Slit and Membrane

The algorithm finished at 03:14 ship-time, which meant nothing to anyone but Lian, who was the only person awake in the boundary-layer physics lab and had been for six hours. She had expected noise. A messy gradient, the usual turbulence maps she’d been generating for two years — the slow-light field’s internal weather rendered as color, as topology, the familiar churn of a medium that was neither vacuum nor matter but something older than both. She’d written the new model to resolve finer structure in the shear zones, the fast-moving streams where the effective speed of light climbed toward something closer to what the old texts called c. She expected detail. She got geometry.

The streams were not random.

She sat with the visualization for ninety seconds, her hands flat on the bench, and the room changed around the image the way a room changes around a body on the floor. The streams formed a lattice. Not approximate, not suggestive — a grid with hexagonal symmetry, spacing consistent to four decimal places across the entire survey volume. Where each line of the lattice crossed another, the field value dropped to a local minimum: a node, a thinning, a place where the membrane between their slow-light domain and the outer vacuum was not thick but threadbare.

Lian zoomed in on one node. The field gradient was not smooth. It had a discontinuity — a phase boundary, sharp as a crack in glass. On one side, the vacuum had the properties they had always known: light crawling at a fraction of its natural speed, signals dying over short distances, the universe reduced to a sphere eleven light-minutes across. On the other side, extrapolated from the gradient’s slope, the vacuum was normal. Lightspeed vacuum. The real universe, with its real distances and its real silence.

She pulled up the refractive index profile. The slow-light domain worked like a metamaterial — not a substance but a structured vacuum, the quantum field arranged in a configuration that bent and slowed electromagnetic propagation the way glass bends visible light, but at cosmological scale. This was known. This was textbook. What was not textbook was the lattice. The refractive index changed not gradually across the boundary but in a step function — a discrete jump from one vacuum state to another. Phase transitions did that. Phase boundaries did that. Engineered phase boundaries did that.

She ran the model again. Same lattice. Same nodes. Same phase boundaries with the same characteristic sharpness that only one kind of physics produced: not erosion, not decay, but construction. These were joints. Seams. The membrane had been assembled from panels, and the seams were where the panels met.

Lian closed the visualization and sat in the dark lab and listened to the ship breathe. The Anwei’s ventilation had a rhythm that every child learned before they learned to walk — a six-second cycle, inhalation and exhalation, the lungs of a world that was not a world. She had breathed with it her whole life. Forty-one years inside a membrane that someone, or something, had built.

Built things have builders. Built things have purposes. And the purpose of a cage is never comfort, regardless of what you name it.

She locked the lab, pocketed the data chip, and walked home through the ship’s intestinal corridors, past the water recycling stations with their permanent smell of ozone, past the darkened windows of the school where she had learned to read and Sovei was currently learning to argue, past the mural on Deck 7 that someone had painted four generations ago — a field of stars, imagined, because no one aboard the Anwei had ever seen a star.

The commons at 03:40 was not empty. The Anwei had four thousand people and the informal rule was that the commons belonged to whoever was awake, which meant the insomniacs and the third-shift recycler workers and the teenagers who had discovered that the adults’ communal spaces felt different at night — less surveilled, the architecture softening into permission. Lian passed a group of them near the hydroponic gallery, five or six kids sitting on the floor eating stolen persimmons, and recognized Sovei’s jacket before she recognized Sovei.

Her daughter did not look up. This was not unusual. Sovei was sixteen and had been practicing a studied blindness toward her mother for the better part of a year, a performance that was itself a political act — or a personal one, or some combination that Lian had stopped trying to parse. She walked past without stopping. The persimmons were probably from Bay 7, which meant someone had picked them before the rotation was scheduled, which meant a conversation at the next food assembly about whether fruit belonged to whoever grew it or whoever was hungry, and Lian could not hold the scale of that argument in the same mind that held the lattice, so she let it go.

