A. Vance had been humming for three days before anyone noticed, including her.
It was T. Okoro in the adjacent bay who logged the first informal complaint — not a complaint, exactly, but a note left on the shared whiteboard in Break Room C that read: whoever is humming on the floor, please stop. some of us are trying to focus. He signed it with his employee number because that was the kind of person Okoro was. He liked things to be traceable.
No one responded. The note stayed on the whiteboard for two days and was eventually erased by the night cleaning crew.
The hum itself was unremarkable. Colleagues who were later interviewed described it variously as “a few notes,” “sort of tuneless,” “like something you’d hear in a waiting room.” One witness, , said it reminded her of a lullaby, though she could not say which one. Another, , said it sounded like a ringtone heard through a wall. None of them could reproduce it when asked.
Vance herself, when first approached by her shift supervisor on the morning of June 9th, denied humming. She was not lying. The supervisor, , noted in a supplemental report that Vance appeared “genuinely confused and slightly embarrassed.” He asked her to be mindful of it and returned to his office. Fourteen minutes later, she was humming again.
The Censor Operations floor of the Office of Memetic Containment occupied the sixth through ninth floors of a building in that had no signage on the outside and a lobby that smelled permanently of industrial carpet cleaner. Tier 3 censors — Vance’s classification — worked in open bays of six, seated at reinforced terminals with partial acoustic shielding. Their job was to review flagged media before it reached the public filtration layer: images, audio, video, text. They looked at the things other people were not allowed to look at.
There were protocols. Tier 3 censors worked four-hour shifts with mandatory 90-minute recovery periods. They wore biometric bands that tracked heart rate, galvanic skin response, and pupil dilation. They submitted to weekly cognitive baselines. Twice a year, they rotated to administrative duty for thirty days so their pattern-recognition pathways could, in the language of the OMC Wellness Manual, “return to resting architecture.”
The system was considered robust. The attrition rate was high — roughly 40 percent of new censors washed out within eighteen months — but those who remained were regarded as resilient. Vance had been on the floor for six years. Her evaluations were uniformly excellent. Her cognitive baselines were stable. Her supervisor described her as “the kind of person who makes the job look boring, which is the highest compliment I can give.”
She had no history of memetic sensitivity. No family predisposition. No prior incident tags.
On June 10th, the day after the supervisor’s conversation, the hum was audible to the entire bay. Okoro filed a formal noise complaint through the OMC internal system. The complaint was routed to Facilities, who sent a technician to check the HVAC ducting on the sixth floor. The technician found nothing. The complaint was closed.
On June 11th, two additional complaints were filed. One from , which was across the corridor and behind a fire door.
That afternoon, Vance’s biometric band flagged an anomaly during her second shift block. Her heart rate had not changed. Her skin conductance had not changed. Her pupil dilation had not changed. What the band flagged was rhythmic microvariation in her jaw tension — consistent, the system noted, with subvocal repetition.
She was humming with her mouth closed.
The shift medic pulled her off the floor at 3:40 PM and administered the standard Tier 3 exposure screen: a 22-item questionnaire, a short cognitive battery, and a four-minute visual-attention test. Vance passed all three. The medic noted: Subject is alert, oriented, and cooperative. Denies awareness of vocalization. No evidence of cognitive impairment. Recommend observation and follow-up in 48 hours.
Vance returned to work the next morning. She told the medic she felt fine. She told Okoro she was sorry about the noise. She told her shift supervisor she thought it might be stress.
By noon she was humming again, and she did not stop for the rest of her life.
The OMC’s Incident Review Board convened on June 14th, four days after the formal complaints began and six days after Vance had first been heard. The board comprised three people: the division’s senior operations manager, a staff psychologist, and a memetic safety officer. Their initial assessment, documented in the file tagged 2073-T-441, was that Vance was experiencing an “idiopathic vocalization pattern, likely stress-related, with no current indication of memetic origin.”
They recommended continued monitoring, a temporary reduction to Tier 2 material, and a referral to the employee assistance program.
The safety officer, , added a dissenting footnote. It read: I want it on the record that I asked what she was reviewing in the 72 hours before onset, and that the answer I received was that the specific material logs for that period were not available due to a system migration. I do not find this explanation satisfactory.
The footnote was included in the file. No action was taken on it.
What Vance had been reviewing in the 72 hours before onset was later reconstructed from backup server logs recovered during the . The material included 340 flagged images, 71 audio fragments, and 16 video files. Of these, 14 audio fragments had been tagged as Priority-A — material suspected of containing active memetic properties rather than merely derivative or referential content.
