The First Step to Understanding the Problem

A field report recovered from the HALO archives

In the winter of 1961, long before anyone used the phrase human factors with confidence, a small interdisciplinary team gathered in a converted aircraft hangar on the edge of the Mojave. Engineers, flight specialists, physicians, and systems analysts worked under a single directive: understand why highly trained people failed at critical moments.

The program would later be called HALO, though the name mattered less than the problem it pursued. The team quickly discovered that their object of study was something unexpected:

They were not studying incompetence. They were studying surprise.

The Early Failure Pattern

While the earliest HALO case reviews covered challenging high risk scenarios. What emerged from the initial studies was not failure in the face of obvious, visible, and inevitable disaster. They documented something more unsettling.

Experienced operators failed after the danger had already arrived.

In transcripts and debriefings, the same phrases appeared again and again:

• “I didn’t realize how bad it was.”
• “I thought we still had time.”
• “It didn’t look dangerous yet.”

The threat was present. The signal existed. But the operator’s mind had not crossed the internal boundary that says, this is now a different problem. Failure occurred in the gap between reality and recognition. Historians would later identify this as one of the seminal insights of the HALO project.

Later it would come to be called the recognition gap.

Early Mistakes: Observation is Not Recognition

One of the program’s earliest misconceptions was the belief that overcoming the recognition gap depended on seeing more information. Engineers built better dashboards. Clinicians added monitors. Team leaders were given increasingly precise instrumentation.

Outcomes did not improve.

Researchers would eventually come to understand that the failure was not perceptual. It was interpretive. Operators observed the signal but had not yet understood its meaning.

It was not until a systems analyst, recently returned from studying aviation mishaps, had an insight capture for all time in a note left written in the margins of one of his notebooks that was eventually discovered in the HALO archives:

“They see it. They just don’t know what it means yet.”

That marginal note is now widely regarded as the key insight shaping everything that followed. Much of the modern understanding of threat recognition can be traced back to it.

The Retrospectoscope

With this new insight, researchers and specialists went back and reviewed all the airway disasters classified as HALO events. New patterns began to emerge:

• Oxygen saturation drifting from 98 to 92 was missed
• Mask ventilation requiring extra hands, not noticed
• Small amounts of blood pooling in the hypopharynx, ignored
• Laryngoscopy becoming “harder than expected” not acknowledged

In the original studies, none of these signals, in isolation, triggered panic. But together, over time, they formed a trajectory toward loss of control that was not recognized.

The danger was not missed. It was misclassified.

The Quiet Interval

From this new understanding of the problem came one of the most important concepts to emerge from the HALO project;s work: what the team came to call the “quiet interval.”

This was the period after an airway began to degrade but before anyone treated the situation as genuinely unsafe. During this interval, options still existed, help could still be summoned, and alternative strategies remained viable.

In airway case reviews, this interval was almost always filled with reassuring internal narratives:

• This is still manageable.
• I’ve seen worse.
• Just one more adjustment.

None of these statements was irrational. All of them delayed recognition.

How It Was Taught

As clarity around the first domain increased, additional insights followed. The historical record makes clear that more technology, more detailed flowcharts, and newer algorithms did little to improve threat recognition on their own. The recovered training manuals from the period indicate that the solution was a distinctly human one. Threat was increasingly understood as a series of recurring cognitive traps, and the researchers sought to pair them with explicit countermeasures. These became known as trap–interrupt dyads, and they formed the foundation of threat-recognition training. 

HALO instructors staged scenarios that felt almost manageable. The goal was not success, but recognition. Learners were debriefed and asked a single question:

“At what moment did this become unsafe?”

Most answers came too late. That delay was the lesson.

Trainees were then tasked with identifying the trap and naming the appropriate interruption. Over time, testing of this model yielded a consistent result. With repeated exposure, airway specialists could be taught to recognize threat well before deterioration became obvious to everyone in the room.

That earlier recognition preserved time, oxygen, and options, allowing teams to choose responses that were more deliberate, more coordinated, and more effective.

This approach was novel and, at the time, controversial. In fact, it would not be until the early 2000s that this type of learning built for humans became more widely accepted. 

It is also notable that HALO instructors insisted that recognition be spoken aloud. Unspoken recognition, they observed, did not become shared recognition, and threats not shared were rarely managed in time.

A Fond Farewell

These early successes created a sense of euphoria within the team. But shortly after the first domain was formalized, the analyst most closely associated with its discovery died unexpectedly in a late-night car crash while driving home from the Mission Center. In his memory, friends gathered in a moving ceremony outside the center to unveil his most frequently cited observation that had been etched above the center’s entrance:

“Once the threat is obvious, the work of recognition is already over.”

Today, it is widely accepted that threat recognition does not solve the problem. It preserves the possibility that the problem can still be solved. In the early 1960s, this idea was revolutionary.

An Uncertain Future

With the completion of the first domain, the HALO Mission Center struggled to define its purpose. Having reshaped how airway specialists approached their craft, the urgency that once sustained the program faded. Scientists began to leave. Funding followed.

One researcher who remained until the Center’s closure in 1973 summarized the situation at the time with characteristic restraint:

“We solved the problem of threat recognition, and failed to recognize how solving it would become a threat to us.”

As it turned out, this was only the first chapter in the HALO program’s longer and more complicated legacy.