Home was a three-room unit on Deck 9, identical to every other three-room unit, which was itself a political statement — the Anwei’s founding charter specified uniform living spaces, no larger, no smaller, the architecture of equality made literal in square meters. Deris was asleep, turned toward the wall, one arm trailing off the bed’s edge. Deris slept like someone who’d been set down carelessly — limbs at odd angles, face pressed into the pillow with an abandon that struck Lian, every time, as an act of trust in the room itself, in the air and the gravity and the walls that held the air in place. She stood in the doorway and did not think about membranes. Then she went to the kitchen table and sat down and thought about membranes.

Phase boundaries. The slow-light field was not uniform — she had known that for years, everyone in her four-person research group knew it, it was the foundational observation of boundary-layer physics aboard the Anwei. The streams were well-documented: rivers of faster light moving through the general medium, shear zones where the field’s dielectric properties shifted and the effective refractive index dropped. She had published three papers on stream dynamics. She had taught a seminar about them every second year. She had used the scissurellid snails in the thermal vents as a teaching metaphor — tiny organisms, smaller than her thumbnail, whose survival depended on maintaining a permanent slit in their shells. The slit let water through, let the organism breathe. A sealed shell is a dead shell, she had told her students. Then she’d corrected herself: A sealed shell is a coffin. The slit is what makes it a house.

But the streams were supposed to be natural. Turbulence in the field, like weather patterns in an atmosphere. What she had found tonight was not weather. It was infrastructure.

She pressed her forehead against the table’s cool surface and listened to the ship breathe and the ship breathed and she breathed with it because breathing was the only decision she could make that did not require choosing a side.

In the morning — 07:00 by the clocks, though the Anwei had no sun and no horizon and morning was a consensus rather than an event — Lian went to the communal kitchen on Deck 9 and ate porridge with dried apricots and said nothing. The kitchen served eighty people in a room built for sixty, which was another political fact: the population had grown but the kitchens had not, and the resulting closeness was either community or crowding depending on your politics. Deris was at the long counter chopping onions for the midday stew, working with the efficient, slightly distracted rhythm of someone whose body knows the task well enough that the mind can wander. Deris taught ecological management — the biomes, the thermal vents, the snail colonies, the fungal networks that decomposed waste and returned it as soil. Deris understood closed systems the way a conductor understands an orchestra: not the individual instruments but the way they breathed together.

“You came in late,” Deris said, not looking up from the onions.

“New model running. Wanted to see the results.”

“And?”

“Interesting. I need to verify it.”

Deris set down the knife and turned to face her. Deris had a way of looking that Lian had once described, in a rare moment of self-awareness, as being weighed by someone who knows your exact density. Seventeen years together and the look still worked. It was not accusation. It was attention, which was harder to deflect.

“Interesting meaning you’ll talk about it over dinner,” Deris said, “or interesting meaning you’ll vanish into the lab for a week and come back looking like you’ve been arguing with the hull?”

“I don’t argue with the hull.”

“You used to argue with Fen.”

The name landed the way it always landed — not like a blow but like a change in air pressure, the room suddenly denser. Fen Audo had been dead for fourteen years. The cooling system patch that killed her and eleven others had been redesigned, rebuilt, the section of Deck 4 renamed, and the assembly had passed a resolution about transparency in structural reporting that everyone quoted and no one enforced. Lian did not talk about Fen. She did not need to. Every decision she had made since — every piece of data held back, every finding she sat with for an extra week before sharing, every conclusion she softened or withheld or framed — was a conversation with Fen’s ghost, conducted in the language of what she should have done and what the assembly should have been and what happens when you give a crowd a fact they cannot hold.

“I need to verify it,” she said again. “That’s all.”

Deris picked up the knife. The onions sizzled when they hit the pan. The kitchen smelled of cumin and rendered oil and the particular warmth of twenty people eating in the same room, breathing the same air, pretending that the morning was ordinary because ordinary was what they had, and what they had was enough, and it had to be enough because there was nothing else.

