Hall Seven was on the third floor of the academic wing.

  Better ventilation than most buildings on this side of campus.

  Worse acoustics than any of them.

  The ceiling was vaulted. The stone old enough that sound moved across it in strange ways—a voice at the front could arrive at the back wall fractionally delayed, certain syllables arriving twice. Kael had noted this on his second day, when he'd come early enough to test it.

  He sat in the fourth row.

  Close enough to see the board clearly.

  Far enough back to not be the obvious center of attention.

  His notebook was open. Pen in hand. He had read ahead to the section Aldea was likely to cover.

  He had been right about the section.

  He had been right about the gaps in it too.

  ?

  Aldea wrote her premises on the board before she spoke them aloud.

  Methodical. Precise. The particular precision of someone who had thought about the same material for long enough that she knew exactly where the edges were—and exactly how to present them to students who hadn't found them yet.

  Forty-something. Dark-haired. Ink on her left hand from a habit of holding her pen between fingers while she thought.

  She did not perform warmth.

  She did not perform authority.

  She moved through her lectures the way she moved through the room—with a purpose you could observe but hadn't been invited to comment on.

  Kael had read her two published papers before the first class.

  The first: a careful examination of Affinity calibration in high-saturation field environments. Thorough, well-cited, the kind of work that established a researcher's credibility without committing her to any particular theoretical position. The paper made no enemies and opened a productive line of inquiry that other researchers could follow. Professionally calculated.

  The second was older—a theoretical piece on the upper limits of Stability measurement, ending with questions she had not answered and had not returned to in print. The questions were specific and pointed: what happened at the theoretical ceiling, whether the dissipation assumption was derived or assumed, whether the calibration instruments were accurate at their own upper bounds. Questions that aimed directly at the framework's foundations.

  She had published them once.

  Then stopped.

  He had found the questions interesting.

  He suspected she still thought about them.

  Whether she had stopped because she'd found the answers, or because she'd understood that continuing was professionally inadvisable, was the variable he most needed to read correctly.

  ?

  Today's topic was conversion efficiency.

  Output equals Capacity times Stability, divided by one hundred.

  She wrote it on the board without comment. Worked through the implications. Why Stability varied under load. How the ranking system translated the formula into comparative scores. Why high Capacity and moderate Stability could outperform the reverse in certain contexts, and why the reverse was also sometimes true.

  She was thorough.

  The material was correct as far as it went.

  Kael watched the board and said nothing.

  He waited.

  She reached the upper-bound section—the theoretical maximum output, the ceiling the ranking system was built to measure against. She wrote the formula. Derived the upper limit.

  She paused.

  Not the pause of someone who had lost her place.

  The pause of someone who had arrived at a conclusion they had stated many times and who was, in this particular instant, briefly aware of its foundations.

  One second.

  Two.

  Kael raised his hand.

  ?

  "The formula assumes the remainder dissipates," he said.

  Aldea turned from the board. "That's correct."

  "What if it doesn't?"

  The room went quiet.

  The specific quiet of a moment that was either very obvious or very unusual, and no one was sure yet which.

  Aldea looked at him. Her expression was neutral in a way that required some maintenance. "Practitioners at maximum Stability still show a transmission ceiling. The energy that doesn't convert—"

  "The formula assumes it leaves the system." Level—not flat, just precise. "The dissipation term isn't derived. It's a premise. The framework treats conversion inefficiency as a physical constant, but the calibration records treat it as observed rather than proven."

  Aldea did not answer immediately.

  He had been watching her for two sessions. She thought before she spoke the same way he did—not pausing to construct a response, but actually running the problem through before committing. It was one of the reasons he had decided to take the risk.

  "The ranking system requires a fixed ceiling," she said. "Without the dissipation assumption, the upper bound becomes—"

  "Unbounded under certain conditions," Kael said. "Yes."

  Half a second of silence.

  A student two rows back said, audibly: "That doesn't make sense."

  Kael didn't look back.

  "The theoretical case you're describing," Aldea said, "would require a practitioner whose Stability doesn't decrease under load. The highest confirmed Stability value in the Academy's records is ninety-one point three percent."

