The notebook is in the kitchen drawer, between the takeout menus and the envelope of spare fuses for the acoustic monitoring junction box on our floor.
I put it there this morning. Not thrown away — accessible. There is a difference between finished and abandoned, and the difference is whether the drawer is labeled. I did not label it.
The notebook is small, hardcover, graph paper. Eighty-seven pages used. The first entry is from fourteen months ago, the day the corridor study began, and it reads: 42 Hz, refrigerator compressor, continuous, amplitude varies with thermal cycling. Clinical. The kind of entry a person makes when they believe they are documenting something external.
The last entry is from this morning. One line, written standing at the counter after the doenjang-jjigae had cooled enough to eat: The study ended. The ears did not. Both of these are fine.
Between those two entries are eighty-five pages of a woman teaching herself to hear a city she had lived in for nine years without listening.
The corridor study was Bok's idea, originally. He noticed the gap — the physical gap, the place where two sections of corridor infrastructure do not quite meet, where the Seam's acoustic monitoring grid has a dead zone roughly four meters wide. The monitoring grid is the neighborhood's nervous system: thin sensor wires strung between buildings, junction boxes with amber status lights, all of it feeding into the District Acoustic Profile that the city updates every seventy-two hours. The Profile determines noise ordinance thresholds, construction permits, even rental pricing — apartments in high-resonance zones cost less because the city considers them acoustically compromised. My apartment is in a medium-resonance zone. My rent reflects this.
Bok noticed the gap. I noticed what was inside it.
Inside the monitoring dead zone, the acoustic environment is unmanaged. No sensors adjusting the District Profile. No algorithmic noise compensation from the Seam's environmental systems. Just raw sound — the building's actual voice, unfiltered and unoptimized. I stood in that gap for eleven minutes on a Tuesday afternoon in January and heard my neighborhood for the first time.
The refrigerator compressor three floors up: 42 Hz, cycling. A water pipe somewhere in the walls: intermittent, pressure-dependent, roughly 100 Hz when active. Traffic from the elevated road to the east, filtered through two buildings and a parking structure: a broadband wash centered around 250 Hz, rising during rush hour, falling at night but never disappearing entirely. The Seam's own infrastructure — the monitoring grid itself — producing a 50 Hz hum from its junction boxes that I had lived alongside for nine years without consciously registering.
I went home and started the notebook.
The study lasted eleven months. I mapped 347 discrete frequency sources across a six-block radius. I categorized them: infrastructure (the monitoring grid, HVAC systems, the automated waste collection that runs at 3 AM), transport (elevated road, the electric delivery vehicles that produce a 2 kHz whine from their regenerative braking), biological (birds — seasonal, mostly 2-8 kHz; human voices — variable, clustered around 300 Hz fundamental), and what I called ambient residuals — the frequencies that exist because other frequencies exist, the interference patterns and standing waves that the Seam's architecture generates without intending to.
I became precise. I bought a portable spectrum analyzer from the Technical Commons — a six-month lend, moderate collateral, my first premium-tier borrowing. I calibrated it against the District Acoustic Profile and found thirteen discrepancies in the first week. The Profile was wrong about my neighborhood. Not drastically wrong — the sensors are good hardware — but wrong in the way that any model is wrong about the thing it models. The Profile captured the statistical average. I was capturing the Tuesday afternoon at 3:47 PM when the compressor cycled off and the water pipe surged and for four seconds the dominant frequency in my kitchen was the 50 Hz hum of the monitoring grid that was supposed to be measuring the silence.
I reported the discrepancies. The Acoustic Governance Board acknowledged them, filed them, did nothing. The Profile updated on its seventy-two-hour cycle and the discrepancies appeared in the next dataset as minor fluctuations within acceptable parameters. My eleven minutes in the gap had produced data that the system absorbed without changing.
This did not bother me. I was not trying to change the system. I was trying to hear the city.
The grocery store was when I understood what had happened to me.
Three weeks after the study ended — the formal end, the day I returned the spectrum analyzer to the Technical Commons and watched my LendScore tick upward from the early return credit — I went to buy groceries. Saturday afternoon. The store on the corner of Jang-su and Eighth, the one with the narrow aisles and the freezer section that everyone complains about because the compressors are too loud.
The compressors are not too loud. They run at 60 Hz with a harmonic series extending to roughly 300 Hz, and their combined output at ear level in the frozen foods aisle is approximately 65 decibels, which is within the Seam's commercial zone noise threshold. I knew this. I had measured this, seven months ago, entry forty-three in the notebook.