She ran the model four more times over the next three days. Each run confirmed the lattice. On the third run, she added a temporal dimension — comparing the current field structure against archival survey data from eight, fifteen, and twenty-two years ago. The lattice had been there in every dataset. It had always been there. The streams she had spent her career studying were not rivers in a landscape. They were gutters in a roof. Drainage channels built into the membrane’s architecture.

What had changed was the seams themselves.

They were widening.

The rate was small. A fraction of a percent per decade, extrapolating from twenty-two years of data. At that rate, the phase boundaries would remain structurally stable for another sixty to ninety years before the first node degraded enough for light to leak through. But the trend was unambiguous, present in every dataset, monotonically increasing. The membrane was thinning at its joints the way ice thins at the cracks when the water beneath it warms.

Lian understood ice. Her first research specialty, before boundary layers, had been the analogous structures in the slow-light field — regions where shear forces thinned the medium, where the field flowed fastest, where the refractive gradient was steepest. She had borrowed the terminology from Earth-era glaciology: ice streams, rivers of solid material moving under their own weight, lubricated by hidden water at the boundary between ice and bedrock. In Antarctic ice sheets, ice streams had accounted for ninety percent of mass loss. Solid matter that moved like liquid because of what lay hidden at the interface. The streams were where the sheet bled.

The membrane was bleeding too. And the streams — her streams, the structures she had built her career on — were the channels through which it bled.

She sat with this for one full day before she opened the ship’s historical archive.

The archive was a physical room on Deck 3, though the data could be accessed from any terminal. Lian went to the room. She wanted walls around her while she read this.

She searched for a phrase she had encountered once in a first-year physics seminar, a piece of pre-departure cosmological theory that the Anwei’s founders had included in the knowledge base the way you include a warning label on a container of something that could kill you: the dark forest.

The dark forest hypothesis had arrived on the Anwei as part of twelve thousand years of human thought, compressed and curated and filed by discipline. It was listed under cosmology, subsection astro-sociology — a field that had not survived the departure because the Anwei had no stars to be social with. Lian had read the abstract once, years ago, the way you read about a disease that afflicts people on a different continent. Now she read the primary sources.

The logic was not complex. It did not need to be. The simplest arguments are the hardest to escape.

The universe is large. Civilizations exist but cannot verify each other’s intentions. Communication takes years, centuries, millennia. Trust is impossible across relativistic distance because the message that says I am peaceful and the weapon that says you are not arrive at the same speed. The only rational strategy is to destroy any civilization you detect before it detects you. The cost of being wrong about a threat is extinction. The cost of being wrong about a friend is guilt, and guilt is cheaper than nonexistence. Therefore every civilization hides. And the ones that failed to hide are dead. And the silence of the cosmos is not emptiness but evidence — the sound of everyone who survived, holding their breath.

Lian read the mathematical formulations. Game theory applied to lightyears. Evolutionary dynamics extended across species boundaries. The chain of suspicion: even if you are peaceful, you cannot prove it across relativistic distance; even if they are peaceful, they might not remain so across evolutionary time; even if both of you are peaceful now, a third civilization might not be, and it takes only one defector to validate the entire model. She read the sociological appendices — how the theory had shaped Earth-era policy, how it was invoked to justify militarization and surveillance and the building of barriers, how dissenters had attempted to counter it with broadcasts of goodwill that were, by the theory’s own logic, either suicidal or irrelevant. How the theory had become, over centuries, not a hypothesis but an architecture. How you could trace the shape of the Anwei’s membrane in the equations.

She read for six hours. When she stopped, the archive room was dark except for the terminal’s glow and her neck ached and she understood something she had not understood before: the Anwei’s membrane was not a shield. It was not a wall. It was a strategic position. The founders had not hidden their people from the universe out of caution or cowardice. They had encoded a cosmological argument into the vacuum structure of space itself and wrapped it around four thousand human lives and called it home.

The membrane was an argument. It said: We are not here. We have never been here. If you find us, we were never here. And for eleven generations the argument had held because the physics had held.

