What the Machine Knows

Chapter Fifty-Six

What the Machine Knows

Gray discovered the next morning that Galarian rain was not interested in being dramatic.

It did not hammer the windows or turn the institute grounds into a theatrical gray blur. It simply persisted, fine and cold, coating stone walkways and darkening the brick until every building looked older than it had the day before. By seven-thirty the clouds had lifted enough to reveal the quarry walls beyond the research compound, but the air still carried that damp chill that seemed capable of working its way through three layers of clothing out of professional spite.

Gray stood beneath the covered entrance of the visitor residence with coffee in one hand and a tablet in the other while Lyra checked the transport schedule.

“You know you can look at the itinerary after we leave,” she said.

“I’m looking at the weather report.”

“You were looking at the itinerary thirty seconds ago.”

“The itinerary contains the weather contingency.”

Lyra glanced toward the pale drizzle drifting across the courtyard. “It contains one sentence saying the field visit may be delayed if conditions become unsafe. Does that look unsafe to you?”

“No.”

“Then what are you actually checking?”

Gray looked at the tablet, then at her. “Whether Bell changed anything.”

“Did she?”

“No.”

“Then I think we’ve solved it.”

Gray took another drink of coffee and slipped the tablet back into his bag before Lyra could continue. He had slept better than expected, which for him meant almost six hours, though he had woken twice thinking about Bramble and once because Bastiodon had shifted heavily enough in the adjoining room to make the floorboards creak. That had turned out to be nothing more serious than Bastiodon deciding the reinforced bed provided by the institute was less comfortable than the floor beside it.

Their first appointment of the day was away from the institute. Bell met them beside the transport vehicle just before eight, dressed in field clothes and waterproof boots, and wasted very little time on greetings before asking whether both of them had eaten breakfast.

“I ate,” Gray said.

Bell looked at Lyra.

“He did.”

Gray frowned. “Why are you verifying?”

“Because your answer sounded defensive.”

“It was accurate.”

“So was hers.”

Bastiodon climbed into the rear compartment with practiced patience while Gray settled beside him. Bell took the front seat next to the driver, leaving Lyra across from Gray in the back. The road carried them steadily away from the institute and into countryside that seemed to grow older the farther they went: dark stone cottages, low walls furred with moss, fields divided by thick hedges, and narrow roads that curved around hills rather than attempting to cut through them.

Their destination belonged to a retired field researcher named Eleanor Vey. Bell had described her as one of the institute’s earliest long-term Dracovish handlers, though the word handler apparently understated the relationship considerably. Vey had restored her partner eighteen years earlier, during a period when Galarian fossil research was only beginning to understand that successfully producing a living Pokémon did not automatically mean anybody knew how to give that Pokémon a good life.

Gray had asked specifically to see an older Dracovish outside an institutional habitat.

Bell had found him one.

Vey’s property sat beside a long, low lake beneath an old limestone ridge. What had probably once been a farmhouse had been expanded over the years with broad reinforced doors, weatherproof storage sheds, and a covered passage connecting the house to the water. Gray noticed the habitat modifications before he noticed the woman waiting for them beside the gate.

Eleanor Vey was short, broad-shouldered, and somewhere in her late sixties. A wool cap sat beneath the hood of her waterproof coat, and her hands had the roughened look of someone who had spent more of her career moving equipment outdoors than sitting behind conference tables.

She looked at Gray, then at Bastiodon, and finally at Bell. “You brought me another paleontologist.”

Bell smiled. “This one is from Sinnoh.”

“That doesn’t make it better.”

Gray stepped down from the vehicle. “I can leave.”

Vey studied him for a moment. “No, you can’t. Marian already told me you have questions.”

“I do.”

“I know. She sent me the list.”

Gray turned toward Bell.

Bell looked entirely unapologetic. “It saved time.”

Vey had already begun walking toward the house. “Come on, then. Mara’s by the water.”

Gray followed without asking who Mara was. He found out less than a minute later.

The lake behind the house was no longer natural, at least not entirely. Its shoreline had been reshaped into a series of gradual access slopes, broad submerged shelves, and deeper channels that connected with the central body of water. Reinforced stone platforms sat just above the surface in several places, broad enough for a large Pokémon to rest without having to haul its full weight onto land.

Gray stopped near one of the slopes and looked over the shoreline. “This was all modified?”