  "The highest confirmed value is whatever the instrument recorded as its highest reliable reading. The calibration documentation specifies an accuracy range. At one hundred percent, the instrument returns an error code." He paused. "There are at least two documented cases in the Academy's own examination records."

  Ensure your favorite authors get the support they deserve. Read this novel on the original website.

  Aldea's pen stopped moving.

  She looked at him. Long enough to be noticeable.

  Not hostility.

  Something more like the expression of someone adjusting a calculation.

  "Your interpretation is not inconsistent with the raw data." Each word chosen for its exact weight. "It is, however, inconsistent with the interpretive framework the Academy uses to evaluate practitioner performance. The examination board's standards require—"

  "I understand that," Kael said.

  The room had gone quiet again.

  Different quality this time.

  This one had an edge.

  Students in the front rows had gone very still—the specific stillness of people who had registered something and were deciding how much distance to put between themselves and it. The students in the back rows were looking at Kael in the particular way people looked at something that had just become interesting in a way they hadn't expected.

  He noted the split.

  Front rows: alignment with the institutional position. They had come from families and backgrounds where the ranking system was not an abstraction—it was the mechanism that had shaped what was available to their parents and would shape what was available to them. Questioning it felt like questioning their own standing.

  Back rows: more genuinely curious. Not because they were better people. Because they had less invested in the answer.

  The dynamic was useful information.

  Aldea turned back to the board. Picked up her chalk.

  "The dissipation term is foundational to the ranking system. The Academy's evaluation protocol treats it as established." She wrote the next line of the derivation without looking at him. "The formula gives correct results within its design parameters. Questions about its theoretical limits are—" she paused— "a separate line of inquiry."

  The chalk moved.

  The class moved with it.

  Kael wrote four lines in his notebook and did not raise his hand again.

  ?

  After class, someone bumped his shoulder in the corridor without looking back.

  He let it pass.

  The exchange had accomplished exactly what he'd intended, and one thing he hadn't planned for.

  What he'd intended: establish that the question existed. That someone who knew the material had noticed the gap. That the gap was real enough to make an expert pause before answering.

  What he hadn't planned for: the speed.

  How quickly the room had moved from uncertain to opposed—before he'd finished the second sentence, simply from the posture of questioning the foundation.

  The framework wasn't just taught here.

  It was load-bearing.

  The ranking system. The dormitory assignments. The merit point allocations. The access hierarchies. Every mechanism he'd spent two days cataloguing ran on the formula Aldea had written on the board.

  Questioning the dissipation assumption wasn't an academic exercise.

  It was a structural challenge to the system that had put him in Block Four.

  They hadn't consciously thought that through.

  But they'd felt it.

  He filed the observation and kept walking.

  ?

  He saw Aldea near the stairwell at the end of the corridor.

  Same direction. Same pace. She didn't slow when she noticed him. Didn't speed up either.

  He matched her for three steps.

  "The intermediate theory texts. The librarian requires instructor approval."

  Aldea looked at him sideways. "I know what they require."

  "I found two cases in the examination records where the instrument returned an error at maximum reading. There may be more in the calibration archives."

  She walked four more steps before speaking. "The archives are restricted."

  "I know. Restricted for students without instructor access." He paused. "I'm not asking for the archives. I'm asking for the intermediate texts."

  Aldea stopped at the top of the stairs.

  Different look than the classroom.

  Not reading him against the class's expectations.

  The look of someone who had thought about a problem before, and was now seeing it arrive in a form they hadn't anticipated. Not surprise exactly—more like the recalibration of someone who had assumed the next instance of this thought would arrive differently, and was now adjusting.

  He had read her second paper carefully. The questions at the end of it had been specific—not broad theoretical gestures but precise, targeted questions about the calibration process, about what happened at the instrument's upper range, about whether the dissipation term was observed or assumed. She had written them as open questions, invitations for future research, the kind of thing you published when you wanted someone else to pick up the thread.

  No one had.

  Or if they had, it hadn't made it into print.

  He was now standing in front of her holding the thread.

  "You believe the formula is wrong," she said.

  "I believe the formula is a useful approximation that has become confused with the thing it approximates." He said it the same way he said everything—not as a performance, just because it was accurate. "The dissipation assumption is load-bearing for the ranking system. That doesn't make it correct. It makes it consequential."