But arguments wear out. And the seams were where the argument was thinnest.

On the fourth night, Lian did not go home. She slept in the lab on a cot that someone had left there a generation ago — evidence that the lab’s gravitational pull on its occupants was not unique to her. She woke at 02:00, made coffee from the illicit kettle that every lab on the Anwei maintained in cheerful defiance of the energy-rationing guidelines, and ran the model she had been avoiding.

She wanted to know what would happen if a seam opened.

The mathematics were straightforward. A phase boundary collapse at any lattice node would create a channel — a cylindrical region, perhaps two hundred meters across, where the slow-light field would normalize to background vacuum. Light would travel at c. Electromagnetic radiation — thermal, radio, the entire spectrum of a living ship — would propagate outward at full speed. For the first time in eleven generations, the Anwei would emit.

Not a message. Not a signal. Just thermal noise — the waste heat of four thousand bodies and their machines and their cooking and their arguments, leaking through a crack in the wall. At interstellar distances, it would be faint. A candle in a dark forest.

But candles are visible. That is the entire problem with candles.

Lian stared at the numbers and thought: Faint is not invisible. Nearly undetectable is not undetectable. And in the dark forest, the distance between nearly and completely is the distance between existing and not.

She ran a second model. A controlled opening. If you could direct the phase transition, collapsing a seam along one vector while reinforcing the surrounding nodes, you could create a directional channel. Not an omnidirectional beacon but a narrow aperture — a telescope. Receive without transmitting. Look without being seen. One-way glass.

The mathematics were possible. The engineering was uncertain. The computational limits of the ship’s systems meant she could model the general case but not the boundary conditions with enough precision to guarantee directionality. To know whether controlled opening was achievable, she would need to run experiments on the membrane itself. On the boundary that kept them alive and hidden and ignorant and safe.

She locked the lab at 04:00 and went looking for more coffee. What she found was Sovei.

Her daughter was sitting on the floor outside the physics wing with two other teenagers Lian did not recognize, reading something on a shared tablet. They were so absorbed they did not hear her approach. Lian stood in the corridor for a moment, watching — the three of them leaning in toward the screen with the concentrated attention of people reading something they believe will change them. She looked over Sovei’s shoulder — close enough to smell the stolen persimmon on her daughter’s breath — and read the header: The Opening Argument: Why Anarres Must Speak.

Anarres. The ship’s youth had taken to calling the Anwei by the name of a fictional planet from the pre-departure literature — an anarchist world behind a wall, a society that had chosen isolation and been shaped by the choice until the choice and the shape were the same thing. The analogy was imperfect — the Anwei had no planet to be exiled from, no Urras to measure itself against, only the membrane and the dark — which the teenagers argued about constantly, which was part of the point. They called themselves the Openists. They wanted to speak. They wanted to drop the wall and send a signal into the dark and wait for an answer, because they were sixteen and had never watched a cooling system kill twelve people and had the terrifying clarity that comes from not yet having a reason to be afraid.

Sovei looked up. Their eyes met. Lian saw her daughter’s face — young, hard with conviction, utterly certain in the way that only someone who has not yet been wrong about anything that matters can be certain — and felt a vertigo that had nothing to do with the ship’s rotation.

She knew the wall had seams. She knew the seams were opening. And her daughter was sitting on the floor outside her lab, three meters from the data that would prove everything the Openists believed, and Lian’s silence was a wall between them that was no less real for being invisible.

“Mom,” Sovei said. The word was flat. An acknowledgment, not an invitation.

“Late night,” Lian said, and walked past, and did not stop, and did not speak.

Back in the lab, she sat in the dark for an hour. Not working. Not thinking, exactly. Processing, the way the ship’s water recyclers processed — taking in what was contaminated and returning something cleaner, though never clean.