Eleven of the fourteen had originated from a single source submission intercepted at the filtration node four years earlier. They had been in the review queue since 2069. They had been assigned to Vance’s bay as part of a routine backlog clearance.
The fragments were, according to their metadata, partial recordings of ambient sound from a public transit platform. They contained crowd noise, mechanical sounds, a public-address system, and, beneath it all, a tonal pattern that the submitting analyst had described as “possibly coincidental, possibly structured.”
Vance had listened to each fragment twice, per protocol. She had marked nine of the eleven as “no memetic content detected” and forwarded them for archival. She had flagged two for secondary review with the note: Unusual tonal layering in background. Likely acoustic artifact but recommend verification.
She filed the flags at 4:52 PM on June 6th. She clocked out at 5:15 PM. Somewhere between the parking garage and her apartment — she lived alone, fourteen minutes away by bus — the humming began.
Within two weeks of Vance’s removal from the floor, three additional censors in adjacent bays reported involuntary vocalization. Two of them had not reviewed the same material. One of them had never worked audio files at all.
The Incident Review Board reconvened on June 28th. The dissenting footnote was no longer dissenting. The file was reclassified from “workplace incident” to “potential memetic exposure event.” Vance and the three secondary cases were placed on indefinite medical leave. The sixth floor was closed for acoustic decontamination — a procedure that, at the time, had been used only twice before in the OMC’s history.
The decontamination team found nothing in the air ducts, nothing in the walls, nothing in the acoustic shielding panels. They disassembled Vance’s terminal and found nothing in its speakers or its wiring. They tested every surface in Bay 6 for residual vibration patterns and found none.
The hum was not in the building. It was in the people.
Vance was transferred to the residential monitoring facility on July 3rd. She went willingly. She told the intake officer she understood the situation and that she was not frightened, which the officer noted as “appropriate affect, though the subject’s continued vocalization during the interview was noted.”
She hummed through the intake process. She hummed through dinner. She hummed in her room that night with the door closed, and the overnight nurse who checked on her reported that the sound was “steady and quiet, like a refrigerator running in the next room.”
She was not in distress. Her cognitive function remained within normal limits. She could hold a conversation, read a book, complete a crossword. She simply could not stop producing a sound that no one could quite describe and no one could reproduce and no one, in the end, could trace to any known stimulus.
The OMC’s classified summary, written three months later and not declassified until , contained the following assessment:
The vocalization pattern observed in Vance, A. and three secondary subjects does not match any catalogued memetic signature. It does not appear to be a direct reproduction of source material. The pattern may represent a novel transmission vector in which the memetic payload is not the sound itself but the impulse to produce it. If this assessment is correct, the implications for current containment doctrine are significant, as the existing protocol assumes that memetic content must be perceived to be transmitted. The Vance case suggests that it may be sufficient for content to be processed — even by a trained censor operating within safety parameters — for transmission to occur.
We do not currently know what Vance is humming. We do not know if it matters.
A. Vance remained at for eleven months. She was described by staff as a cooperative and pleasant resident. She read extensively. She took walks in the facility’s enclosed garden. She wrote letters to her sister, though the letters were reviewed and, in several cases,
She never stopped humming. Not while sleeping. Not while eating. Not while speaking — the hum threaded beneath her words like a second voice, so faint that new staff sometimes didn’t hear it until someone pointed it out.
The three secondary cases recovered within eight weeks. Their humming ceased as suddenly as it had begun. They returned to work. Two of them transferred out of Censor Operations within the year.
Vance did not recover.
On June 2nd, 2074 — almost exactly one year after onset — she was reclassified from “medical leave” to “permanent memetic injury, Class 2.” The reclassification entitled her to a pension, housing assistance, and ongoing monitoring. It also removed her security clearance and barred her from entering any OMC facility for the rest of her life.
She moved to a supervised apartment in . She lived quietly. She kept a small balcony garden. She saw her sister on weekends. She hummed.
The last entry in file 2073-T-441, dated August 2076, is a brief note from a field officer conducting a routine welfare check. It reads:
Subject in good health. Apartment clean and well-maintained. Subject reports no changes in condition. Vocalization continues at baseline volume and pattern. Subject asked me, politely, to stop writing about her. She said she was not a file. I told her I understood.
When I left, she was standing at the kitchen window. I could hear her from the hallway. I could hear her from the stairwell. I was in my car before I realized I could still hear her, and then I realized I couldn’t — I was hearing myself. I had been humming. I don’t know for how long.
I am filing this note and requesting a cognitive baseline screen at my earliest convenience.