Dead Reckoning

The second read was slower.

The first pass through RN-0047-NL's file, I'd been tracking the shape of it, the arc from early sessions through the adjustments and then the truncation point. The second pass was more deliberate. I went entry by entry, session log open beside the file, comparing milestone dates against my own.

Rate separation put their session at thirty. Mine at thirty-one. Close enough to be coincidence, not far enough apart to mean anything on its own. Concurrent-threshold framing landed at session fifty-two. Mine landed somewhere in the mid-forties, depending on how I counted the attempts that had produced partial results without clean repeatability. Their numbers ran slightly ahead of mine across the board, but the structure of the progression, the sequence of failures and adjustments, the specific variables they'd chosen to isolate first, matched mine at a level that moved past coincidence into something I couldn't name with one word.

I made a table in the session log. Their milestone, their session number, my corresponding milestone, my session number. Twelve rows filled the table. The average gap between their sessions and mine was 2.3 sessions, theirs ahead, consistent across all twelve data points. That was the detail I kept returning to, because a consistent ratio implied a consistent cause. Either we were very similar practitioners working the same problem under similar conditions, or something outside both progressions was shaping them from the same pressure point. Neither option was comfortable to sit with.

I closed the table, reopened it, read the columns again. The numbers did not rearrange themselves into anything more neutral.

The disc was still at eight degrees. The session log was still open. Outside the Annex's east corridor, the texture anomaly was roughly four centimeters wider than it had been five days ago, which I knew because Nadia had measured it with a three-point reference system precise enough that I'd copied her methodology rather than develop my own. The measurement was in the filed secondary administrative report, cross-referenced against my session log, submitted to Practitioner Record with a routing note to Facilities. That report was somewhere in the administrative channel now, either in a queue or already forwarded or sitting in a folder no one had opened yet. I had no practical way to find out which.

The sixth linked record in the predecessor's file had returned the same two lines as the other five. Classification Level 5, Tier 3 Required. I'd tried five of them the night before. The sixth I'd left because I was tired, and because opening one more locked file that told me nothing would have made the tired worse in a specific way I recognized as unproductive. Now I pulled it up.

Classification Level 5, Tier 3 Required. The same two lines with no variation. I stared at the access denial longer than was useful, then closed it. The table was still in the session log, my twelve-row comparison sitting there with its 2.3-session average gap. I added a thirteenth row. Predecessor's Tier access at truncation point, unknown. My Tier access at current session, Level 4, Tier 2. The gap in that column had no number I could fill in, so I left the cell blank and wrote "locked" in the margin.

Thirteen rows filled the table. Twelve with data, one with a blank, and the disc at eight degrees, the cord straight, the tile flat.

The real problem was that I was procrastinating, and knowing it didn't make it stop. The session log was open, attempt fifty was queued, and I was comparing rows in a table I'd already read three times. The reluctance came from a specific informational gap, not from fear in any clean sense. I knew the predecessor had reached approximately my position. I knew the object in their hall had changed behavior around the same point. I knew two practitioners before them had documented the same shift. And I knew that all three of those practitioners were no longer here. What I didn't know was whether the causal chain ran through the practice itself or through something else, and I couldn't find that out by sitting still.

Attempt fifty meant null state from simultaneous initialization, both states brought up together, no ordering priority. Duration target, beyond forty-three seconds, which was the predecessor's confirmed ceiling before day eighty-one. My personal ceiling was forty-eight seconds, achieved on attempt forty-eight.

I initialized.

The null arrived at roughly the same speed it always did with the simultaneous method, which is to say fast enough that the first two seconds were mostly transition. The inverse came up alongside it without the ordering friction that the old sequential approach always produced. By three seconds I had both states, which the disc confirmed at its usual angle, and I moved into the duration phase.

Forty-three seconds was what I was watching for, a reference point rather than an endpoint.

Thirty seconds in, I ran the standard internal audit: null stable, inverse stable, neither feeding into the other in ways I hadn't authorized. The disc held. The cord along the east wall was straight.