The dramatic irony was not lost on her. Her daughter was reading a manifesto about opening the wall. Her mother had the data proving the wall could be opened, was opening, would open whether anyone manifested about it or not. Between the two of them stood Lian’s silence, which kept the knowledge on one side and the people who needed it on the other, and Lian stood at the boundary pretending she was a door when she was, in fact, a lock.

She wrote a letter to Deris on the terminal. She did not send it. She knew what it said — that she had found something that changed everything, and she was holding it alone because she believed she was best equipped to understand what it meant, and she knew this was the same conviction Fen had held about the cooling data, and she was doing it anyway, because the alternative — trusting the assembly, trusting that four thousand people could hold a truth larger than any of them — felt like handing a child a loaded tool and hoping they build something instead of breaking something. She knew that metaphor made her the person who thinks she’s the only adult in the room. She knew that person is always wrong. She was doing it anyway.

She deleted the letter. She went back to her models. The lattice stared back with its hexagonal patience, its 1,447 nodes, its seams that were widening whether she watched them or not. She ran the numbers again. She would always run the numbers again. The numbers were the part of the problem she could control, and control was the drug she had been taking since she was nineteen years old and watched Fen Audo learn that sharing the truth and being heard were two different things.

On the seventh day, Deris came to the lab. They brought soup. This was an act of care that also functioned as a siege weapon — after seventeen years, Lian recognized the dual purpose of Deris’s cooking, the way food became an argument you could not refuse without refusing to eat.

“You haven’t been home in three days.”

“I know.”

“You’re holding something.” Deris set the container on the bench between two stacks of printouts. “I’m not asking what. I’m telling you what I see. I see my partner disappearing into this room. I see her sleeping on that cot. I see her looking at our daughter like Sovei is a variable in an equation rather than a person who misses her.”

“Sovei doesn’t miss me. Sovei is angry at me for reasons she hasn’t articulated and I haven’t asked about because I’ve been —”

“Hiding.”

The word sat between them. Lian opened her mouth and closed it.

“She’s angry because you’re absent,” Deris continued. “She’s angry because the Openists talk about community and participation and transparency, and you — her mother, the boundary-layer physicist — won’t even attend their meetings. She thinks your silence is a judgment.”

“My silence is —” Lian stopped. Started again. “Some knowledge is dangerous, Deris. Some findings, when you release them into a community that hasn’t built the structures to hold them, don’t illuminate. They detonate. Fen brought her cooling data to the assembly. The assembly panicked. Three factions in a night. She was pressured to hedge, to soften, and the compromise was so weak that the patch failed and twelve people burned. I was nineteen. I watched it happen. I watched Fen’s face when she realized the assembly wasn’t going to do what the data required.”

“I know the story, Lian. Everyone knows the story.”

“Then you know why I’m careful.”

“I know why you think you’re careful. I think what you call careful is something else.” Deris picked up a printout from the bench, glanced at it — a graph of field gradients that meant nothing without context — and set it back down. “In our structure — in this structure that we have spent eleven generations trying to build and failing to build and rebuilding — no one has the right to hold knowledge alone. That’s the rule. That’s the thing we agreed, or that was agreed for us, or that we’ve spent three hundred years trying to agree on. The food isn’t yours. The air isn’t yours. The data isn’t yours.”

“Some knowledge is a weapon, Deris. And you don’t hand a weapon to a crowd.”

“You sound like the people who built the wall.”

The sentence landed and Lian could not tell whether it had arrived at the wrong moment or the right one, and the not-telling was itself a kind of answer. She looked at Deris. Deris looked back. Neither of them flinched, which was not a contest but a recognition: they had arrived at the place in the argument where both of them were right and the two rights were incompatible and the incompatibility was not a problem to be solved but a condition to be lived in, like gravity, like the slow-light field, like love.

“Maybe I do,” Lian said. “Sound like them.”

Deris left. The soup cooled on the bench. Lian ate it an hour later, standing at the window that overlooked the thermal bays, watching the dark channels where the scissurellid snails lived — their slit shells facing currents they had never chosen, water flowing through the openings in their armor because a sealed shell was a coffin and an open shell was a risk and a slit shell was the negotiation between the two that allowed you to keep breathing.

