The Thursday methodology discussion had become, over the past three weeks, one of the more reliable features of Sicheng's working week.

This was information he processed with the same neutral precision he applied to other forms of data that had begun to behave consistently: noted, filed, not examined too closely for what the consistency implied. The discussions were productive. They generated output that was measurably improving the collaboration paper. These were objective facts about the meetings, and he held them at the front of his thinking and the other things — the way the hour passed differently than any other hour in his week, the specific quality of the silence between exchanges, the fact that he had started arriving to the conference room six minutes early to arrange his notes in a way that was probably not necessary — at a careful distance behind.

He had his notes open and his laptop positioned and his reasoning about the week's central structural issue partially assembled in his head when she walked in at exactly two o'clock, notebook in hand, and sat down across from him with the particular settled attention of someone who had also been thinking about the week's central structural issue.

"The framework assumption," she said, by way of greeting.

"Yes," he said.

They had been working together long enough that this constituted sufficient setup.

Three weeks. He had been in Edinburgh for four years and the Thursday meeting had become, in three weeks, one of the things he looked forward to without deciding to look forward to it, which was a category of experience he had been careful to avoid for most of his adult life and which was proving, as he had always suspected it would, entirely resistant to carefulness.

---

The first forty minutes went well. They moved through the agenda in the efficient parallel rhythm they had developed: one person talking while the other took notes, then switching, then a period of working in adjacent silence that had, at some point in the past three weeks, stopped requiring any effort to maintain. He had identified the points of structural tension in his framework. She had identified the corresponding issues in hers. The overlap was productive.

Then she said: "Your assumption about network directionality."

He looked up.

"The model assumes directionality is static across the observation window," she said. She was not looking at him. She was looking at her notes, which meant she had thought about this before today and was choosing her words with the care she used when she had a point she considered important. "But in the biological systems I'm modeling, directionality is dynamic. It changes in response to perturbation events. If your framework can't account for that, the integration we've been building doesn't hold."

He was quiet for a moment.

"The static assumption is a necessary simplification," he said. "Without it, the computational cost—"

"I know the computational cost argument." She looked up now. "I'm saying the simplification produces results that diverge from biological reality at the exact time points we're most interested in. The perturbation events."

"The divergence is within acceptable parameters—"

"Acceptable to whom?" She set down her pen. "The Henderson framework was built for stability analysis in closed systems. We've been extending it to open systems, which was my point in the first seminar. Directionality is another version of the same problem."

He had, in fact, been thinking about this. He had been thinking about it since Tuesday, when he had read a paper that had made a related point in a different context and had found himself returning to it repeatedly in the hours afterward.

"You're describing a version of the temporal variance problem," he said.

"I'm describing a more fundamental version of it." She picked up her pen again. "The temporal variance extension addresses timing. Directionality addresses the underlying assumption that the system structure is consistent across time. They're related but distinct."

He looked at his notes. He had approximately four responses available to him, and three of them were forms of defense and one was an acknowledgment that she had identified something real. He went, after a moment, with the fourth.

This was, he recognized, becoming a pattern. In the first three Thursday meetings he had defaulted to the defensive responses more often than not — not because they were better, but because they were familiar, because they kept the conversation on terrain where he felt he could control the outcome. She had responded to this by finding the gap between the defense and the actual question and asking about the gap directly, which was both methodologically correct and, he had to admit, deeply irritating in the way that accurate criticism was always more irritating than inaccurate criticism.

"Walk me through the divergence pattern," he said.

She did. It was precise and well-evidenced and built to a conclusion he could not easily dismiss, because she had anticipated the three most obvious objections and incorporated them into the argument rather than leaving them as vulnerabilities. He followed it, asked two clarifying questions, received two clear answers, and arrived at the same place she had started: the static assumption was a problem.

"Fine," he said.

She looked at him.

"You're right about the directionality. The static assumption holds within the Henderson framework's original parameters and breaks outside them. We need a dynamic version."

"That's not a small revision."

"No." He pulled his notebook toward him and began writing. "But it's solvable. The dynamic directionality model would require — " he calculated " — probably two weeks of additional work on my end. Possibly three."

"And on mine, recalibrating the integration points." She was writing too now. "Call it two weeks if we work on it in parallel."

