The Repair Lab was, as Sera had assessed from her first visit, a hazard environment.

This was not a criticism, exactly. She had spent enough time in functional creative spaces to understand the difference between disorder that was accumulating without purpose and disorder that was the byproduct of active making. The lab fell definitively into the second category: everything in it had a logic, even if the logic was not immediately apparent to a visitor, and the disorder was the kind that could be traced backward to specific decisions if you spent enough time looking.

She spent enough time looking.

She was here on a Saturday morning because Riku had sent her a message the night before: *the primary recorder is fully functional. Come get it whenever — I’ll be in the lab.*

Fully functional. The words she had not let herself expect.

She arrived at nine-fifteen and found the lab as it always was: Riku at the bench, something in pieces in front of him, three fans running for reasons she had given up asking about, the ambient smell of solder and machine oil and the remnants of whatever he had eaten at the bench for dinner, which appeared to have been convenience store pasta.

On the corner of the bench nearest the door, placed with conspicuous care relative to the surrounding disorder, was her recorder.

She picked it up.

The casing was clean — someone had removed the pond sediment entirely, buffed out a small scuff on the side she had not noticed before, and the screen, when she pressed the power button, came on immediately. The familiar startup sequence. The waveform display, the file directory, the settings she had configured two years ago and never changed.

She navigated to the file directory and checked: all the intact files present, the reconstructed grandmother file in the front pocket folder where she had put the card, the project files organized as she had left them.

She pressed play on the sound_050 file — the petal path recording from Tuesday — and listened to ten seconds through the built-in speaker.

The frequency response was correct. The mid-range clarity, the bass response, the spatial imaging — it was her recorder. Not a copy, not a replacement. The original, returned.

“The condenser element was partially water-damaged,” Riku said from the bench, not turning from his work. “I had to source a replacement from a discontinued model, which took some time. The new element is technically superior to the original — wider frequency range, lower noise floor. It should perform better in the low-frequency sessions.” He paused. “I kept the casing because I thought you’d want the original casing.”

She looked at the scratch on the upper left quadrant of the screen.

Still there.

Of course it was.

“Yes,” she said. “Thank you.”

He turned from the bench then, wiping his hands on a cloth, and looked at her with the direct attention he used for things that mattered. “The scratch is the same. I didn’t — I wasn’t sure if you wanted it repaired or not, and I decided it was better to leave something I wasn’t sure about than to change it.”

“You decided correctly,” she said.

He nodded.

She stood in the disorder of the Repair Lab holding the recorder and felt something settle that had been slightly wrong since the first Tuesday of the semester, since the pond, since the specific sound of the recorder hitting water and the particular quality of that absence.

The recorder was back. The file was recovered, as much as it was going to be recovered. The project had fifty-two sounds and forty-eight remaining and five and a half weeks of spring.

She set the recorder down and looked around the lab.

This was the first time she had looked at it without an immediate task drawing her attention. The first three visits had been for specific purposes — the diagnostic, the data recovery, the equipment briefing — and she had been focused on those purposes. Now she looked.

The lab was structured around three large workbenches running parallel, each at a different stage of occupation. The nearest was the active bench — current project, active tools, the kind of productive disorder that was uncomfortable to look at but clearly in use. The middle bench was storage and reference: materials organized by type, equipment cases, the sub-bass transducer in its carrying bag, a row of completed small-scale models that she recognized as iterations of the haptic chamber design, each one slightly different from the last.

She walked over to the row of models.

There were seven of them. She looked at them left to right. The first was clearly early — the proportions were exploratory, the materials inconsistent, the kind of model that was thinking through a problem rather than representing a solution. Each subsequent model refined something: the proportions, the material specification, the surface treatment. The seventh was the one she had seen in progress over the past weeks, and it was different in kind from the first six — not just refined but arrived.

“How long have you been working on this?” she asked.

“Two years,” he said. He had come to stand beside her, not too close, the comfortable distance they had developed over weeks of standing next to each other during recording sessions.

“All of these are the same project?”

“All iterations of the same question. The question changed as I understood it better.” He looked at the row of models. “The first version was about translating sound for people who can’t hear it. Which is a… not wrong, but incomplete framing. It assumed hearing is the standard and everything else is accommodation.”

“And the current version?”

“The current version is about designing a space where sound exists as a physical experience regardless of how you receive it. Not translation. Just — presence.” He picked up the seventh model and turned it in his hands. “So that the person who feels the frequency in their chest and the person who hears it through their ears are having the same experience. Just through different channels.”

She looked at the model in his hands. It was, she thought, one of the better pieces of design work she had seen in three years at this university, and she had seen a considerable amount of it. The proportions were correct in a way that was hard to describe technically but was immediately perceptible: it sat in the hand the way things sat when their geometry had been fully understood.