She slept in the lab again. She dreamed about Fen Audo standing at the front of the assembly, holding a number that the room was not ready to carry, and woke up knowing that the dream was about herself.

She found it on the ninth night. The data that collapsed the timeline from decades to days.

She had been mapping the widening rate at sixty different nodes, looking for spatial patterns. Most nodes were degrading slowly, at the gentle rate she had already documented. But the forward cluster — a group of twelve nodes in the ship’s direction of travel — was different. The widening there was not just faster. It was accelerating. And the acceleration was not smooth.

It came in steps. Discrete jumps every 73.2 hours, as regular as a heartbeat. As regular as a probe cycle.

Between each step, the rate held constant — the membrane at rest, absorbing the disturbance, reaching a new equilibrium. Then another step. Another increase. The pattern was unmistakable: something was applying force at the forward seam, measuring the membrane’s elastic response, waiting exactly 73.2 hours, and applying force again. Calibrated. Incremental. Patient.

The membrane was not just degrading. It was being tested.

Lian ran the numbers three times. Checked the sensor calibration twice. Cross-referenced with navigational data to rule out internal vibrations, thermal fluctuations, computational artifacts. The signal was clean. Something outside the membrane was probing the lattice with the systematic precision of a researcher mapping an unknown material — or a locksmith testing a lock — or a surgeon locating the thinnest tissue — or something for which she had no analogy because the dark forest did not offer analogies. It offered outcomes.

At the current rate of acceleration, the forward seam would reach critical phase-boundary threshold in eleven days. The vacuum state at the forward nodes would undergo spontaneous symmetry breaking. The slow-light phase would collapse to normal vacuum. A channel two hundred meters wide would open in the membrane.

Eleven days until the candle lit itself.

She sat in the lab and held the number. Eleven was a small number. It fit in your hand. It weighed nothing. It meant everything. She thought about Fen Audo, who had held a number like this and brought it to the assembly and the assembly had flinched and the flinching had killed twelve people. She thought about Fen’s face in the dream and wondered whether Fen had sat alone with her number too, and for how many nights, and whether the sitting alone had been the thing that killed the twelve or whether the telling had.

She thought about Deris, who was right about knowledge and wrong about danger and right about the wall and wrong about what the wall was for.

She thought about Sovei, who wanted to open a wall she did not know was already opening.

Three options. She could present the data to the assembly. She could act alone — redirect the thermal systems to reinforce the forward seam, kill the ecology to save the wall. Or she could do nothing and let the probe succeed.

Each option was a different shape of violence. Against the community’s autonomy, against the community’s ecology, against the community’s existence. There was no fourth option. The universe did not offer options that preserved everything — it offered trade-offs, and the only question was who got to make the trade. And she sat in the lab with eleven days pressing against her like a headache that was also a countdown and thought: I am the people who built the wall. I have been standing on the inside of it my entire life, and the only thing I have ever done with the view is refuse to share it.

She walked to the communications panel. She called an assembly.

The assembly met in the commons on Deck 5 — dining hall and gymnasium and theater and clinic, depending on the hour and the need. Lian had called it, which anyone could do, which was the rule, which meant that by 20:00 ship-time there were eleven hundred people on the floor and along the walls and more watching from screens in their units, and the room smelled of dinner and carried the particular electric tension of a community that knows something has broken but not what.

Lian stood at the front. She had the visualizations loaded on the main display. She had not prepared a speech or built a narrative or constructed an argument or decided what she thought they should do. She had data and she had options and she was finished deciding for other people what they could carry.

She showed them the lattice. She showed them the seams.

“The membrane around this ship is not a natural phenomenon,” she said. “It was constructed. The slow-light field has internal structure — a hexagonal lattice of phase boundaries, engineered joints between vacuum-state panels. It was built.” She let the word sit. “I don’t know by whom. I don’t know when. But it was built, and built things degrade.”