Forty seconds came and went. I kept the audit running. The predecessor had documented this range as unremarkable, methodologically speaking. Their notations from sessions in the forty-second window were brief and professional, nothing that flagged instability or object-behavior change. I was tracking the disc closely enough that I caught it when it shifted a half-degree toward the wall at forty-one seconds, then corrected back to eight degrees at forty-two.

Forty-three. Past the marker. Both states still present. The disc was at eight degrees and staying there, consistent with all my prior data and inconsistent with nothing. I kept going.

Forty-eight seconds, and I held. That was my personal ceiling. I moved through it without crossing anything, just continued duration, inverse and null coexisting at whatever threshold they'd settled into.

Fifty seconds. The disc shifted again, three degrees toward the east wall, then back. I logged it mentally and held the states.

Fifty-two seconds arrived. The predecessor had reached fifty-two-second coexistence by their session fifty-two, according to the file. I was at attempt fifty. Another data point for the table.

Fifty-five seconds, and something changed.

Not in me and not in the states. The wall.

The east wall of my hall had a texture anomaly approximately twelve by eight centimeters in size, located about forty centimeters from the floor, which I had measured and logged and cross-referenced in a secondary administrative report. I had its location fixed in my sense of the room, background presence, not active attention.

At fifty-five seconds, it pulled my attention the way nothing had in the session. It made no sound. I couldn't describe the movement cleanly. The anomaly was simply present in a way it hadn't been three seconds earlier, its location in my peripheral awareness shifting from background to foreground without any clear mechanism.

I held the states for two more seconds while I decided what to do, then dropped both cleanly at fifty-seven.

The disc returned to its neutral position. I wrote in the session log: Attempt 50. Duration: 57 seconds. Both states achieved simultaneously. Null and inverse stable through 57 seconds. At approximately 55 seconds, east wall anomaly entered active awareness without clear perceptual trigger. Disc shift noted at 41-42 seconds (half-degree east, corrected) and at 50 seconds (three degrees east, corrected). No physical contact with anomaly area. Session concluded voluntarily.

I sat with the log for a minute before adding one more line. The attention-shift at 55 seconds does not have a clear perceptual explanation. Not sound, not movement, not light change. Will note as data point, not as cause.

Then I sat back against the wall and looked at the disc.

Nadia was in the east corridor when I came out. Either it was coincidence or she'd been timing my sessions from somewhere she could hear the door, and knowing Nadia, it was probably both: she kept to her own schedule, but her schedule had a way of intersecting with things she wanted to check on.

"Fifty?" she said, without looking up from a measurement she was taking.

"Fifty-seven." I leaned against the doorframe. "The disc moved twice. Brief corrections."

That made her look up. She had a thin measuring rod in one hand and was standing with her back to the anomaly's section of wall, which meant she was either measuring something else or she'd already finished and was waiting.

"Toward the anomaly?" she asked.

"East wall. Yes. Both times it corrected without me doing anything."

She wrote something on the notepad she kept on a loop around her wrist. Her handwriting in the field was the same as her formal notation, smaller and more uniform than mine, numbers clear enough to read at arm's length. She finished writing and looked at the anomaly section, then back at me.

"My measurement today is four and a half centimeters wider than Monday," she said. "Which is slightly outside the four-centimeter interval we had."

The secondary administrative report I'd filed used the four-centimeter figure. If the rate had shifted, the filed figure was already inaccurate. "Does it feel linear to you, or is it accelerating?"

"I only have twelve days of data. I can't distinguish between linear with noise and a slight curve." She capped her measuring rod. "But four-point-five against four doesn't read as noise."

It probably wasn't. I filed the new measurement mentally alongside the disc corrections, the predecessor's object-behavior change at a similar duration threshold, the forty-centimeter-from-the-floor position of the anomaly. Everything was east wall. Everything oriented toward the same location.

"I'm going to request a direct Facilities inspection," I said. "Not just the forwarding note. An actual request, named to whoever handles physical anomalies in the Annex."

