Noise Floor

Dr. Yuki Tanaka arrived at the control room at 7:58 AM, two minutes before her shift officially began, carrying the specific expression she reserved for mornings when the overnight log queue was longer than usual. She found Rowan exactly where she had left him the previous evening, which told her everything she needed to know before he said a word.

"You didn't sleep."

"I slept some."

"Your definition of *some* has been medically concerning for fourteen months, Row." She set her bag down and looked at his screen, at the waveform still open from hours ago, at the log entry timestamped 04:53. She read it. She read it again. Then she pulled a chair over without being asked, which was, in the specific grammar of their working relationship, the equivalent of rolling up her sleeves.

"Walk me through it," she said. "Slowly. Assume I'm stupid."

"You're not stupid."

"Assume it anyway. It's a useful exercise."

He walked her through it. The anomaly buried in the 23 GHz band, flagged below threshold. The decoder — their decoder, the one they'd built together in the years before Yuki had joined the project, back when it had been Rowan and Lena arguing about consciousness signatures over a whiteboard that still, technically, had some of Lena's handwriting on it under three subsequent layers of dry-erase. The match score. The triangulation across three additional dishes. The two hours and five minutes of transmission before the signal cut off mid-pattern.

Yuki listened without interrupting, which was itself unusual for her, and when he finished she sat for a long moment with her hands folded on the desk, looking at the number on the screen the way a person looks at a diagnosis they have not yet decided whether to believe.

"Okay," she said finally. "Let's break it."

This was the thing Rowan had always respected about her, the reason he'd fought to bring her onto the project four years ago over Miriam's initial hesitation — Yuki did not react to extraordinary claims with awe or dismissal. She reacted to them with the specific, patient hostility of someone whose entire professional identity was built on the conviction that most extraordinary claims turned out to be ordinary claims wearing a costume. She wanted to find the seam. She always wanted to find the seam.

"Calibration drift," she said, pulling up a terminal. "First and easiest. If the decoder's baseline comparison has drifted, we'd get inflated match scores across everything, not just Lena's archive."

"I checked that. Ran it against the composite and against my own archive first."

"Run it again. While I watch."

He ran it again. His own archive: 1.4%, statistically identical to the previous run, well within the noise band a healthy person's data should produce against a pattern that wasn't theirs. The composite: 0.8%. Lena's archive: 99.997%, the number arriving with the same flat, indifferent precision as it had eleven hours earlier, as if the intervening time and the growing horror of two more people looking at it made no difference to the mathematics at all.

Yuki's jaw tightened almost imperceptibly. "Do it against Fenwick's."

Rowan didn't have Fenwick Abara's archive locally, but Yuki did, on her authorized credentials, and she pulled it into the comparison herself rather than ask him to. 2.1%. She ran her own. 0.6%. She ran three more archives from the original forty-nine-participant study, pulled from the aggregate research set, methodically, the way a person checks all four burners on a stove before leaving the house.

Every archive but Lena's produced numbers between zero and three percent. Lena's produced 99.997%, consistently, on every pass, against every attempt to explain it away as an artifact of the tool itself.

"Signal spoofing," Yuki said next, not because she believed it but because the list demanded completeness. "Someone with resources fakes the origin coordinates, fakes a neural-pattern match, sends it as — I don't know. A message. A threat. A test."

"I already ran that literature."

"Run it with me."

They ran it together. The math on interstellar signal spoofing was not, strictly, impossible — nothing about physics forbade a sufficiently advanced and sufficiently motivated actor from constructing a false transmission that appeared, from Earth's receiving end, to originate from a specific point in the sky. But the resource requirements scaled with distance in ways that made the numbers, at 13.8 light-years, essentially theological. You would need transmission infrastructure either already present at Wolf 1061c — which no human mission had ever placed there — or so powerful from a closer vantage that its own signature would have been unmissable on every survey instrument pointed anywhere near that patch of sky for the last several decades. Neither existed in any confirmed record. Yuki checked the survey archives herself, cross-referencing three independent observatories, and found nothing.

