Im Soojin walked into Lee's Kitchen carrying a file box that was heavier than it looked and wearing an expression that said she'd already decided whether this meeting was worth her time but hadn't told anyone the verdict.
Mid-forties. Short hair. The kind of short that said "I cut it myself because hair appointments are time I could spend on research." She wore reading glasses pushed up onto her forehead and a sweater with a coffee stain on the left cuff that she either hadn't noticed or didn't care about. The file box went on the table before she sat down.
"Director Kwon told me you have evidence of pre-Awakening mana infrastructure in the Seoul geological substrate," she said. No greeting. No handshake. "I've been publishing papers about anomalous substrate signals for eight years. My last three submissions were rejected by the Association's research committee with the note 'insufficient empirical basis for the claimed phenomenon.' I have more empirical basis in this box than their committee has ever seen."
"Dohyun," Dohyun said. "Kang Dohyun."
"I know who you are. Kwon briefed me. B-rank Field Commander running an off-book infrastructure repair operation with a cell-structured organization. She also told me you're The Prophet from the hunter forums, which I don't care about. The forums are noise." She looked at the operational board. Her eyes moved across it the way a reader moves across a page: left to right, top to bottom, absorbing. "Your pressure readings show a decline from seventy-seven to fifty-five percent over six weeks. Your integrity metric is at sixty-one percent. Fourteen active repair batteries on randomized frequency protocols. Four cells: Containment, Repair, Intelligence, Security."
She read the entire board in ninety seconds.
"You've built a functional repair operation with roughly a dozen people and a budget that wouldn't fund my old lab's coffee supply. Either you're very good or very lucky." She sat down. Opened the file box. "I'm guessing both, because nobody gets this far on luck alone and nobody with only skill would be stupid enough to try."
Dohyun looked at Kwon's security clearance memo on his phone. Clean profile. No modification markers. Separation from the Association was political. Published papers on substrate dynamics that contradicted the official position. Pushed out.
"What was the official position you contradicted?" he said.
"That the mana substrate is passive. The Association's geological division treats the substrate as a medium. Mana flows through it the way water flows through pipes. The substrate doesn't do anything on its own. It's a conduit." She pulled a folder from the box. "I published findings showing the substrate has active properties. Signal generation. Pattern storage. Frequency-selective transmission. The substrate doesn't just carry mana. It processes it. The Association told me I was wrong, and when I provided additional evidence, they told me I was disruptive. Disruptive is the word they use when they can't say wrong anymore."
"What kind of signal generation?"
"Low-amplitude, sub-audible frequency patterns originating from deep geological formations. I detected them using modified seismographic equipment calibrated for mana-frequency ranges. The patterns are organized. Structured. They repeat on cycles that correlate with lunar tidal forces acting on the substrate's crystalline components." She opened the folder. Graphs. Waveform traces. Geological cross-sections annotated in handwriting that was precise and small. "I found these signals in twenty-three locations across the Seoul metropolitan area. Each location corresponds to a convergence point in the natural mana-conductive substrate. The signals at each location have the same frequency range but different phase patterns. Like twenty-three instruments playing different parts of the same piece."
Twenty-three locations. The ring circuit had seventeen primary channel junctions. The difference was six, which could represent the four keystones plus two junction points that Dohyun's team hadn't yet mapped.
"Show me the locations," Dohyun said.
She pulled a map from the folder. Seoul and greater Gyeonggi, with twenty-three points marked in red. Dohyun held it next to the operational board's infrastructure map.
Seventeen of the twenty-three points overlapped exactly with the ring circuit's known junction locations. The remaining six fell at positions that, based on Taehyuk's surface-channel mapping, could represent the root network junctions that connected to the mana-gathering system.
"You mapped the infrastructure from the surface," Dohyun said. "Without knowing it existed."
"I mapped anomalous substrate signals. I didn't know they were infrastructure. I didn't know the word infrastructure applied to geology. The Association's position was that the substrate was passive. I was measuring something active and nobody would engage with the data."
"When did you start this research?"
"Eleven years ago. Before the Awakening. I was a geophysicist studying tidal effects on deep-substrate crystalline formations. The mana-frequency signals appeared in my data set as artifacts that I couldn't explain with standard geophysics. After the Awakening, when mana became a recognized phenomenon, I reanalyzed the data using mana-spectrum instruments. The artifacts resolved into structured signals."
She'd been studying the infrastructure's output since before the System activated. The same signals that Taeyang's sensor network tracked. The same frequencies that the watcher generated from its crystal formation. The same patterns that the architects had built their weapon around.
"The Security cell needs someone who understands the Association's data systems from the inside," Dohyun said. "Every database, every monitoring protocol, every communication pathway. The gardener has already exploited the monitoring database as an intelligence pipeline. I need to know what other pathways exist that we haven't identified."