She showed them the widening data. Twenty-two years of archival evidence.

“The degradation is accelerating. In the forward quadrant, the rate has increased by a factor of six in nine days. The pattern is periodic — 73.2-hour cycles. Something outside the membrane is probing the seam structure.”

The room shifted. Not a sound but a change in the quality of the silence — a held breath becoming a different kind of held breath.

She showed them the dark forest hypothesis. Not her interpretation — the original sources, the mathematics, the game theory. She watched eleven hundred faces encounter the idea that the universe was a place where being seen was the same as being targeted, and she watched the idea land differently on each face, because fear is personal even when the threat is cosmological.

“If the forward seam breaches, light escapes. We become detectable. In the context of the dark forest, that means we become a target. Not because we’re dangerous. Because we exist.”

She showed them the options. Reinforcement: redirect thermal energy to the forward seam, restabilize the boundary, cost the ship’s ecology years of recovery. Controlled opening: attempt a directional channel, receive without transmitting, theoretically possible but unverified. Inaction: let the probing continue, accept the outcome.

“I have eleven days of data and forty-one years of training and I do not know which of these is right,” she said. “I brought you the data because it belongs to you. Because Fen Audo brought her data and this assembly failed her and I spent twenty-two years believing that the failure meant you didn’t deserve the truth, and I was wrong. The assembly failed because it was ambushed. I’m trying not to ambush you. I’m trying to give you time to be what the charter says we are.”

She sat down.

The room held for three seconds. Then it broke open the way rooms break open when the pressure has been building for eleven generations and someone finally names the shape of the container — not with a single crack but with fissures running in every direction at once, everyone speaking to a different person about a different fear.

The survivalists spoke first, because fear moves faster than thought: reinforce the membrane, redirect the thermal systems, accept the ecological damage, hide. Ged Tomás, who had run structural engineering for twenty years, stood and said every generation ship ever theorized worked on the same principle — protect the hull first, grieve the garden later. A woman from the ecological team shouted back that the garden was the hull, that without the fungal networks there was no air cycling, and without air the membrane was academic. Deris stood — Lian had not seen Deris until that moment — and said, “The thermal redirect buys us decades. The ecological collapse costs us the same decades. You’re spending the currency you’re trying to save.”

The Openists spoke next. Pava Osei argued that the probing was evidence of intelligence, that intelligence implied the possibility of communication, that the dark forest was a theory and theories could be wrong. A physicist from Lian’s own group — Rais, who had not known about the lattice until tonight — countered that the theory did not need to be wrong for the probing to be non-hostile, but the theory did not need to be right for the probing to be lethal.

A third faction assembled without a name — engineers, ecologists, people who could not bear either extreme. They argued for the controlled opening. Listen first. Crack the door an inch. Put your ear against it.

The debate went for two hours. It became circular, then personal, then ugly in the way that communal arguments become ugly — not through cruelty but through the exhaustion of discovering that the people you eat with and sleep near and share air with have completely different ideas about what survival means and what it costs. Someone accused Lian of withholding the data for weeks, which was true. Someone else defended her, which was kind but changed nothing. A man proposed a vote and was shouted down by people who said you don’t vote on survival, and others who said you don’t vote on anything, consensus or nothing, and others who said consensus takes time and eleven days was not enough time for consensus and maybe they should talk about that, and someone else said if they talked about talking they would die talking about talking.

Sovei stood up.

The room was loud. Sovei was not loud. But sixteen-year-olds broadcast on a frequency that cuts through adult noise — not authority, not volume, but the pitch of someone who has not yet learned to qualify.

“I want to say one thing,” Sovei said. “One thing, and then I’ll sit down.”

The room did not go quiet. But it went quieter.

