She arrived with her notebook.

This was deliberate. She had spent two hours the previous evening preparing for the first training session, filling twelve pages with the most relevant material she could identify from the Reedfall archive: foundation principles, the lineage of techniques she believed would be most applicable to the gaps Ilyun had identified, cross-referenced notes from Minhwan's demonstration and her own analysis of what her body had failed to execute. She had organized the pages in what she considered a logical teaching order and secured them with a cord.

It was not a small notebook.

Ilyun looked at it when she set it on the railing of the practice area. He picked up the top three pages and read them with the attention he gave everything, which was complete and unhurried. Then he set them back down on the railing. He did not look at the remaining nine pages.

"When did you prepare these?"

"Last evening."

"How long did it take?"

"Two hours."

He was quiet for a moment. He looked at the practice area rather than at her. Then: "Put the notebook away."

She held the notebook for a moment longer than necessary before setting it in her bag. The feeling this produced in her was not simple frustration, which she had experienced many times and could manage. It was the more complicated feeling of having prepared something carefully and had it set aside without assessment. She could not argue with a thing that had not been evaluated.

"All right," she said.

"Show me something you do well. Anything."

She had anticipated this, at least. She chose the most technically complex sequence she could execute with reasonable accuracy: the third form of the Crane series, the one she had been working on from observation and which she had drilled with the students at Reedfall for four months before setting out. She moved through it slowly and then at pace. She knew it was not clean. She knew where the gaps were. She did it anyway and let him see all of it.

When she finished, he said nothing for a moment. She waited.

"Stance," he said.

She stood very still.

"Everything else is downstream from stance. The form you just did has three distinct points where the technique fails because the foundation is incorrect. Everything built on the wrong foundation carries the error forward." He paused. "We start with stance."

She had read two chapters of the Foundation Manual on stance. She had read the original methodology text's treatment of stance three times. She knew, specifically, what the correct Reedfall stance was supposed to produce in terms of weight distribution, joint alignment, and available range of motion. She was aware that her current stance deviated from the ideal in certain respects. She had assumed these deviations were the kind of thing that could be addressed after the more significant execution gaps were closed.

"The manual covers stance in the first chapter," she said. "I've read it twelve times."

"I know."

"The second chapter covers the transition from stance to the first movement sequence. The third chapter covers, "

"I know what the manual covers," he said. It was not unkind. It was the same flat quality as everything else he said. "Come here."

She came.

He looked at her from the side, then from the front, with the specific quality of assessment she had recognized from the first day at the fence: reading the structure rather than the surface. He reached out and placed two fingers on her left shoulder and applied light pressure.

She lost her balance at the contact. She caught herself, but it was close.

"That is the consequence of the stance you currently have," he said. "I used less force than most opponents will use in the first exchange of any match."

She stood with this. She did not have a rebuttal because he had not made an argument. He had demonstrated a fact.

"All right," she said again.

---

The first session was two hours of standing.

Not in one position, not statically, but in continuous small adjustments as he directed, working through the specific mechanical corrections that the Reedfall manual described correctly in principle and incorrectly in executable detail. She had read the manual's description of the correct stance many times. She was beginning to understand that reading the description and inhabiting the described position were two processes that did not necessarily produce the same result.

He did not explain anything she did not ask about. She asked specific questions and he gave specific answers. The economy was similar to her father's, the quality of a person who does not spend words on things that can be demonstrated.

At one point she quoted directly from the Foundation Manual to support a question about hip alignment.

He listened to the quote and then placed his hand briefly on her hip and adjusted the angle by a few degrees.

She felt the entire weight of the stance shift and resettle into something more stable.

"The manual is correct," he said. "The description is not precise enough to tell you where the angle is in your specific body. That is what the manual cannot do." He removed his hand. "That is what I can do."

She stood in the corrected position and thought about her grandfather's transmission chapter. The gap between the written description and the inhabited position, which the methodology text had spent eighteen pages describing and which she had read and understood and which she now understood again, differently, from inside.

By the end of the two hours, her hip flexors and posterior chain ached in ways she had not felt from any previous training. She was producing the correct stance. She was producing it by holding a specific physical tension that the correct stance required, and the muscles responsible for that tension had not been trained for it.

