysis should not be abandoned. It remains valuable when a failure is relatively linear, the evidence is strong and a technical or procedural weakness can be clearly demonstrated.

The problem is not the term “root cause.” The problem is the confidence with which one selected factor is sometimes presented as the only possible explanation.

If the purpose of investigation is learning rather than explanatory closure, we must move beyond asking only, “What was the root cause?” We should also ask:

• How did the situation develop over time?
• What combination of conditions made the occurrence possible?
• What did the people involved believe was happening?
• Why did their actions make sense at that moment?
• Which barriers should have prevented or contained the event?
• Why were those barriers absent or ineffective?
• Where else might similar conditions exist?
• What is supported by evidence, what is reasonably inferred and what remains uncertain?

Aviation investigates occurrences because it believes they can be understood. It investigates responsibly when it also recognises the limits of that understanding.

Every occurrence has a history, but that history rarely resembles a single chain leading back to one root. It passes through physical events, human perceptions, local decisions, organisational conditions and wider system influences.

Recognising the limits of our knowledge does not weaken aviation safety. It makes our conclusions more honest and our learning more meaningful. We do not need perfect knowledge before taking action, but we should distinguish carefully between what the evidence demonstrates, what we reasonably infer and what remains unknown.

The real danger is not that we occasionally fail to find the root cause.

It is that we stop learning because we believe we have already found it.