"Kwon told me about the database exploit. The Gwacheon server farm sits on the southern arc. Any data stored on that server leaks into the substrate at sub-threshold levels. It's the same effect I was measuring in my research." She took off her reading glasses. Cleaned them on the coffee-stained sweater. "The monitoring database is the obvious pipeline. There are at least three others. The Association's real-time gate telemetry system transmits data through a fiber network that crosses four substrate convergence points. The personnel registry server is co-located with the geological survey database on a server farm in Bundang, which sits above the eastern arc. And the financial transaction system, which processes hunter compensation payments, routes through a payment processor in Gangnam whose data center is thirty meters above the southern primary channel."
Three more pipelines. Three more pathways that the gardener could potentially use to extract data about the operation.
"I'll have a full audit of every Association data pathway that touches the substrate within a week," she said. "But that's the Security cell work. That's not why I asked Kwon to bring me in."
She reached into the file box. Pulled out a different folder. This one was older. The paper was yellowed at the edges. The label was handwritten in a style that predated the Association's standardized filing system.
"My research has a predecessor," Soojin said. "When I was compiling my initial findings eleven years ago, I ran a literature search for prior work on anomalous geological signals in the Korean substrate. The search returned zero results in peer-reviewed journals. But the Korean Geological Survey's rejected submissions archive had one entry."
She opened the folder. Inside was a printed copy of a paper. The formatting was decades old. Manual typeface. The kind of document that had been written on a word processor in the era before modern publishing software.
"A single paper, submitted to the Korean Journal of Geological Sciences thirty years ago. The paper described anomalous electromagnetic-like signals detected in the Seoul metropolitan substrate using modified seismographic equipment. The methodology was similar to mine. The findings were similar to mine. The paper was rejected with the note 'no known physical mechanism for the described phenomenon.' Which was true at the time. Mana wasn't a recognized scientific concept."
She handed the paper to Dohyun.
The title: "Anomalous Sub-Audible Frequency Patterns in the Seoul Metropolitan Geological Substrate: Evidence for an Active Signal-Processing Function in Deep Crystalline Formations."
The abstract described exactly what Soojin's research had found. Structured signals in the substrate. Active processing properties in the crystalline geology. Frequency patterns that suggested organized, non-random behavior in the deep rock.
The author: Choi Donghwan. Independent researcher, Seoul.
Dohyun read the name twice.
"Choi Donghwan submitted a paper about mana infrastructure signals thirty years before mana was a recognized phenomenon," Soojin said. "Using equipment that shouldn't have been able to detect mana-range frequencies, because the instruments weren't calibrated for a spectrum that didn't officially exist. Either he was extraordinarily lucky with his equipment configuration, or he knew what he was looking for before he started looking."
Dohyun turned to the author biography at the end of the paper. Two lines. "Choi Donghwan is an independent researcher based in Seoul. His research interests include sub-surface signal dynamics and geological information storage."
Geological information storage. The watcher's function, described in two words, in an author bio written thirty years ago.
There was a photo. Small. Black and white. The printing quality was poor, the image reproduced from what looked like a passport photo. A man in his fifties. Korean. Wire-frame glasses. The face was unremarkable except for the eyes. The eyes were the eyes of someone looking at a camera in a world that didn't understand what he knew, with the expression of a person who'd given up trying to explain.
"I tried to find him," Soojin said. "When I discovered the paper ten years ago, I searched every database available to me. The Association's personnel records. University staff listings. Research institution rosters. Immigration and residency records." She took the paper back. Looked at the photo. "Choi Donghwan doesn't exist in any current database. No Association record. No university affiliation. No residency registration in Seoul or anywhere else in Korea. The geological survey journal has no contact information beyond the mailing address he used for the submission, which traces to a postal box in Jongno-gu that was closed twenty-six years ago."
A man who knew about the infrastructure thirty years before the Awakening. Who published findings that matched both Soojin's research and the team's empirical discoveries. Who then disappeared from every record system in the country.
"He knew," Dohyun said.
Soojin looked at him. "Knew what?"
"He knew about the infrastructure because he'd seen it before. In another timeline. He regressed to a point thirty years before the Awakening and spent his time studying the substrate because he remembered what the substrate contained." Dohyun took the paper back. Looked at the photo again. Tired eyes. Fifty-something. "He's the person who wrote the log entries in the watcher's crystal."
Soojin's hands went still on the file box. The gesture of someone encountering a hypothesis that exceeded her current model. "Kwon didn't mention log entries."
"The Intelligence cell recovered a message from the northern keystone's geological formation. A personal log entry, written in System-encoding protocol, dated approximately thirty years before the Awakening. The entry describes future events in past tense. It references the ring circuit. It references previous failed repair attempts."
