The RehabTrack alert arrived at 6:14 AM, before Mitsuki had finished her first cup of barley tea. Patient 7, session 6: reconsolidation index holding at 0.73, twelve percent above pre-study baseline, plateau confirmed across six consecutive sessions. The system flagged it as a positive deviation requiring annotation.
She set down her cup on the clinic's narrow counter — recycled polymer, slightly warm from the building's thermal regulation — and pulled up the full longitudinal view on her display. The curve was unmistakable. Flat for fourteen months, then a step function upward beginning exactly when Gu disconnected his piezoelectric sensor from the corridor wall.
Mitsuki had never met Gu. She had read his post-study document on the notice board outside the administrative wing — a single page, printed on standard clinic stock, describing the conclusion of a three-year acoustic frequency study conducted in the east corridor of Building 4. The document was precise and modest: measurements taken, baseline established, study terminated, equipment removed. It mentioned a final reading of 18.4 Hz, up from the long-term average of 18.3. It did not mention what that shift might mean for the seven reconsolidation therapy rooms whose environmental substrates had been calibrated against the old number.
That was Mitsuki's problem now.
Patient 7 arrived at 9:00 AM. Dae-won, thirty-four, a former experience composer who had decoupled eighteen months ago after his own compositions triggered Qualia Dependency Disorder symptoms — the particular cruelty of building emotional states so precisely engineered that your own unaugmented affect felt like static afterward. His CouplingScore had been 94 with compositional AI, which meant studios still sent recruitment messages to his dead coupling address. The RehabTrack system logged each one as a potential relapse trigger.
"The room feels different," Dae-won said, settling into the reconsolidation chair. The chair's haptic array — seventy-two independently addressable pressure points designed to simulate the somatic component of emotional memory — hummed at its initialization frequency. "Warmer. Not temperature-warm. Space-warm."
Mitsuki noted the word. Warm. Not a term Dae-won had used in any previous session. She pulled up his linguistic profile on her secondary display: eighteen months of session transcripts analyzed for vocabulary drift, emotional range markers, and coupling-residue phrasing. The system tracked words that decoupled patients stopped using — usually coupling-specific technical vocabulary like gyeolhab, donggihwa, gongmyeong that carried hedonic associations — and words they started using, usually spatial and somatic terms that signaled reintegration with unaugmented perception.
"Warm" was new. It appeared nowhere in his prior sessions. But it appeared in Chae's notes — Nalgeot Chae, the physiotherapist who had participated in Gu's corridor study. In her final protocol annotation, the one written in blue ink between Sections 2 and 3, Chae had described the corridor after the study ended as "warm with accumulated listening."
Two people who had never met, using the same word to describe the same building, for reasons neither could articulate.
Mitsuki did not mention this to Dae-won. Instead, she initiated the standard reconsolidation protocol: controlled re-exposure to the emotional memory traces that coupling had distorted, with the haptic array providing somatic anchoring. The chair's frequency — calibrated, until last week, to 18.3 Hz — now ran at 18.4. The difference was 0.1 Hz. The difference was everything.
"Tell me about the compositions you miss," Mitsuki said. This was the prompt for trace activation — the moment when decoupled patients re-encountered the emotional architecture they had built while coupled, now experienced through unaugmented neural substrate alone.
Dae-won's hands moved on the armrests. Composer's hands — they still shaped air when he talked about compositions, sculpting the phantom qualia he could no longer generate. "There was one. Haeje — dissolution. I designed it for twelve participants, synchronized to within forty milliseconds. The emotional payload was grief experienced simultaneously with its own resolution. You could feel your loss and your acceptance of the loss in the same neural moment. That is not possible in uncoupled cognition. The circuits are mutually inhibitory."
"And now?"
"Now I feel grief. Just grief. And it takes as long as it takes." He was quiet for a moment. The haptic array held its frequency. "Is that recovery?"
"What do you think?"
"I think recovery is when you stop comparing." His eyes were closed. "I haven't stopped comparing."
The RehabTrack system registered his galvanic skin response as elevated but within therapeutic range. The reconsolidation index ticked from 0.73 to 0.74. Mitsuki noted it. One hundredth of a point. One session's worth of a man learning to feel things at the speed they actually happen.
At 10:30, between patients, Mitsuki walked to the east corridor.
She had no clinical reason to be there. Her reconsolidation rooms were on the third floor of Building 4's west wing; the east corridor was ground level, administrative, a passage between the intake area and the records archive. But she had been walking here every morning for the past week, ever since she noticed the step function in Patient 7's data and traced it backward through the building's environmental logs to Gu's final measurement.
The corridor was empty. It was always empty at this hour — the administrative staff started at 11:00 on Saturdays, and the archive was accessible by keycard only. Mitsuki stood in the passage and listened.
18.4 Hz was below the threshold of conscious hearing. You could not hear it the way you heard a voice or a door closing. You felt it as a presence — a subtle pressure in the chest, a quality of the air that made silence feel inhabited rather than empty. Mitsuki had read the acoustic literature: infrasound at this frequency could induce feelings of unease, awe, or spiritual presence, depending on the listener's state and the resonance characteristics of the space. But the corridor did not feel uneasy or awesome. It felt settled. Like a room that had been well-used and was resting.
