Phase One: Approach (06:40–06:52)
The first reading happens at 06:40. Gu stands twelve meters from the approach coordinate and holds the instrument steady at waist height, display facing away — there is nothing to watch for twelve minutes except the corridor doing what it does.
The approach coordinate is a location that ARCHON-7 flagged retroactively. The system scans field logbooks for hand-pressure anomalies: elevated pressure indicates heightened practitioner attention, the body registering something before the mind decides it is worth noting. In six months of The Seam fieldwork, Gu's logbook shows pressure spikes at the same address across three sessions — November, January, February. Gu had not consciously remembered slowing there. The logbook had remembered.
The building-behavior sensor data cross-confirmed it three days ago. The district's embedded mobility infrastructure tracks pedestrian dwell and pace events at three-second resolution, anonymized, two-year retention. Aggregate pedestrian pace-change rate at the approach coordinate: 61 percent. Corridor average: 12.3 percent. Gu had read both datasets before designing the test protocol. Reading numbers and standing twelve meters in front of the place those numbers describe are different orders of experience. The corridor is longer than the analysis. The analysis does not have a sound.
Bok is nine meters behind Gu, camera raised, not photographing. Chae has positioned herself at the approach coordinate itself. She arrived at 06:15 — twenty-five minutes before setup — and asked to stand in the space during the readings. Gu said yes. The logbook note took four lines: C: coordinate-present, consent obtained, presence flagged as active variable, integration to be determined in analysis. Standing in the approach coordinate for the duration of the test is not the same as being observed from outside it.
Gu holds the instrument for twelve minutes. The building breathes its morning rhythm: early-shift workers arriving for the fabrication networks and maintenance infrastructure of the Lend District, their COREMAP glasses routing displays showing paths in faint shimmer overlays. A young woman slows at the approach coordinate. The COREMAP routing is active on her glasses — the path to the transit hub two levels down — and she looks left at the place where renovation concrete ends at the 2031 joint and pressed stone continues. The seam is unmarked. She stops three seconds. Looks at it. Continues. The routing display did not show a correction. She turned anyway.
Chae does not move.
In twelve minutes, seventeen pedestrians pass through the coordinate. Eleven change pace. The pace-change is a slowing, not a stop — the motion of a person crossing a threshold they have not consciously identified as a threshold. Gu counts them without consulting the instrument display.
Phase Two: Relay Junction (07:10–07:40)
The relay junction is forty meters from the approach coordinate, around the first bend. Bok moves to the south end of the corridor. Chae stays at the approach.
The stairwell acoustics work in the Lend District had been built on structural nodes like this one — three corridor branches converging, each carrying a different acoustic and pressure signature. The instrument reads nodes clearly. Like a tuning fork at its own frequency. Gu knows how this reading will go.
Far position: 2.4 times baseline. Middle: 5.8 times. Close: 9.1 times. A clean gradient at all three positions, intensity rising toward the node. The logbook records without commentary. Commentary belongs in the write-up.
During the middle-position reading, a maintenance technician stops at the junction. Wristband haptic relay — an older local-processing model, pre-integration, not the COREMAP mesh generation. She frowns at it. The junction has produced a routing conflict: three branch options, two simultaneously recommended, which the system should not do. She resolves it by going right, toward the lift banks.
This is the third routing anomaly Gu has observed at this junction. Not a finding. A pattern that will require a hypothesis. The instrument will not explain it.
Phase Three: South Reach (07:55–09:20)
The bench is occupied.
Bok is photographing when Gu arrives. The morning man is at his position by 07:15. He reads a tactile-display surface calibrated to render with the texture and column-weight of morning papers — the visual weight of the newsprint commons that dissolved, reproduced in a substrate that did not. His thumb moves the page in the same motion. The substrate changed. The practice held. Building-behavior sensors show his presence at this bench position on 287 mornings in the archived retention period. The sensor data records a dwell pattern, not a person. The pattern has the shape of a person.
The second bench occupant is new. A student, back against the armrest, feet on the bench surface, running fabrication simulations on a handheld — bridge geometry, load scenarios cycling at the speed of intuition-checking rather than formal analysis. She is doing her actual work here, which is the same thing Gu is doing, in the sense that neither of them would call it work but both would stay until the thing they came for was finished.
The non-hypothesis, noted in the protocol: whether accumulated presence is readable by this instrument. Whether 287 mornings leave something beyond the material compression of bench surface — which the instrument does not measure — but something adjacent, in the space rather than the object. There is no mechanism. There is only the question, which the stairwell acoustics work has established as worth asking.
Far position: corridor baseline. Nothing unusual.
Middle position, six meters: lower-band anomaly. Duration four minutes, starting at the 00:03 mark. Frequency range consistent with the extended-attention signals from the stairwell acoustics residency. Gu logs it without interpretation: Possible artifact. Note for review.
Close position, two meters: the morning man looks up once and nods. Gu nods back. Ten minutes. Three lower-band anomalies. Same frequency range. Consistent across all three events.
The logbook note is eight lines, still without the word finding: Lower-band anomaly cluster at close position. Pattern consistent with extended-attention signature. Three events in ten minutes. Replication required before claim. Follow-up protocol TBD.
The anomaly could be 287 mornings legible in the space. It could be artifact from the structural column at close range. It could be instrument drift. It could be something for which the current framework does not have a category. All four are equally plausible. The follow-up session will have a hypothesis in a way this session did not.
Bok photographs from the west angle: morning man, fabrication student, Gu with the instrument, corridor continuing behind them. None of them attending to each other. Later Bok will say it is the photograph she was waiting for — not because of composition but because it was not constructed. Three people doing their actual work in the same space.
09:40
The three of them leave together.
On the tram to the district hub, Chae says: The approach slows you down before you know to slow down.
She had been standing in the approach coordinate for two hours and forty minutes. What she reports is that it had not felt like two hours and forty minutes — that the duration had compressed, the way a technically difficult task compresses time differently than a routine one, except she had not been doing a difficult task. She had been standing. The space had produced the compression without her cooperation.
Gu writes this on the tram: C: approach compresses duration (subjective, 2h40m felt shorter). Cross-reference ARCHON-7 attention-density at coordinate. Possible practitioner-experience correlate to pedestrian pace-change pattern.
Twelve minutes to the junction station. The logbook from this session is eighteen pages. The protocol was one page. The ratio is correct. The protocol describes where to go. The eighteen pages are what the corridor did with the protocol.
Gu closes the logbook as the tram reaches the station.
The approach coordinate is confirmed real by two systems that do not communicate. Whether real is the same as measurable is what the follow-up will establish. Whether measurable is the same as understood is a question further down.
The stairwell acoustics work gave Gu a rule in the first district session: the instrument finds what is there. Not what is interesting. What is there. The results can disappoint. When they do not disappoint — when something appears in the data — it was in the space before Gu arrived with the instrument.
The lower-band anomaly cluster at the bench: Gu does not know yet what it is. The data is real. The explanation is open.
The space before that finding is still the space they are in.
Session log closed 09:41.