But standing in the aisle with a bag of frozen mandu in one hand, I did not think about the measurement. I thought: the freezer aisle is a drone note. The entire section is a continuous 60 Hz foundation, and everything else — the cart wheels on tile at 12 Hz, the announcement system cycling through its queue, the compressor relay clicking every forty seconds — sits on top of that drone like overtones on a fundamental. The grocery store is an instrument. The frozen foods section is its lowest string.
I stood still for eleven seconds. Someone said excuse me. I moved.
In the oil aisle I picked up a bottle of sesame oil and felt it vibrating. The bottle was glass, thin-walled, and the shelf it sat on was connected to the same structural frame as the freezer compressors. The vibration was tiny — subliminal, maybe 0.1 millimeters of displacement — but my hand knew it was there. My hand had learned to feel things that my hand had always been able to feel but had never had permission to notice.
I put the sesame oil in the cart. I paid. I walked home. The streetlights had turned on by the time I reached my building — the old sodium vapor ones the city has not yet replaced on my block — and when the one at the corner cycled off for its automatic restart I heard the ghost. The 50 Hz hum it had been producing disappeared and left a shape in the air, an acoustic afterimage, and I stood on the sidewalk for six seconds listening to the absence of a sound I had been ignoring for nine years.
The study was over. My ears were not.
I tried to draw a line. This is the thing I spent the last three weeks doing, the thing the notebook documents in its final twenty pages: the attempt to separate hearing from living.
The logic was simple. The study had a purpose — to map the acoustic environment of The Seam's monitoring gap. The purpose was fulfilled. The data was collected, the discrepancies reported, the analyzer returned. The study was a bounded project and bounded projects end. Therefore the hearing should end. I should go back to the person I was fourteen months ago, the one who lived in a medium-resonance zone and paid rent accordingly and did not know that her refrigerator was 42 Hz.
This did not work.
The hearing is not a skill I acquired. It is not a tool I borrowed and can return. It is a permanent alteration — the way learning to read is permanent, the way you cannot look at a word and not read it. I cannot stand in the frozen foods aisle and not hear the drone. I cannot hold a glass bottle and not feel the building's infrastructure humming through the shelf and into my fingers. I cannot walk past a streetlight and not know that its absence has a shape.
I tried logging. If I could not stop hearing, I could at least contain it — turn it back into data, clinical entries, the safe format of frequency and amplitude and source. Entry 312: Kitchen, 2 AM, compressor off, acoustic afterimage, duration approximately 4 seconds. Entry 326: Elevated road, Tuesday 5:30 PM, broadband wash, estimated 72 dB at apartment window. Entry 341: Park bench, Sunday, diesel bus idle at 85 Hz, children, ball bounce frequencies not measured.
Entry 341 was when I stopped logging.
The child dropped a ball. I watched the child, not the ball. I watched the child laugh and chase the ball across the grass and I did not calculate the bounce frequency. I heard it — I could not not hear it — but I did not calculate. The hearing happened. The analysis did not. The child was a child and the ball was a ball and the grass was grass and the bus was 85 Hz and all of these were true at the same time and none of them needed to be written down.
Bok closed his exhibition today. He texted: one line, no details. I did not ask how it went. I know how it went. He pointed at something until the pointing became the thing, and then he stopped. I have done the same thing with frequencies. The study was the pointing. The city was always making these sounds. I was always capable of hearing them. The notebook was the finger, and now the finger is in the drawer.
Tomorrow I will go to the grocery store. Not for data. For sesame oil.
The bottle will vibrate in my hand. The freezer aisle will drone. The cart wheels will rattle at 12 Hz on the tile and someone will say excuse me when I stand still too long in the frozen foods. I will hear all of it. I will not log any of it.
The monitoring grid will update the District Acoustic Profile and my neighborhood will be quantified and my rent will or will not change and the 50 Hz hum from the junction box on my floor will continue whether or not I acknowledge it. The city does not need me to listen. The city has sensors.
But the sensors do not stand in the freezer aisle and feel the sesame oil vibrating. The sensors do not hear the ghost when the streetlight cycles off. The sensors do not lie awake at 2 AM and listen to the compressor and know that the silence after it stops has a shape.
The sensors measure. I hear. The notebook is in the drawer. The ears are not.
I cook dinner. Rice, sesame oil, the last of the doenjang from the jar I bought three weeks ago at the frequency-saturated grocery store. The meal is good. The kitchen hums. I eat standing up at the counter, the same counter where the notebook used to sit, and I do not miss it. The counter is a counter. The meal is a meal. The 42 Hz is 42 Hz and the evening is the evening and both of these are fine.