The occupancy display on the wall showed zero in amber numerals. The building's SustainGrid — the integrated environmental management system that regulated temperature, humidity, and air exchange across all four buildings of the clinic complex — reported nominal across every metric. It did not report frequency. Frequency was not a parameter the SustainGrid managed. Frequency was what the building did on its own.
She thought about what Gu's document had not said. It had not said why the frequency had risen. It had not speculated about whether the building was responding to the removal of the sensor, or to the absence of the observer, or to some structural shift unrelated to the study entirely. It had simply recorded: 18.4 Hz. Context unknown. Interpreter departed.
Gu had left the number and walked away. The number had stayed.
At 11:15, Practitioner Seo knocked on Mitsuki's office door.
"I have a question about your memo," Seo said. She was holding the recalibration document Mitsuki had submitted to Director Yun the previous week — the one proposing that all reconsolidation protocols in Building 4 be re-baselined against the current environmental frequency rather than the historical one. Yun had approved it on Thursday, adding a note about quarterly reviews. The first formal review was scheduled for April.
"The sustained contact technique," Seo said. "Section three of your memo references a manual intervention method for patients whose reconsolidation plateaus. You call it sustained contact."
Mitsuki blinked. She had not written about sustained contact. She pulled up her memo on her display and searched. The phrase was not there.
"I didn't write that," she said.
Seo frowned, then checked her own copy. "Sorry — I read it in Park's case notes this morning. I assumed it came from your protocol revision." She paused. "Park calls it sustained contact. Kim calls it threshold hold. I've been calling it maintained pressure but I think we're all describing the same thing."
Mitsuki knew what they were describing. It was Chae's technique — the overhang, the moment of holding contact with tissue past the point where clinical protocol said to release, allowing the body's resistance to become information rather than obstacle. Chae had developed it during the corridor study, refined it through 58 logbook entries, and declared it complete in a protocol document that sat on a shelf in the physiotherapy department. She had never presented it. She had never published it. She had never named it for anyone else.
But her hands had done it in front of patients, and patients had bodies, and bodies talked to other practitioners' hands.
Three different practitioners. Three different names. One technique that had migrated without its name, without a paper, without a presentation. Through practice. Through the body's memory of what worked.
Mitsuki wrote this down in her session notes: Post-study propagation. Findings that spread through practice rather than publication. The technique arrives before the terminology. The body teaches faster than the document.
She looked at what she had written and felt a specific vertigo — the recognition that she was not observing propagation from outside. She was part of it. Her memo to Yun, her recalibration protocol, her daily walks to the east corridor — these were also the study's findings spreading through a practitioner who had never been part of the study.
At 2:00 PM, Patient 7 returned for his afternoon session. Dae-won sat in the reconsolidation chair and closed his eyes.
"The word I used this morning," he said. "Warm. I've been thinking about it."
"Tell me."
"When I was composing — when I was coupled — I could generate warmth. Precise warmth. I could target a specific neural population in an audience and activate exactly the warmth response I wanted. I knew the frequency, the amplitude, the duration. I could make a room of eighty people feel the same warmth simultaneously, synchronized to within forty milliseconds." He opened his eyes. "This morning, when I said the room felt warm, I couldn't tell you what I meant. I couldn't quantify it. I couldn't reproduce it. I couldn't even tell you if it was the room or me."
"And?"
"And it was better." He closed his eyes again. "The warm I made was perfect. This warm is real. I don't know the difference yet. But my body does."
Mitsuki adjusted the haptic array — a minor shift, barely perceptible, bringing it 0.1 Hz closer to the building's current baseline. The RehabTrack system logged the adjustment as a protocol deviation. She accepted the deviation and moved on.
The system would flag it in tomorrow's report. Director Yun would see the flag and note it against the quarterly review timeline. The quarterly review would establish new calibration standards. The new standards would reflect a building that had been changed by three years of someone listening to it.
And none of this — the 0.1 Hz shift, the plateau, the technique migrating through three practitioners' hands, the word "warm" arriving in two unconnected vocabularies — none of it would appear in Gu's study. His study was over. His document was complete. His sensor was disconnected, his cabinet closed, his logbook ended at Entry 58.
Gu's study had produced findings. Mitsuki was living inside the findings' consequences.
At 4:00 PM, she walked the east corridor one more time. The light had shifted — afternoon sun through the frosted polymer panels turned the passage amber, the occupancy numerals on the wall displays showing zero in soft orange characters. The SustainGrid had lowered the corridor temperature by two degrees for the evening cycle. The air moved differently at this hour — slower, as if the building were breathing out.
She placed her palm flat against the corridor wall. The surface was cool polymer over concrete over structural steel, and beneath all of it, the low hum that was not a hum but a frequency, not a sound but a presence, not a measurement but a fact about the building that existed whether or not anyone held a sensor to it.
Gu knew this number. Chae carried this number in her hands without knowing it. Three practitioners used techniques shaped by this number under three different names. Patient 7 called this number "warm" and could not explain why. Director Yun had approved a protocol acknowledging this number as institutional reality. And Mitsuki — Mitsuki was documenting the propagation of a finding through systems that had never been part of the study.
She was not the researcher. She was not the subject. She was the evidence.
The corridor held its frequency. She took her hand away and walked back to her office, where the RehabTrack system was already preparing tomorrow's alerts, calibrated against a baseline that had quietly, permanently shifted.