“Most of it,” Vey said. “The original lake was deep in the middle and useless around the edges. Too steep in some places, too shallow in others.”

Gray crouched and rubbed two fingers lightly across the edge of the access surface. “Textured composite under the gravel?”

Vey nodded. “Exactly.”

“What grade?”

“Eight percent at the steepest. Six through the main access.”

Lyra looked at him.

Gray chose not to acknowledge that.

Movement disturbed the water near the far end of the lake. Something surfaced slowly, lingered for several seconds, and then began moving toward shore.

Mara was larger than Bramble, though age had changed the impression of her body. The musculature through her rear limbs remained powerful, but she moved with more deliberation, wasting less energy with every stroke. Old scars marked one side of her body, pale lines running through restored tissue that had healed years ago. She floated briefly before turning toward the sound of Vey’s whistle.

As Mara approached, Gray watched with the same concentration he had given Bramble the day before. The differences were obvious once he stopped comparing silhouettes and paid attention to movement. Bramble had been explosive and almost recklessly energetic. Mara moved economically, allowing the water to carry her weight until the final possible moment before transitioning onto the shallows.

When she reached the sloped edge, she stopped beside Vey and pressed her head against the woman’s hip. Vey rested one hand against the side of her face with the casual familiarity of someone who had done it thousands of times.

“This is Mara.”

Gray moved closer only after Vey nodded permission. “She’s eighteen?”

“Eighteen years and four months.”

“Original restoration records still intact?”

“In the house. Marian told me you’d ask.”

Bell folded her arms. “I did warn you.”

Gray ignored both of them and focused on Mara.

The Dracovish looked back at him with clear, attentive eyes. Her breathing was steady, and although Gray could see stiffness in one rear articulation when she shifted her weight, it was not severe enough to prevent her from moving comfortably.

He indicated the joint. “Left pelvic articulation?”

Vey nodded. “Degenerative changes. First showed clearly on scans around fourteen.”

“Pain management?”

“Low-dose anti-inflammatory after heavy activity. She gets hydrotherapy too, though mostly because she enjoys it.”

Gray looked toward the enormous lake.

Vey laughed. “Yes, I know how that sounds.”

“It sounds efficient.”

“That too.”

Mara moved past them toward Bastiodon. Gray stiffened for a moment out of reflex, then relaxed when Bastiodon remained calm. The introduction proceeded much more slowly than Bramble’s had the previous day. Mara sniffed along the edge of Bastiodon’s shield, paused with her head tilted slightly, and waited while he answered with one low rumble.

Then she touched the side of her head against his armor.

Gray glanced toward Bell. “Better manners.”

“Age.”

Vey heard him and shook her head. “No. She was worse when she was young.”

Mara continued ignoring the discussion as though she had long ago accepted that humans would say embarrassing things within earshot.

They spent the next hour moving around the property while Gray asked questions at a pace Vey tolerated remarkably well. He wanted to know how much time Mara spent on land, what temperatures she preferred, whether activity changed seasonally, how her appetite shifted with age, which surfaces caused the most joint strain, how often she chose human company, and whether she ever avoided areas she had used comfortably when younger.

Vey answered nearly everything without complaint. Around the thirtieth question, however, she looked toward Bell and said, “You undersold him.”

Bell smiled. “I didn’t think you’d believe me.”

Gray had already moved on. “How much of the lake does she use during winter?”

“Less than half. She favors the eastern section because the water stays warmer.”

“Artificial heating?”

“Geothermal exchange system.”

Gray turned toward her more sharply.

Vey pointed toward the house. “Installed twelve years ago.”

Lyra waited until Gray had looked back toward the lake before speaking. “How deep is your southern pond?”

He answered without thinking. “Three-point-eight meters at center depth, two-point-one across the western shelf, though the shelf needs regrading because sediment accumulation has reduced the usable—”

Gray stopped.

Lyra smiled.

He stared at her. “You did that deliberately.”

“I asked a question.”

“You already knew the answer.”

“I knew approximately.”

“You knew enough.”

Vey looked between them. “You have a pond?”

“Several,” Gray said.

“How much land?”

“Twenty-seven acres total. Less usable water than that.”

Vey considered the number, then glanced toward Mara before returning her attention to Gray. “You could build for one.”

Gray looked toward the lake.