"Previous failed—" She stopped. Recalibrated. The academic mind working through the implications at a speed that would have impressed Minhee. "You're describing a regressor. Someone who experienced the future, returned to the past, and used foreknowledge to study the infrastructure before anyone else knew it existed."
"Yes."
"And you know this because..."
"Because I'm the same thing."
The kitchen was quiet. Soojin looked at Dohyun across the back table of a restaurant in Bucheon, the file box between them, the rejected paper from thirty years ago spread on the surface. She looked at the operational board. At the cell assignments. At the pressure readings that tracked a containment operation built on the foreknowledge of a twenty-year-old who'd lived to forty-two in a timeline that no longer existed.
"Kwon didn't mention that either," she said.
"Kwon doesn't know. The Intelligence cell leadership knows. Junseong knows. Now you know." He put the paper on the table. The photo of Choi Donghwan face-up. "The previous regressor used the infrastructure channels to store log entries in the watcher's crystal formation. We've recovered one entry. There are at least six more. The entries describe failed repair attempts across multiple regression cycles."
"Multiple cycles."
"At least four. Possibly more."
Soojin picked up the photo. Held it close to her face. Her reading glasses were still on top of her head. She squinted at the image, the printed face of a man she'd been looking for since she'd found his paper a decade ago.
"Choi Donghwan," she said. "I looked for you for ten years and you were a time traveler."
She put the photo down. Opened her laptop. Started typing.
"What are you doing?" Dohyun said.
"Building a search profile. If Donghwan existed in Seoul thirty years ago, he had a life. An address. Utility bills. Employment records. The paper was submitted from a postal box in Jongno-gu. That's a starting point. Even if his current records have been wiped or expired, historical archives might contain traces. Library cards. Hospital records. Tax filings from the era before digital systems." She typed without looking up. "A person doesn't just disappear. They leave footprints in paper systems that nobody digitizes because nobody looks."
"He might have died. Thirty years is a long time."
"Then there's a death certificate. A cremation record. A grave. Dead people leave more paperwork than living ones." She looked up from the laptop. "His log entries might tell us what he knew. But his life might tell us something more useful: what he did. What he built. Whether any of his infrastructure from the previous cycle survived into this one."
Whether the previous regressor had left something behind. Not messages in crystal. Physical assets. Equipment. Research facilities. Things he'd built during his years of operation that might still exist in the real world, buried in archives and forgotten storage units and the paper systems of a pre-digital era.
"The Security cell just got its first assignment," Dohyun said.
"The Security cell just got its first researcher," Soojin said. She closed the laptop. Put the photo of Choi Donghwan in the folder. Put the folder in the file box. "I'll need access to the national archives. The pre-digital geological survey records. And whatever historical address databases the Association maintains for hunter research."
"Kwon can authorize archive access."
"Good. And I'll need the full Intelligence cell briefing on the watcher, the log entries, and the infrastructure topology. Compartmentalization is fine for combat operations. For research, I need the complete data set."
"Minhee will brief you."
Soojin stood. Picked up the file box. Held it against her hip the way she'd held it when she arrived, the box of research that eight years of institutional rejection had failed to discredit.
"One more thing," she said. "The previous regressor. Choi Donghwan. He studied the same substrate signals I studied. He reached the same conclusions I reached. He submitted a paper that said the same things my papers said, thirty years earlier, and got the same response." She adjusted the box on her hip. "He was alone. No team. No cell structure. No institutional support. And he couldn't even get a paper published."
She looked at the operational board. At the names. At the cells.
"I was alone too. For eight years. An ex-Association researcher with findings nobody would read, studying signals nobody believed in, looking for a man who'd vanished from every database in the country." Her jaw set. "I'm not alone anymore. Don't waste that."
She left. The door closed. The file box's weight had left an impression in the wood grain of the table, a rectangular mark where the sharp corners had pressed.
Dohyun sat with the rejected paper. The photo of Choi Donghwan. Tired eyes behind wire frames, looking out from a passport photo taken thirty years ago by a man who'd already lived and died and come back and tried and failed and left breadcrumbs for whoever followed.
The previous regressor had a name. A face. A paper trail that Soojin could follow through thirty years of archives.
Somewhere in those archives was the answer to what Choi Donghwan had built, what he'd tried, and how the gardener had beaten him.
Dohyun put the photo in his jacket pocket, next to the decoded log entry, next to his mother's texts, next to the phone that carried the operational picture of a war he was winning.
Two soldiers' documents in the same pocket. One alive. One probably dead. Both carrying memories of futures that had ended and trying, with whatever tools they had, to make the next attempt different.
Soojin's footsteps faded down the alley. The kitchen was quiet.
Choi Donghwan. Independent researcher. Seoul. The man who came before.
What did you build that we haven't found yet?