Ilyun looked at her one final time before leaving.

"Tomorrow," he said. He walked away.

She stood in the practice area alone for another minute, holding the stance because stopping felt like a retreat. Then she sat on the railing and opened her notebook and found the twelve pages she had prepared.

She read through them slowly.

They were accurate. All of them were accurate. The information was correct, the cross-references were correct, the organization was logical.

She turned to a new page and wrote one sentence: he looks at the place where the technique fails, not the place where it is correct.

She was not entirely certain what she meant by this yet. She was certain it was true.

---

The ache in her muscles that evening was specific enough that she could map it: the left hip flexor was doing more compensatory work than the right, which meant her previous stance had been unequally loaded in a way she had not noticed. The notation she had developed for documenting technique had no mark for this kind of information, the kind that lived in the body rather than in observation.

She sat at the desk in her guesthouse room and looked at the twelve prepared pages and thought about the problem of documentation.

Her grandfather had written: the practitioner who believes she can learn solely from watching has confused the map for the territory. She had read this and understood it. She had not understood, until today, what it felt like to stand inside the gap between the map and the territory with someone who knew exactly how large the gap was.

She was going to have to learn to read the map differently. She was going to have to learn to read it as a set of directions that required her body to complete the translation, not a set of instructions that the body should execute directly.

This was a different problem than the one she had come to Seven Roads to solve. Or it was the same problem described more precisely.

She added a note to the twelve pages: start over from the foundation. The foundation is not what I thought it was.

Then she went to sleep with her hip flexors still aching and the sensation of the stance still in her body, the way significant things sometimes stayed.

The ache in her posterior chain the next morning was specific and diagnostic. She lay on her sleeping mat and catalogued it: the left hip flexor doing more compensatory work than the right, which meant her old stance had been unequally loaded in a way she had never measured. The left side of her lower back, where the muscles connecting the pelvis to the spine had been asked to do a new kind of stabilizing. Her right heel, where she had been holding tension she had not known she was holding.

She sat up.

She had read the Foundation Manual's stance section twelve times. She had retained every word of it, every diagram, every notation in the margins. Not one of those twelve readings had told her which muscles to recruit to hold the correct position, because the manual had been written for practitioners who already had those muscles trained, who were refining an existing foundation rather than building one. The manual was correct. The manual was not for her.

She thought about this while she dressed and went downstairs.

She thought about the twelve pages she had prepared and set aside on the railing, and what they represented: her attempt to approach a physical problem with a text-based solution. The pages were accurate. They were also beside the point, because the point was not information. The point was the body that had to act on the information, and she had spent twenty-two years developing one of those things and neglecting the other.

She was not sorry she had prepared the pages. She needed them eventually. She was sorry she had not understood earlier that eventually was the right word.

She went to the practice area earlier than necessary and stood in the correct stance for thirty minutes before Ilyun arrived. She was not doing this to demonstrate preparation. She was doing it because the stance needed to become a physical fact rather than an intellectual understanding, and the only way to move between those states was repetition, and she had time for repetition, and she was going to use it.

When he arrived, he checked her stance without comment and moved her right foot a quarter of an inch. She felt the correction and adjusted to maintain it and stood in it while he prepared the first drill.

She thought: yesterday he pushed me over with two fingers. Tomorrow I will be harder to push. Not because I have become stronger, but because I will be standing correctly.

She thought: this is the work. Not the notebook. Not the archive. This.

---

The first drill was a simple sequence, the kind designed to build the body's record rather than the mind's: a repeated pattern of stepping and turning and arriving in a specific stance configuration, run twenty times, then thirty, then with variation. She ran it until the shape of it began to arrive ahead of her decision to produce it. That was the threshold he was training her toward and she knew it, and she drove toward it with the same attention she gave everything she had decided mattered.

He was watching her run the drill from a position at the edge of the practice area, the position she had come to understand as his assessment position, the angle and distance from which he could read the full structure of her movement rather than its surface.

On the twelfth repetition he said: "Breathe."

She had been holding her breath through the turn. She had not noticed. She released it and ran the next repetition with the breath present and felt the difference immediately: the turn was more complete, the arrival in stance more stable, the connection from the ground cleaner.