Mara had moved several meters out again and was floating comfortably in water deep enough to take almost all the strain off her joints. Gray found himself mapping the ranch without intending to: the southern pond, the lower drainage area, the slope beyond the western fence, the section Marcus had once mentioned could be excavated without interfering with the existing water table.

“I know,” he said.

It was the first time he had admitted that aloud.

Until then, the possibility of bringing home a Dracovish had remained safely abstract, something that belonged to research papers, restoration ethics, and a future decision he could continue delaying. Standing beside Mara’s lake made it physical. Habitat could be measured. Water could be deepened. Slopes could be changed. Heating could be installed. None of that meant Gray had decided anything, and it certainly did not amount to permission.

But for the first time, he was no longer asking whether it could be done.

He was beginning to understand what doing it properly would require.

Mara returned to the water after half an hour on land, descending the sloped bank with enough care that Gray could see exactly where the stiffness in her rear joint mattered. The moment the water took her weight, the difference was obvious. Her movements lengthened, the hesitation disappeared, and within seconds she was gliding away from shore with a confidence that made her age much harder to see.

Gray watched until only the top of her head remained visible above the surface.

Inside the house, Vey had organized Mara’s medical history better than some institutions Gray had worked with, which earned immediate respect. She spread old scan reports, behavioral logs, dietary adjustments, environmental changes, and several handwritten notebooks from Mara’s first five years across the kitchen table while Gray worked through them in chronological order. Bell stood near the window with a cup of tea, occasionally glancing toward the lake as though checking that Mara had not somehow decided to leave the property while nobody was looking. Pasted markdown

The early records were worse than the later ones, though not in the dramatic way Gray had half expected. They were worse in the way ignorance usually was: gradual, reasonable at the time, and obvious only once somebody knew enough to recognize the pattern. Mara had been given too little water access during her first eighteen months, and much of her terrestrial exercise had been modeled on rehabilitation plans for large-bodied Dragon-types. When she resisted remaining on land, early handlers had treated that as poor conditioning or avoidance behavior rather than considering that land itself might be the problem.

Gray found the first note where Vey had questioned that assumption.

Increased agitation after extended dry exercise. Resolves rapidly after water access. Consider environmental mismatch rather than avoidance behavior.

He read it twice before looking up. “You wrote this?”

Vey set down the kettle. “Yes.”

“How long before anyone listened?”

“About three months.”

Gray’s expression tightened.

Vey caught it. “Could’ve been worse.”

“That is not a defense.”

Unauthorized tale usage: if you spot this story on Amazon, report the violation.

“I didn’t say it was.”

He returned to the records. Once aquatic access expanded, the difference became difficult to dismiss. Stress markers fell, voluntary activity increased, appetite became more consistent, and joint inflammation decreased enough that Mara’s medication could be reduced. The improvement was not subtle enough to excuse how long it had taken anyone to recognize the problem.

Gray looked toward the window. Mara was visible near the middle of the lake now, only occasionally breaking the surface.

“When did you decide to keep her?”

Vey paused over her tea.

The question had not been on Gray’s list, and from the way Bell glanced over, she knew it.

“About six months after restoration.”

“Why?”

“Because she’d already decided.”

Gray looked at her.

Vey settled into the chair opposite him. “She followed me everywhere she was allowed to. Got irritated whenever another handler moved her somewhere without me, and once stopped eating properly when I left for a week. Eventually we had to admit that pretending the relationship was temporary was causing more disruption than acknowledging it.”

Gray rested one hand on the open notebook. “Any regret?”

Bell looked at him. Lyra did too.

Vey did not need time to think. “No.”

She wrapped both hands around her cup and glanced toward the lake. “I regret what we didn’t know when she was young. I regret treating land exercise like some developmental milestone she needed to meet instead of asking whether she preferred the water because being in the water felt better. I regret the first enclosure, and I regret two years of feeding schedules built around what the literature said instead of what she was actually telling us through her behavior.”

Gray waited.

Vey’s gaze remained on the window. “I don’t regret her.”

The distinction landed harder than any grand defense of restoration would have. She was not pretending their mistakes had been harmless, and she was not using Mara’s current health to retroactively excuse them. She simply refused to confuse regret over poor care with regret over the Pokémon who had lived through it. Pasted markdown

Gray looked back at the notebooks. “I understand.”