They worked in silence for a few minutes, both of them mapping the problem from their respective ends. The conference room had its usual early-afternoon quality: the low hum of the building's ventilation, the faint sound of Edinburgh beyond the window, the smell of coffee from the pot someone had left running in the hallway. He was aware of these things the way he was aware of most background variables — peripherally, as context rather than content.

He became aware, at some point, that they had stopped writing in English.

He was not sure when it had happened. He had been working through the directionality problem verbally as he wrote, thinking out loud in the way he sometimes did when a problem required both the speaking and the noting simultaneously, and she had responded, and at some point the responses had been in Mandarin — not because either of them had decided to switch, but because the argument had moved faster than the decision about which language to use, and Mandarin moved faster for both of them on terrain this specific.

He looked up. She was reaching for her water bottle, and she glanced at him as she did, and he could tell from the slight pause in the gesture that she had noticed it too.

"我们在用中文说话," he said. We're speaking Chinese.

She set down her water bottle. Looked at him with the specific attention she used when something had landed that required processing.

"你不介意吗？" she said. You don't mind?

He considered this for a moment. The question was simple. The answer was also simple, and was also, he was aware, more than the question.

"不介意," he said. I don't mind.

She held his gaze for a beat — the particular quality of looking that meant she was checking whether the answer meant what it appeared to mean. Then she nodded, very slightly, and picked up her pen.

They continued in Mandarin.

The shift was subtle and total at the same time. He had noticed this about language before — the way a conversation in a different language was not the same conversation translated, but a different kind of conversation, shaped differently by the different weights of words and the different associations they carried. His Mandarin was not the Mandarin of the conference rooms where he had learned to be professional. It was older than that, less managed, more himself in the specific way that the first language you thought in was always more yourself than the ones you learned later.

---

The dynamic directionality model took shape over the next twenty minutes in a way that neither of them had entirely predicted: not one person's solution extended to accommodate the other's constraints, but a third approach that neither of them would have arrived at independently. He recognized this kind of outcome. It was what happened when two frameworks that had been built in isolation were genuinely allowed to talk to each other rather than merely coexist, and it was rarer than most people in collaborative research admitted, because it required both parties to hold their own approach loosely enough to let it change.

She had, he had noticed over the past three weeks, a particular gift for this. She could argue for a position with complete conviction and abandon it with equal completeness the moment a better position was available. He found this difficult to emulate and increasingly admired it.

"This works," she said, looking at the whiteboard they had filled in the process of working through the new approach. She said it in Mandarin, still, without thinking.

"Yes," he said, also without thinking.

A moment. The conference room had the particular after-argument quality of a space where something had been resolved — lower tension, slightly warmer, the specific ease of people who have been through something together and come out the other side with the work in better shape than it started.

He looked at the whiteboard. Then, without quite deciding to, he asked: "你在大陆哪里长大的？"

The question came out in the same register as everything else they had been saying for the past hour — conversational, unguarded, as if it were simply the next thing. He was aware, a fraction of a second after saying it, that it was not the next thing. It was a departure. It was the first question he had asked her about herself, as herself, that had nothing to do with the work.

She looked at him.

He did not walk it back. Walking it back would have been more revealing than leaving it.

"四川," she said. Sichuan. And then, after a pause: "长大是在成都。后来去深圳读大学。" Grew up in Chengdu. Then went to Shenzhen for university.

"I know Shenzhen a little," he said, in Mandarin. "I was there for a conference three years ago."

"It's changed a lot in three years," she said. "It changes every three months, really."

"Taipei is like that in some neighborhoods." He paused. "Other parts feel like they haven't changed in fifty years."

"Chengdu is like that." Something in her voice shifted — not much, but enough that he noticed. The careful composure she maintained in professional contexts had relaxed slightly, in the way that things relaxed when a language changed and the associations changed with it. "My grandmother's neighborhood. The same tea house on the same corner."

"Does she still live there?"

"Yes. She refuses to move." A small pause. "My mother is the same. Neither of them has been further from Chengdu than Chongqing."

He thought about this. He thought about what it meant to be the first person in a family to go this far — not just from Chengdu to Shenzhen, but from China to Edinburgh, carrying a scholarship and a research question and the accumulated expectations of everyone who had believed she would get here.

"That's a long way to travel," he said. He did not specify from where to where. He thought she understood he meant more than geography.

