Read Chapter 11: MARGARET HALE

Margaret Hale believed in evidence the way Luc believed in physical law.

She was forty-one, born in Pennsylvania, trained as an electrical engineer at a time when most firms treated women engineers as temporary mistakes. NASA had learned to use her because she was difficult to intimidate and impossible to charm.

“You are emotionally attached to the theory,” she told Luc during their third interview.

“I am emotionally attached to not killing astronauts.”

“That is not the same thing.”

“What theory?”

“That Vale’s group knowingly accepted dangerous hardware.”

“They falsified data.”

“Yes. But intent matters. Did they believe they were hiding harmless anomalies, or did they know the relay could fail in mission conditions?”

“The lot report says it could.”

“Did Vale read it?”

“He requested reclassification.”

“His office requested it. Find me proof he saw the rejection.”

Luc disliked the distinction because it was correct.

Hale pushed harder. “And Delaney’s death?”

“The pressure bottle had an inspection flag that was cleared.”

“That may be negligence. It does not prove retaliation.”

“I did not say it did.”

“You imply it every time you mention him.”

Luc was silent.

She leaned back. “You are useful because you notice inconsistencies. You become dangerous when you make them tell the story you want.”

The rebuke followed him for days.

He returned to the technical record and separated what he knew from what he suspected.

Known: MidSouth lot 62-B failed combined heat and vibration.

Known: Asterion’s qualification file contained a modified summary showing a pass.

Known: Galveston full-level vibration produced dropouts.

Known: a reduced-level rerun replaced the failed result in final reporting.

Known: thermal-vacuum testing later reproduced a dropout.

Known: Pike signed the rerun request.

Known: Vale’s office had created a test reconciliation process.

Suspected: Vale personally ordered data falsification.

Suspected: Delaney had intended to expose it.

Suspected: the home search was related.

Suspected: Delaney’s fatal accident involved more than schedule-driven negligence.

The distinction made Luc feel less certain and more effective.

In March, Hale received a break from an unexpected source.

Susan Bell asked to meet privately.

She brought a file of carbon copies she had kept at home.

“I signed waivers I should not have signed,” she said. “I told myself the engineers understood the risk. Then Frank died. Then the fire happened. I am done telling myself things.”

One memorandum carried Vale’s handwritten instruction:

DO NOT OPEN RELAY REDESIGN. USE VENDOR REQUALIFICATION. SCHEDULE IMPACT UNACCEPTABLE.

Below it was Bell’s note documenting that no approved vendor requalification existed.

Hale read the page, then looked at Luc.

“There is your intent.”

Luc felt no satisfaction.

“What happens now?”

“Now the lawyers arrive.”

Hale rebuilt the case from sources Luc had never seen.

She interviewed procurement clerks, contract administrators, laboratory schedulers and secretaries. One secretary remembered typing a memo twice because Vale disliked the phrase “failed qualification.” A scheduler recalled an instruction to label a rerun as “continuation” rather than “repeat” so it would not trigger automatic review. A clerk produced a routing slip showing that the MidSouth rejection report had entered Vale’s office three days before the reclassification request.

None of those facts proved the relay would fail in flight.

Together they proved knowledge had been managed.

Luc was impressed by the difference between engineering investigation and institutional investigation. Engineers traced energy, signals and material. Hale traced authority.

“Every system has a schematic,” she told him. “Organizations just pretend theirs is invisible.”

She drew one on a whiteboard: supplier quality to Asterion procurement, procurement to project quality, project quality to Vale’s reconciliation group, group decisions back to test and reporting.

“Here is your fault path,” she said.

Luc looked at the arrows.

“Where is Pike?”

“Downstream.”

“Responsible?”

“Yes.”

“Primary cause?”

“Probably not.”

The distinction reminded Luc of accident analysis. The component that finally opened a circuit was not always the root cause. Sometimes the root cause was a decision made months earlier that placed the component there.

Vale had not touched the relay.

He had touched the conditions under which everyone else evaluated it.

The lawyers' arrival created an unexpected technical problem: people stopped writing candidly.

Meeting minutes became vague. Engineers who once wrote “relay opens under vibration” began writing “continuity behavior under review.” Every sentence was shaped by fear of discovery requests and headlines.

Luc complained to Hale.

“We are investigating falsified clarity by producing legal ambiguity.”

She nodded. “Common disease.”

“How do we stop it?”

“We separate factual logs from legal interpretation. Make engineers record observations in plain language before counsel edits anything.”

Together they drafted a temporary procedure. Test operators recorded event facts, time stamps and raw channel references immediately after a run. Legal review could annotate but not replace the original entry.

The policy seemed minor compared with indictments, yet Luc considered it essential. Fear of liability could become another incentive to hide uncertainty, recreating the very problem the investigation sought to solve.

One evening Hale asked him why he cared so much about wording.

“Because words are interfaces,” he said.

“Between what?”

“Reality and the next decision.”

Hale wrote the phrase in her notebook.

“Annoyingly good,” she said.

“Thank you.”

“Still not praise.”