“You shifted the framing after the pump room,” she said. Not a question.

He looked at her.

“The notes you took that day. And the gymnasium channel. And the way you’ve been positioning yourself during sessions lately — you’re not just holding the microphone. You’re attending to your own perception.” She looked at the model. “You’re doing primary research.”

A pause. Then: “Yes.”

“On yourself.”

“On how it feels to receive sound through a different channel than I usually use,” he said carefully. “Yes.”

She was quiet for a moment.

“It’s a good project,” she said.

He looked at her as if she had said something unexpected. Which, she supposed, she had — she did not often offer unprompted assessments of other people’s work, and when she did they tended to be qualified or conditional. This had been neither.

“Thank you,” he said.

“I mean it technically. The project has a coherent concept and a clear research basis and the material execution in this model is—” She stopped.

“Is?”

“Is correct,” she said. “In the way things are correct when the person making them understands what they’re actually making.”

He looked at the model in his hands for a moment. Then he set it back in its place in the row, with the same care she had noted him applying to her equipment.

“Your recorder,” he said. “Take it for a test this afternoon. Let me know if anything needs adjustment.”

“I will.” She picked it up from the bench and put it in her bag. The familiar weight. The familiar shape.

“Also,” he said, “the third bench has a specific acoustic property you might be interested in. The metal surface is at a thickness that resonates at around two-seventy hertz when struck, which isn’t a spring sound exactly, but it’s a sound that’s been in this room for four years of work and I thought—” He stopped. “Condition four. Sorry. I was just mentioning.”

She looked at the third bench. Then she took the recorder out of her bag and held it up.

“Strike it,” she said.

He picked up a small mallet from the tool rack and struck the metal surface of the third bench once, a clean single blow, and the note rang out in the lab — a clear metallic tone at approximately the frequency he had described, sustained for several seconds before fading into the ambient hum of the fans.

She pressed stop.

“Sound_053,” she said.

She wrote it in her notebook: *sound_053: repair lab bench, metal surface resonance, 270hz approx, single strike, 0932hrs.*

And below that, because she had been doing this more often lately, a second notation: *the recorder sounds like itself again.*

She tested the recorder that afternoon.

The north slope of the campus, just below the dormitory buildings, had a particular quality in the middle hours of the day: the sun hit the retaining wall at an angle that warmed the stone, and the warmed stone created a slight updraft that moved through the ivy on the wall in a way that was audible if you knew to listen for it. The ivy had been on the wall since at least the late eighties — it was in several of the archive photographs from that era — and its movement was a sound that had existed on this campus for decades.

She positioned the recorder and let it run for four minutes.

The frequency response was, as Riku had described, improved. The low-end clarity was better, the noise floor lower; she could hear the ivy movement cleanly without the slight electronic hiss that had been a feature of the old condenser element. The wide frequency range he had mentioned showed up in the upper registers, where she could hear detail she had previously been reconstructing from inference rather than direct capture.

She played back the four minutes and made notes.

Then she walked down to the ornamental pond and sat on the stone ledge — the same spot as the first morning — and held the recorder and thought about what it meant that it was back.

The recorder had been gone for three weeks. In those three weeks the project had added thirty-two sounds and gained access to spaces she had not been able to reach independently and developed a working rhythm with another person that had not disrupted her process but had altered it in ways she was still mapping. The spring had continued. The blossoms had given way to early leaves. The light had changed.

The recorder was back and it was better than before and the scratch on the screen was exactly where it had been.

She thought about what Riku had said: *I decided it was better to leave something I wasn’t sure about than to change it.*

This was, she thought, a very particular kind of attentiveness. Not the attentiveness of someone who asked what you needed. The attentiveness of someone who paid close enough attention to infer what you needed, and then chose not to act on the inference unless they were certain, because acting on an uncertain inference about what someone wanted was a form of presumption.

She had not previously encountered this in any category of person she had spent time with.

She filed it. The same folder as the notation about sub-bass localization, the same folder as the thirty-second recording on his phone, the same folder as the image of him standing in a pond in clean clothes because he had broken something.

*Ongoing.*

The evening came in from the sea, cooler than the afternoon, the light going to the particular gold that existed in the last hour before sunset. She stayed at the pond until the light was nearly gone, recording the transition — the specific sound of a campus going from afternoon to evening, the shift in bird population, the change in traffic patterns as the evening activities began and the afternoon activities ended.

When she finally stood to leave, she had four more sounds in the recorder and a number that had shifted again.

Fifty-seven sounds. Forty-three remaining.