This was something she had not read in any text. The texts she had read addressed breath in combat in terms of timing and endurance. None of them had described what she had just experienced, the structural contribution of breath to the stability of a movement pattern.

She filed this. She would have to find the language for it later. For now she ran the drill.

By the twentieth repetition her hip flexors were telling her they were being asked for something new and sustained. She continued. By the thirtieth repetition the fatigue had reached a plateau and she was still producing the correct form because the correct form had become cheaper to produce than she had expected: the muscles were doing less work when the mechanics were right.

She had read about this. Efficient technique conserves energy because the load is distributed correctly. She had understood it.

Now she knew it.

The following morning she arrived before him again. She had been doing this since the third session, arriving early enough to run through the previous day's corrections before he came, not to impress him but because the corrections were better when her muscles were already warm and the body had been reminded of its previous work.

She ran the stance sequence, then the transition into the first Foundation movement, then the hip alignment check. All of it was present. Not perfect, the stance was never quite the same twice, but the core of it was there, the connection from the ground that had gone from being a concept to being a physical fact.

He arrived and watched her complete the sequence without speaking.

They began the session.

She had been thinking about something since the previous evening and she asked about it midway through the second drill, in the pause between repetitions: "When you assess a practitioner, what do you look for first?"

He was quiet for a moment. "The part that is most wrong," he said.

She ran this. "Not the part that is most correct."

"The correct parts will remain. The wrong parts will compound. The wrong parts are where the work lives."

She thought about this while she ran the next repetition. She thought about how this was the inverse of how she had been reading her own training: she had been assessing herself by what she was doing correctly and noting the gaps. He had been assessing her from the gaps and working backward to what the gaps revealed. Both approaches produced accurate information. They produced it in different orders.

"My father works the same way," she said. "He corrects the specific thing, not the general level."

"Yes."

"I thought that was his particular method."

"It is the method," he said. "The general level is an average. The specific thing is where the change can actually happen."

She ran the transition again and thought about this and thought about how it applied to more than drilling.

---

The afternoon session was footwork, a full two hours of stepping patterns designed to build the lower body's vocabulary of positions. She ran the patterns until she was too tired to run them correctly and he told her to stop and she was aware, as she stopped, that this was the first time he had told her to stop rather than pushing through the fatigue.

She stretched and cooled down and thought about the difference between training through fatigue and stopping at it, the information each approach generated, the different purposes they served.

"Why stop now?" she said. "We have stopped earlier than the previous sessions."

"You were no longer producing the correct pattern. Continuing would have been training the incorrect version." He put on his jacket. "Fatigue is information. What to do with it depends on what information you are after."

She sat with this on the walk back to the inn and thought about the way the day's two sessions had been structured and what the structure revealed about how he understood the process of building capacity. She was starting to be able to read his training design the way she read a complex text: not just the surface, the architecture under it, the choices being made about what to develop and when and why.

Reading his training design required the same quality of attention as reading him, which was the attention that moved toward understanding rather than recording.

She had been paying that kind of attention to both things for some time now.

She thought about the inverse, too: if he assessed from the worst part rather than the best, what did that mean about how he understood progress? It meant progress was not the addition of correct things but the removal of incorrect ones. Every incorrect thing that was identified and corrected created space for the correct thing to exist without having to compete with the incorrect version.

She had been reading progress the other way for years. She had been building up correct knowledge beside incorrect knowledge, assuming the correct would eventually outweigh the incorrect by volume. He was removing the incorrect at its source, which was more efficient and more permanent.

She thought: this is the same logic as the three-inch adjustment. Not adding a new position to the repertoire. Finding the correct position and removing the incorrect one from the foundation so the correct one has nowhere to compete against.

She ran the drill again and thought about the things that were most wrong in her current practice and what it would mean to remove them rather than to supplement them.

She thought about the way the day had ended: the two-hour footwork session, the specific fatigue, the way he had told her to stop. She had been expecting to be pushed through the fatigue. She had been prepared to be pushed through the fatigue. His decision to stop at the point where continuation would have built the wrong pattern was the same kind of intelligence as the three-inch adjustment: not more but correct, the distinction between work that built something and work that reinforced something it should not reinforce. She was learning to tell the difference. She was starting to apply it to her own training decisions as well as to receive it from his. This too was part of the ground being built.