Vey studied him. “Do you?”

“I think so.”

“You’re considering restoring one.”

Gray glanced toward Bell.

Bell looked entirely innocent, which convinced him she was responsible.

“This is spreading internationally.”

Vey laughed. “You came to my house to measure a pond.”

“That is not proof.”

“You asked the depth before you asked her name.”

“That is not true.”

Lyra looked up from the records she had been skimming. “It is nearly true.”

Gray ignored her, which was becoming the most efficient available response.

They finished the visit shortly before noon. Mara returned to shore when Vey walked them out, though her interest in Gray had not increased much during the morning. She inspected him briefly from a respectful distance before deciding Bastiodon remained more interesting.

Vey rested one hand against Mara’s side. “You’re asking the wrong question if you’re trying to decide whether she should exist.”

Gray looked up. “What should I be asking?”

“Whether you can give the next one what Mara should’ve had from the beginning.”

That followed him through most of the drive back.

The field site came next.

The weather had improved enough that Bell refused to cancel, which Gray appreciated. The rain had stopped entirely by the time they left the main road, though the quarry tracks remained slick and the wind had strengthened across the higher ground.

The excavation area lay inside a broad cut through exposed sedimentary rock several kilometers from the institute’s primary fossil storage facilities. It was no longer an active commercial quarry, though parts of the site had once been worked heavily. Research teams now occupied portions of the exposed faces under a permit system Bell described as “an administrative nightmare nobody has successfully killed.”

Gray liked the place immediately.

Within five minutes, he also understood why Roark would have hated missing it.

The strata were beautiful. Pale sedimentary bands alternated with darker compressed layers, some tilted sharply enough that millions of years of geological history could be followed across a single exposed face. Tagged excavation grids covered several accessible sections, while field tents farther downslope sheltered technicians cleaning and documenting finds before transport.

Gray took out his phone.

Lyra noticed immediately. “For Roark?”

“No.”

She looked at him.

Gray sighed. “Yes.”

He started recording before she could say anything else.

Bell led them toward one of the historical Fish-component recovery zones along a marked path that should have taken twenty minutes. It took almost forty because Gray kept stopping at bedding planes, exposed inclusions, and old excavation markers.

“You know we have photographs,” Bell said after the fourth delay.

“I know.”

“And scans.”

“I know.”

“And stratigraphic maps.”

Gray looked up from the rock face. “I want context.”

Bell smiled. “There you are.”

The Fish recovery zone sat low in the exposed sequence. Bell pointed out the relevant horizon and showed Gray where several important components had been recovered over the years. Some had been isolated finds. Others appeared near aquatic fossil material dense enough to suggest a complex ancient ecosystem rather than a narrow single-species deposit.

Gray crouched beside one of the old markers. “How confidently can you correlate this horizon with the Drake site?”

“Locally?”

“Yes.”

“Not confidently enough for the answer you want.”

He looked up.

Bell continued before he could ask. “Regional correlation is strong. Exact temporal equivalence is not.”

“What range are we talking about?”

“That depends on which dating model you prefer.”

Gray frowned. “That is a terrible sentence.”

“It is also true.”

They spent nearly an hour in the Fish zone before driving to the Drake locality. The second site lay farther inland and higher in the exposed sequence, and Gray understood the problem before Bell explained it.

He stood near the edge of the excavation grid and looked from the marked layer to the geological notes Bell had handed him. “They’re not from the same bed.”

“No.”

“Not even the same immediate formation.”

“Related depositional system.”

“That is not the same thing.”

“I know.”

Wind pulled at the front of Gray’s coat as he looked across the exposed rock. “So there is no direct evidence that Fish and Drake were parts of one organism.”

“No.”

“Any associated skeleton?”

“No complete one.”

“Partial association?”

“Debated.”

“Trackways?”

“Nothing diagnostic.”

Gray thought for another second. “Coprolites?”

Bell stared at him. “You’re reaching.”

“I’m being thorough.”

“You are asking whether anyone found fossilized shit that could rescue the hypothesis.”

“It has happened before.”

Bell sighed. “That is unfortunately true.”

Gray looked across the site again.

The problem felt worse here than it had inside the institute. A restoration machine could make the evidence look coherent because the machine ended with a living Dracovish. The ground did not care about coherent conclusions. Fish and Drake components appeared within overlapping stretches of Galar’s ancient history, but not in any pattern that allowed a paleontologist to confidently reconstruct one animal from both.