“Everyone’s arguing about what to do with the membrane. Seal it or open it or crack it. But the membrane is already opening. Something out there is opening it. It’s been doing it for nine days and it doesn’t care what we vote.” She looked at her mother — not at the room, at her mother, and Lian felt the look land like a phase boundary, a discontinuity she could not smooth. “So maybe the question isn’t what we do to the membrane. Maybe it’s what we do with ourselves while it does what it’s going to do. We’ve been hiding inside a wall we didn’t build for eleven generations. I don’t think it’s made us better. I think it’s made us careful. And I don’t know if careful and good are the same thing.”

She sat down. She had not said what the answer was. A man in the back row started to speak and then stopped, as if he’d realized his sentence was no longer the one that was needed.

The conversation resumed, but different now. Not about physics or strategy but about identity — what the hiding had done to them, whether the wall was their protection or their shape and whether those were different things. Lian listened to people she’d known for decades argue as if they were strangers, and realized that on this question they were — she had never had this conversation with any of them, because the wall had made the conversation unnecessary, and the wall was coming down.

Lian sat in the back row and listened. Deris was beside her — she had not noticed when Deris sat down. At some point — she could not have said when — Deris had taken her hand. Not a reconciliation. Not a gesture. A fact, like the ship’s gravity, like the recycled air. Deris’s hand was warm and rough from years of tending the thermal systems and it held hers the way the ship held them all — not because anyone had decided to hold on, but because letting go was not yet the thing that was happening.

The assembly continued. No consensus was reached. No vote was taken. People argued and fell silent and argued again, and somewhere in the arguing a woman whose name Lian didn’t know stood up and said something about her grandmother and the first-generation colonists and what they had feared, and the room went quiet, and then someone else said something about children, and the room went loud, and there was no resolution, and the lack of resolution was not a failure but not a success either — it was the process, and the process was ongoing, and the membrane did not care about their process but they did, and that caring was the only thing they had that the dark forest did not account for.

Afterward, when the commons had emptied — not adjourned but dispersed, people carrying their arguments home like heat — Lian went to the thermal bays.

The channels ran beneath the hydroponic racks in a network of ceramic pipes, heated by the reactor’s secondary loop, cooled by the hull’s radiative surface, cycling water through the temperature gradient that powered the ship’s ecology. In the warm sections, clinging to the pipe walls, the scissurellid snails carried on as they had carried on for longer than any human generation aboard — small, pale, their ridged shells no larger than a child’s fingernail, each one bearing a slit through which the thermal water entered, passed over the animal’s respiratory membrane, and exited. The slit was the central engineering problem of their existence: too wide and the shell collapsed; too narrow and the organism suffocated. Every snail was a negotiation between armor and breath, between the safety of enclosure and the necessity of exchange.

Lian crouched at the edge of a channel and watched them. Dozens, oriented into the current, alive without commentary, without justification, without knowing that the boundary they inhabited had a name or a history or a crisis bearing down on it. They breathed through their slits and the water flowed through their shells and they did not know about the membrane or the lattice or the dark forest or the assembly or the eleven days. They lived at the boundary between warm and cold, inside and outside, protection and exposure, and they had never needed to know what the boundary was for because knowing was not required.

The water was warm. The current was steady. Somewhere above, something pressed against the membrane’s forward seam with a patience that might have been curiosity or appetite or something without a human name. The assembly was still arguing in hallways and kitchens. Deris was somewhere. Sovei had said what she said.

Lian thought about calling another assembly for the morning. She thought about running the controlled-opening model one more time with tighter boundary conditions, about whether Rais could verify her phase-boundary calculations independently, about whether any of it mattered when the probing would continue at its 73.2-hour intervals regardless. She thought about porridge with dried apricots and about Fen and about Deris’s hand and about nothing.

The water flowed. The shells held. The slits faced the current — the smallest doors on the ship — and through them the water moved, in and out, and beyond the ceramic and the hull and the slow-light field and the vacuum, something continued its work at the forward seam with a patience that had already outlasted every argument in the room above. The water flowed. The shells held. The current moved through the slits whether the slits understood the current or not.