"Nobody sent this from somewhere else and made it look like it came from Wolf," she said, mostly to herself. "There's no infrastructure. There's no plausible actor. There's no motive that isn't insane."

"So."

"So it's not spoofing." She said it like an admission of defeat, which, in a sense, it was — Yuki did not like crossing possibilities off a list unless she was certain, and she had just crossed off the possibility she'd found most comforting. "Software artifact from your side, then. Decoder malfunction, presenting as a real match because the code itself is broken in a way we haven't found yet."

They spent the next hour on this one, because it was the most technically thorough option and because Yuki, Rowan suspected, wanted very badly for it to be true. She rewrote the comparison algorithm from scratch in a sandboxed environment, isolated from the original codebase entirely, using nothing but the raw waveform data and a clean reimplementation of the matching logic that she built line by line while Rowan watched. It took her fifty minutes. When she ran the reconstructed signal against Lena's archive using code that had never touched the original decoder, that owed nothing to eleven years of accumulated assumptions and possible bugs, the match came back at 99.994%.

Not identical. Close enough that the difference was almost certainly rounding artifacts between two independently written implementations doing the same fundamentally sound mathematics.

Yuki stared at her own code for a long time.

"I want it to be broken," she said quietly. "I want there to be a bug. I've been trying to find one for fifty minutes and I can't, and I need you to understand that I am not saying this because I believe—" She stopped. Started again. "I am saying this because I have exhausted the debunking options I know how to run, and I don't like where that leaves us."

"Where does it leave us?"

"It leaves us with a real signal, from a real location, carrying a pattern that is, by every rigorous test I know how to apply, indistinguishable from Lena." She rubbed her eyes with the heels of her hands. "Which is impossible. Which I want to say very clearly is impossible, Row, because no mission has ever been sent there, because there is no mechanism by which a person's neural pattern gets transmitted across space, because she has been dead for fourteen months and dead people do not transmit anything, and I need you to hear me say all of that and also hear me say that I cannot find the flaw."

Rowan had known, somewhere beneath the part of him that had spent the night hoping otherwise, that this was where it would end. He had wanted Yuki to break it. He had wanted, more than almost anything in fourteen months, for someone with a colder head and no personal stake to sit down with his data and dismantle it methodically, to hand him back a clean, sad, ordinary explanation he could file away with all the other reasons grief made people see things that weren't there.

She had not been able to do it. That fact sat between them in the control room with a weight that neither of them immediately knew how to hold.

"There's something else," Yuki said, after a silence that had gone on long enough to become its own kind of statement. She pulled up a secondary display, one Rowan hadn't been looking at because he'd been entirely focused on the match scores. "I've had a background process running a continuous scan of that coordinate range since you flagged this at five AM. In case it started again."

"And?"

"It started again. Forty minutes ago. It's still going."

Rowan looked at the display. The signal was live, updating in real time, the waveform scrolling across the screen in the same structure as the recording he'd analyzed all night, except that Yuki's overnight comparison — automated, continuous, running the same match algorithm every thirty seconds against the incoming stream — showed something the earlier transmission hadn't.

The match score was climbing.

99.997% had been the ceiling on the first transmission, the number that had held steady across every replay and every independent verification. The live signal, forty minutes into its second appearance, was already at 99.998%, and as they watched, in the specific silence of two people who understand exactly what they are looking at and do not yet have language for it, it ticked upward again.

99.999%.

"It's getting closer," Yuki said, very quietly.

"Closer to what?"

Neither of them answered that, because neither of them had an answer that didn't sound, out loud, in the flat fluorescent light of a control room at eight in the morning, like the kind of thing a rational scientist was not supposed to be willing to say. But Rowan was thinking it anyway, and from the way Yuki's hand had gone still on the desk, fingers no longer moving across the keyboard, he suspected she was thinking it too.

The signal held at 99.999% for six full minutes. Then, without warning, without the gradual attenuation either of them expected from a source thirteen point eight light-years away, it stopped.

The silence that followed felt different from the silence before, though Rowan could not have explained, in any language his training gave him, exactly why.