  Aldea was quiet.

  "The examination board requires the formula to be taught as established theory." Neutral—not opposed, not endorsing. Stating conditions. "A student who produces work contradicting the established framework will receive a notation on their evaluation record."

  "I understand."

  "The notation affects merit point allocation."

  "I know."

  She looked at him for one more moment.

  Then she turned toward the stairs. "The intermediate section has a chapter on measurement methodology. The third text covers historical calibration standards." She paused at the top step. "I'll leave a note with the library desk."

  She walked down.

  Kael stood at the top of the stairwell for exactly two seconds.

  Then he went back to the library.

  ?

  The notation was on his evaluation record by the following morning.

  He found it at the merit point terminal in the dormitory block's ground-floor office. Small entry. Neat official script. Processed overnight—automated, he suspected, triggered by Aldea's submission of the class record. She hadn't added the notation herself. She had submitted the record through the standard channel, and the channel had flagged it.

  *Non-standard interpretive framework. Assessment: pending review.*

  It didn't affect his current total.

  But it was in the system.

  He stood at the terminal for thirty seconds. Read it twice. Confirmed it said exactly what it appeared to say.

  The notation was a cost. He had known there would be a cost—questioning the framework publicly was not free, and it was not supposed to be free, because systems that allowed their foundations to be questioned without consequence were systems that understood their foundations might be wrong. This system was not that kind of system.

  What he'd needed to know was how large the cost would be, and how fast it would arrive.

  Now he knew.

  Overnight. One session. A single exchange in a lecture hall.

  The window was shorter than he'd estimated.

  He went to breakfast, ate in eleven minutes, and walked to the library.

  The intermediate texts were waiting at the library desk.

  He picked them up without comment. Sat at the reading table furthest from the door. Opened the first one.

  The chapter on measurement methodology was exactly where Aldea had said.

  He read until the bell rang. Not finished. But he had found what he was looking for.

  Two footnotes, seventeen pages apart, citing the same calibration study. Conducted forty years ago. Superseded by newer standards that had adopted the dissipation assumption without examining the original study's reservations.

  The reservations were still there.

  In a footnote. In a superseded document. In a restricted section of the library.

  He closed the text, marked the pages, and wrote three lines.

  Then he went to his next class.

  The framework was wrong. The system knew it had gaps and had filed them in places where students without instructor access couldn't find them.

  That was useful information.

  ?

  What it told him was this: the gaps were known.

  Not by everyone. Not openly. But by someone, at some point in the past forty years, who had read the original calibration study's reservations before the newer standards had superseded it, and who had decided—or been instructed—to move those reservations into a restricted section rather than integrate them into the framework.

  The question was why.

  The most charitable explanation: the reservations were theoretical edge cases with no practical implications for the ranking system's daily function. Files correctly. Accessible to researchers. Not worth disrupting an entire institutional evaluation structure over a footnote about calibration accuracy at extreme readings.

  The less charitable explanation: the reservations had practical implications that the people who built the ranking system had decided were better left unexamined.

  He had, this morning, produced a set of training hall results that were not explainable under the more charitable reading.

  The platform inscriptions had gone white.

  He was in Block Four because his Capacity was 38—the lowest in his cohort. The formula said: 38 units maximum, nothing more, no variation possible.

  If the dissipation assumption was wrong, the formula was incomplete.

  If the formula was incomplete, the number on his dormitory assignment slip was an artifact of a measurement system that couldn't see what he actually was.

  Not a fact.

  A gap in the instruments.

  He opened his notebook, wrote two more lines, and went to the training hall.

  He had work to do.

  The footnotes had pointed him toward an older study that the current framework had superseded without resolving. The reservations in that older study were still there, filed in a restricted section, accessible to researchers with instructor authorization.

  Aldea had given him the intermediate texts.

  She had not given him the restricted section.

  He had not asked for it.

  He would ask for it when he had something to offer in return—not just a question about the framework, but a demonstration. Something that couldn't be dismissed as theoretical. Something that had already been witnessed by three students and a hall manager with eleven years of baseline data.

  He had that.

  He needed to present it correctly.

  He walked into the training hall, stepped onto platform three, and began.