Nadia tilted her head. "Do you know who that is?"

"No. I know the routing channel."

She didn't respond right away, her measuring rod still in her hand. "Fass might know the office. He's been here longer than us."

That was true and practical. Fass had joined the Annex intake four months ahead of our cohort and had spent considerably more time working out the administrative geography of the place. If anyone knew which office actually received Facilities reports rather than which channel they were nominally filed to, it was Fass.

"I'll ask him," I said.

Nadia turned back to the wall and pressed two fingers lightly against the surface a few centimeters from the anomaly's edge. She did it the way a carpenter checks a joint, not to diagnose anything but to register what's there. Then she pulled her hand back.

"The texture is different at the boundary," she said. "Not just the anomaly itself. The six or seven centimeters adjacent to it." She wrote that down too. "I didn't note that before. I don't know if it's new or if I wasn't testing for it."

"Log it as newly observed, best current date." That was the methodology the predecessor had used for exactly this situation: date of observation, not inferred date of origin. It kept the record accurate without inventing history.

"Right." She finished writing, looped the notepad back around her wrist. "Fifty-seven seconds."

"Fifty-seven."

"And the anomaly pulled your attention at fifty-five."

"Something did. I can't tell you the mechanism."

She nodded slowly, the way she did when she was categorizing something as important without yet settling on which category it fit. Then she said, "Did you look at it? During the session, I mean. When it pulled your attention."

I hadn't. I'd stayed focused on the states and dropped them when I decided I had enough data for the session. "No. My eyes were on the disc."

"So you didn't see whether the disc movement and the attention-shift were simultaneous."

Going back over the sequence: disc had moved at forty-one, corrected at forty-two. It moved again at fifty. The attention-shift came at fifty-five. The disc hadn't moved at fifty-five, as far as I'd tracked. Those events weren't simultaneous. But that didn't mean they weren't related, only that the relation wasn't direct.

"The disc moved at fifty seconds," I said. "The attention-shift was at fifty-five. Five-second gap."

"Okay." She wrote that down too, the exact gap, and underlined it once. Nadia reserved underlining for things she expected to matter later. "I'll add it to my record."

We stood in the east corridor with the anomaly section of wall between us. The building ran its usual background sounds: ventilation, something in the floor of the level above, the faint pressure-equalization hiss that came through the corridor twice daily at intervals I'd started tracking by instinct. None of those were new sounds.

"The field boundary," Nadia said. "What if it's not the anomaly expanding, but the boundary itself moving? The anomaly might be a fixed point, but its edge condition is changing."

She was working through something she hadn't finished thinking, which was rare for her. Usually she kept quiet until the thought was complete. The fact that she was saying it half-formed meant she wanted me to test it against what I knew.

"The predecessor's register plate changed behavior," I said. "Tracking something other than their produced state. Two practitioners before them documented the same. If the plate tracks the boundary of a field condition rather than the practitioner's output, then a boundary that's moving would explain why the plate's behavior changed at a duration threshold."

Nadia's pen was already moving. "That would also explain the disc corrections. If the disc is sensitive enough to track the field edge, and the field edge moved east at fifty seconds, the disc would correct toward east and then back when it re-established."

We were both extrapolating well past the data. She knew it, and I knew it, and neither of us stopped.

"I'm logging this as a hypothesis, not a conclusion," I said. "Labeled clearly."

"Same." She finished writing. "But Eli. If the boundary is moving, and the rate is increasing, and the anomaly is in your hall's east wall, then the Facilities report isn't just administrative anymore."

She didn't say what it was instead. She didn't need to.

Fass was in the common room two floors up, eating something that had started as a grain bowl before he'd added things to it until the category became unclear. When I entered, he pushed the bowl aside, which meant he'd already registered I had something to say.

"The Facilities channel," I said. "Do you know the actual receiving office? Not the routing designation. Who reads it."

He leaned back in his chair. "Depends what's in the report. Routine maintenance, that's Office Three, building services. Structural or spatial anomalies, they go to a subdivision under the chief administrator's office. I've only ever seen one report go there. It took four days to get an acknowledgment."