Thirty-eight days of spring.

She walked back through the evening campus with the recorder in her hands rather than her bag, held out in front of her slightly, the way she had carried it since she was nineteen and had learned that the world was worth documenting. The way her grandmother had shown her, in the last summer, in the amber light of a persimmon tree.

Record everything.

She was recording.

The evening campus gave her: the light going out in the language labs, one window at a time; a group of third-years sitting on the steps outside the student union, their voices carrying in the cooling air; the specific sound of the groundskeeper’s cart going back to the maintenance building for the night; the bell at the gate, rung at seven for no reason anyone had ever been able to explain, a tradition so old that the tradition’s origin had been forgotten, kept alive simply by its own continuity.

She got all of it.

At the entrance to the dormitory building, she met Sumi coming out with a bag of convenience store purchases and the expression of someone who had been looking for her and had given up and was now pleasantly surprised.

“Your recorder’s back,” Sumi said, looking at the device in Sera’s hands.

“Yes.”

“Riku fixed it.”

“Yes.”

“And improved it,” Sera added, because that was accurate and accuracy mattered. “The condenser element is new. Better frequency range.”

Sumi looked at her. The look that stored information.

“Dinner?” Sumi said.

“Dinner,” Sera agreed.

They went inside. The dormitory was full of its evening sounds: televisions and music and the particular echo of a building with many people in it, all of them doing different things, all of them producing different sounds, layering together into the specific acoustic of a place that was also a home.

Sera recorded none of it. Some sounds were not for the project. Some sounds were just for being in.

But she kept the recorder in her hand until she reached her room, and she set it on the desk where she could see it from the bed, and she went to sleep that night to the familiar sound of Sumi’s breathing and the distant sea and, faintly, the bell at the gate that rang at seven and at two and at times no one had ever documented.

The scratch on the recorder’s screen caught the nightlight.

Exactly where it had always been.

Aki had a photograph of Riku’s face taken at nine thirty-two on a Saturday morning in the Repair Lab, which Aki would later describe as one of his more technically accomplished and personally useful images of the year.

The photograph captured Riku standing at the bench with the seventh iteration of the haptic chamber in his hands, looking at the space where Sera had just been standing. The expression in the photograph was not, on the surface, dramatic. There was nothing in it that would look like anything to a stranger.

To Aki, who had been photographing Riku for three years and had extensive comparative data, it was interesting.

He had been in the lab for a different reason — returning a cable he had borrowed, which he could have done at any time but had done at nine-thirty because he had a feeling about nine-thirty Saturday mornings in the Repair Lab lately, and his feelings about timing had been good for as long as he had been taking photographs, which was most of his life.

He had arrived at nine-twenty-nine. He had taken two photographs of the lab’s general state of productive disorder, which was genuinely interesting from a compositional standpoint. He had been preparing to take a third when Sera Lin had come in and picked up the recorder and the whole texture of the room had changed in a way that was not dramatic and was not visible exactly but was completely legible to someone who paid the kind of attention Aki paid.

He had not photographed Sera. That was not his to document.

But after she left, he had taken the one photograph. Riku at the bench with the model, looking at the space she had occupied.

Aki looked at the photograph on his camera screen as he walked back to the photography building.

The expression was: careful. Concentrated. The expression of someone trying very hard to understand the properties of a material they had not worked with before, and finding it more interesting than any material they had previously encountered.

Aki scrolled through his library to a photograph from two weeks earlier: Riku at the Glass Cafeteria, mid-conversation, looking across the table at something that was not in frame. The same expression, he noted. The specific and careful quality of attention directed at something being studied not as an object but as a phenomenon.

He stopped walking.

He thought about the photograph from the cafeteria, which had been taken three weeks into whatever was currently happening, and the photograph from this morning, which was five weeks in.

He took out his phone.

Aki to Riku: *how’s the haptic chamber.*

Riku: *good. floor panels are done. meeting with adachi this week.*

Aki: *good. how’s the recording project.*

A pause. Longer than necessary for a text message.

Riku: *57 sounds. 43 to go. she got the recorder back today.*

Aki looked at the message for a moment.

*she got the recorder back today.*

Not *Sera got her recorder back.* Not *the primary recorder is repaired.* The possessive was entirely natural, the way you said *her recorder* when you knew whose recorder it was in the way you knew things that mattered.

Aki put his phone away.

He filed the two photographs in his personal archive, in a folder he had labeled with the date range at the start of semester, and he walked the rest of the way to the photography building thinking about the temporal structure of certain kinds of attention and how long they took to declare themselves and whether the person doing the attending was usually the last to understand what it was.

In his experience, yes.

In his experience, almost always yes.