If anything, the geology made it easier to imagine them as remains from distinct organisms that had inhabited the same broader ancient ecosystem.

Lyra came to stand beside him. “What are you thinking?”

“That Rowan was right to be annoyed.”

Bell, close enough to hear, nodded. “He usually was.”

Gray turned toward her. “If they weren’t one organism, why does the machine consistently make one?”

Bell slipped her hands into her coat pockets and looked toward the excavation wall before answering.

“That’s the question.”

Gray hated how much he liked that answer. Pasted markdown

They returned to the institute after three, where Bell’s promised engineering specialist was waiting in the restoration wing.

Her name was Dr. Priya Shah. Within ten minutes Gray learned two things about her: she preferred Priya, and she disliked being called a restoration engineer because, in her words, half her job involved correcting assumptions made by paleontologists who thought machines were magical boxes.

Gray liked her immediately for that.

She did not return the favor quite as quickly.

“You’re Professor Linden.”

“Gray.”

Bell glanced between them. “Careful. He only lets people use that after they’ve argued with him.”

Gray frowned. “That is not true.”

Priya looked at Bell. “How long did it take you?”

“Three hours.”

“Efficient.”

Gray folded his arms. “Can we discuss the machine?”

Priya smiled. “There it is.”

She led them into the control suite, where the restoration chamber looked almost entirely different from this side of the glass. The previous day Gray had focused on the visible hardware. Today Priya brought up the reconstruction architecture itself across several layered displays.

“Traditional restoration begins with a dominant biological pattern,” she explained. “One fossil, one inferred organism. The fossil can be incomplete, sometimes extremely incomplete, but the reconstruction system still has one internally coherent biological identity around which it resolves missing structures.”

Gray nodded. “That part is familiar.”

“Galarian combinatorial restoration is not.”

“I had noticed.”

Priya ignored that. “When we input multiple fossil sources, the system does not classify one as primary and the other as accessory. It builds what we call a shared viability field.”

Gray frowned. “Define viability field.”

“Not the public-relations version?”

“Please.”

Priya expanded the model until several overlapping lattices of information filled the display. “We extract whatever preserved pattern data remain available from each fossil source: structural information, biochemical signatures, developmental constraints, energetic patterns, restoration markers associated with the original organism. Some categories preserve far better than others.”

“So far, normal.”

“Then the system maps compatibility between the surviving patterns.”

Gray leaned closer. “Compatibility according to what?”

Priya’s expression changed. “That is where things become unpleasant.”

“Good.”

She looked at him. “You say that now.”

Several portions of the model became opaque as she adjusted the display. “Missing biological information creates possible solution spaces. Traditional restoration constrains those spaces around one inferred biological identity. Multiple-source restoration instead constrains them around all viable inputs simultaneously.”

Gray stared at the model. “That sounds like extrapolation.”

“It is.”

“How much?”

“Variable.”

“Based on stored species templates?”

“No.”

“A species database?”

“No.”

“Comparative developmental models?”

“Partially, but not enough to generate the organism.”

That finally made Gray turn toward her.

Priya rested one hand against the console. “The machine is not selecting Dracovish from a menu. There is no complete Dracovish blueprint hidden somewhere in the software. Give the system sufficiently intact Fish and Drake pattern sets, and it resolves those inputs toward a stable biological configuration.”

“Why that configuration?”

“We don’t know.”

Gray straightened.

Priya looked almost pleased by his irritation.

“We understand many of the constraints and part of the reconstruction pathway. We can predict outcome ranges extremely well. What we cannot do is point to one buried instruction and say, there, that is Dracovish.”

Gray looked back at the model. “So the machine is inventing.”

Priya shook her head. “That gives the system more freedom than it has. It cannot create whatever it wants.”

“Creating, then.”

“Still too broad.”

“Resolving.”

“Closer.”

Gray moved around the display slowly, following the layers as Priya separated them. “What happens if you alter one component?”

“The outcome shifts within limits. Different Fish material produces individual variation. Different Drake material does the same. But sufficiently preserved Fish plus sufficiently preserved Drake converges very strongly toward the morphology we classify as Dracovish.”

“Which means the relevant information exists in the fossils.”