"Four days."

"Four days for acknowledgment. Not response. Acknowledgment." He watched me process that. "What's the rate now?"

He already knew about the anomaly. Fass had been the one to make the connection between the east corridor, my hall's east wall, and the floor discontinuity. He'd gotten there faster than I had, which was the kind of thing you noticed about a person.

"Four and a half centimeters in five days. Nadia measured today." I pulled up the session log on my interface and turned it toward him. The disc-correction entries, the fifty-five-second attention-shift. "Attempt fifty was today."

He read the entries without touching the interface. Fass had a habit of keeping his hands flat on whatever surface he was sitting near when he read something someone else had written, which I'd interpreted early on as a courtesy: your document, your space. "The disc corrections correspond to the anomaly's east direction."

"Yes."

"And the attention-shift at fifty-five seconds." He looked up from the interface. "The predecessor's plate changed behavior at roughly this duration range too."

He'd read RN-0047-NL's file. I'd shared the relevant sections with him and Nadia six days ago, before the secondary administrative report, and he'd tracked the session-duration data alongside his own observations. That was Fass.

"Their plate started tracking something other than their produced state around the same window," I confirmed. "My disc is still tracking correctly. But the corrections are new, and they're directional."

He was quiet for a stretch. Outside the common room's narrow window, the Annex's courtyard was visible, the paving stones running in a grid that had been laid at an angle to the building, which always made the courtyard look like it had been assembled from two different plans. I'd never thought about that before. My brain wanted to be somewhere else for a second, and it took what was available.

"You need to name the subdivision," Fass said. "In the report. 'Spatial anomaly in active practitioner hall' will get it routed correctly. If you write it as a maintenance issue, it goes to Office Three and sits."

"Do you know who heads the subdivision?"

"Administrator Cress. Or she was, when that report I saw went through. I don't know if she still is." He picked up his fork, set it down again without eating anything. "Eli. If this is what the predecessor's file suggests it might be, a Council review would come before a Facilities inspection. That's how it worked before."

He was telling me something without saying it directly, which was more caution than I usually saw from Fass. He was careful in how he spent words, but he wasn't usually circuitous.

"You're saying file the Facilities report," I said. "But know that the Council channel moves faster."

"I'm saying if you wait for Facilities to respond and the rate keeps accelerating, the decision about what happens next may not be yours to make."

That was clear enough. I closed the session log. "I'm not stopping the sessions."

"I know." He pulled the grain bowl back toward him. "I'm not saying stop. I'm saying document everything you have, today, with Administrator Cress's office named in the routing, and copy Hael as observer of record on the companion note. That way there's a formal acknowledgment trail that isn't just filed to a general channel."

Copying Hael was the part I'd been circling around. She'd told me she couldn't formally supply conclusions about the predecessor's status. But she was still the observer of record for my sessions, and a companion note wasn't a request for conclusions.

"You've done this before," I said. "Filed something that needed an acknowledgment trail."

Fass looked at his bowl. "My predecessor's file wasn't locked. It was just brief. Three weeks of sessions, then a transfer notation and a new name." He ate something. "The transfer notation was institutional courtesy. The person didn't request it."

That was more than he'd told me before. I sat with it, then stood.

"I'll file today. Before the next session."

He nodded, already reading something on his own interface. He did that, gave you what you needed, then went back to his own work, which was a way of treating the exchange as resolved without making a production of it. I walked back to the stairs and took them down two floors to my hall, where the disc was still at eight degrees, the cord straight, the tile flat, and the east wall held its anomaly the way it had yesterday and the day before, slightly wider now, its boundary condition shifting at a rate Nadia had underlined once in her field notes.

The report would name Administrator Cress's office. It would specify spatial anomaly in an active practitioner hall. It would copy Hael by name, as observer of record, with a routing note that was not a request for conclusions but a formal notice that the acknowledgment trail existed.

I opened the filing interface and started writing.