She looked at him for a moment. "Yes," she said. "It is."

They sat with this for a moment that lasted slightly longer than most silences in the conference room. Outside the window, Edinburgh continued to be Edinburgh: gray sky, the suggestion of rain that had been suggesting itself since morning and had not yet committed.

"Taipei?" she asked.

He told her. The neighborhood, the specific quality of autumn there that was nothing like Edinburgh's autumn and which he had stopped missing sometime in year two and had started missing again in year four when he had gotten close enough to the end to start thinking about after. His father's study, his sister's complicated feelings about him leaving, the family obligation that had sat differently on his shoulders after four years of being far enough away to see its shape.

He had not talked about any of this to anyone in Edinburgh. He had not talked about most of it to anyone at all. The people who knew him well enough to ask were in Taipei, and he spoke to them in the register of someone who had done fine and was doing fine and would continue to do fine, which was accurate in most respects and incomplete in the ways that mattered. The people in Edinburgh knew the version of him that had arrived here already assembled — the research, the framework, the productive professional who stayed in the lab until eleven and read three papers a day and never needed anything from the city he was living in except its libraries.

She was watching him tell her about his father's study and his sister's complicated feelings and she was watching him with the specific quality of someone who was adding new information to a file they had already been building for a while, and he was aware of this and kept talking anyway.

She listened with the quality of attention that did not perform itself — not leaning forward, not maintaining steady eye contact as a signal, just actually attending. He became aware of the clock on the wall when it showed four-seventeen, which meant the meeting had run fifty-seven minutes over its scheduled end time, and they had used none of those minutes on protein network dynamics.

"I should — " he started.

"Yes," she said, already gathering her notebook.

They packed up in the specific efficient silence of two people who had just had a conversation they had not planned to have and were now returning to the version of themselves that had arranged a meeting for a specific professional purpose. He held the door. She went through. In the hallway, they paused at the point where his direction and hers diverged.

"Same time next week?" she said. In English. The deliberate switch was its own kind of signal.

He looked at her. She was watching him with an expression that was not quite neutral — that had, in fact, nothing neutral about it, when he looked at it directly, which was something he had been avoiding doing in corridor lighting.

"Same time," he said.

They went their separate directions.

---

He was at his desk for twenty minutes, not reading what was on his screen, before the notification appeared. Shared document, new entry, timestamp 4:03pm:

*Sichuan has better hot pot than Taipei. This is not up for debate.*

He read it.

He had not expected this. He should have, perhaps, because this was how she communicated when she had decided to communicate something — in the shared document, tucked between professional entries, in a register that was deniable if she needed it to be and unmistakable if she didn't.

He typed: *I've never been to Sichuan.*

He looked at what he had typed. He added: *That's not a concession.*

He sent it.

Her reply came in ninety seconds: *It's not. But it could be an invitation.*

He read this.

He read it again.

Outside the Henderson lab window, Edinburgh was doing what it did in November: rain, the early dark, the orange of the streetlamps coming on in the wet. He had been in this city for four years and two months. He had spent most of that time building something specific here — a thesis, a framework, a way of being in a room and not needing anything from it beyond the work. He had been, by most available metrics, successful at this.

He looked at the screen. At *it could be an invitation*, sitting in the shared document between a methodology note and his own response, in her register, which was the register of someone who said what they meant and left you to decide what to do with it.

He did not reply.

He did not close the document.

The rain continued. He sat in the Henderson lab with the shared document open on his screen and thought about Sichuan and Taipei and the specific quality of the last two hours and the question of what, exactly, he was still waiting for.

He was waiting, he thought, for the version of events in which this was a complication he could manage rather than a thing he wanted. He had been waiting for four years in various forms. He had become very good at waiting. The difficulty was that *it could be an invitation* was sitting on his screen in her register, which was the register of someone who did not hedge and did not leave exits, and waiting was beginning to feel less like prudence and more like a choice he was making every day, consciously, against a countervailing pressure that had been building for approximately eleven weeks.

He did not reply. He did not close the document. Outside, Edinburgh kept raining. The document kept being open. He sat with both of these facts and thought about what it would mean to be the kind of person who replied. He thought about it for a long time. The rain was patient. So was the document. So, apparently, was she.

At eleven o'clock he closed the document without replying. This was also a kind of answer, he thought. He was not sure yet which question it answered.