“Yes.”

“Enough information to produce Dracovish.”

“Yes.”

“But not enough to prove that exact organism historically existed in that form.”

Priya smiled. “Now you understand why paleontologists hate our department.”

Bell leaned against the counter. “And why Rowan kept writing letters.”

Gray ignored both of them and returned his attention to the projected lattice.

For the first time since finding the fossils, something shifted into place. It was not an answer. If anything, it created more questions. But the shape of the problem changed.

He had been asking whether the machine reconstructed an authentic body.

That might not be what it was doing at all.

“The machine may not actually be reconstructing anatomy,” Gray said slowly.

Priya tilted her head.

“Not directly. It may be reconstructing viability.”

Bell’s expression changed slightly.

Priya looked back at the model before answering. “That is not how I would phrase it in a technical paper.”

“Why?”

“Because reviewers would murder me.”

“Cowards.”

“But as a conceptual shorthand?” She considered it for several seconds. “Yes. That is closer than most.”

Gray looked at the overlapping lattices again. “Then historical authenticity becomes a separate question. The fossil information is authentic, the resulting Pokémon is biologically real, and what remains unresolved is whether that particular viable arrangement ever existed naturally.”

“Exactly.”

That distinction did not solve the Dracovish problem.

It made the problem much more interesting.

“The unresolved claim is whether that specific viable arrangement existed naturally.”

Priya nodded.

Gray looked back at the layered model hovering above the console. “That distinction should be in every public explanation.”

Bell laughed from where she was leaning against the counter. “You have been in Galar for two days.”

“That does not make me wrong.”

“No,” she said. “It makes you ambitious.”

Gray remained at the console for another hour, and the conversation became less comfortable the farther he pushed it. Success interested him, but failure interested him almost as much, particularly because failed restoration attempts revealed where the limits of the system actually were.

Low-quality fossil inputs were sometimes rejected before tissue development began. Certain fragment combinations never achieved a stable reconstruction field at all. More troubling were the cases where the system reached a technically viable developmental pathway that medical review still considered unacceptable because the projected anatomy carried too high a risk of severe dysfunction or suffering.

Gray’s expression tightened as Priya explained it.

“So you terminate before full restoration.”

“If the projected outcome fails welfare review, yes.”

“Mandatory?”

“Now it is.”

The word caught immediately. “Now?”

Priya exhaled and rested one hand against the console. “Historical practices were worse.”

Gray rubbed at his jaw. “That keeps happening.”

“Yes.”

There was nothing defensive in her answer, which made it harder to be angry at her personally. Gray shifted instead to another part of the problem that had been bothering him since the previous day.

“What about reproduction?”

Priya’s expression changed. “Very limited data.”

“Why?”

“Low population numbers, ownership fragmentation, lack of deliberate breeding programs, and ethical objections to creating a breeding study around Pokémon whose basic developmental history is already controversial.”

Gray glanced toward Lyra.

She caught the look immediately. “Don’t.”

“I didn’t say anything.”

“You were about to.”

Priya looked between them. “Is this common?”

“Yes,” Lyra said at the same time Gray said, “No.”

Bell laughed.

The reproductive records gave Gray very little satisfaction. There were scattered observations of pair-bonding behavior and a handful of reports suggesting restored fossil Pokémon were capable of ordinary reproductive interest, but nothing approaching a robust dataset. Nobody could confidently say whether offspring would preserve the same morphology, whether restored combinations bred true, or whether developmental factors outside the restoration process changed the result.

Gray added three questions to his notes, reconsidered one of them, and then added four more.

By the end of the session, his original decision framework had begun to resemble the opening section of a research grant.

Bell noticed before he did. “You realize you cannot answer every one of those before deciding.”

Gray looked down at the list. “I know.”

“Do you?”

“Yes.”

“You added ‘comparative offspring morphology’ ten minutes ago.”

“That is a legitimate question.”

“It is.”

“I’m glad we agree.”

“It is also a question that could take twenty years to answer properly.”

Gray frowned at the notebook.

Bell smiled. “That’s what I thought.”

They left the control suite close to sunset. Bastiodon had spent most of the afternoon in one of the institute’s outdoor enclosures, where the staff had apparently discovered that he enjoyed pushing a reinforced enrichment block around with his shield. Gray found him beneath a covered area near the edge of the enclosure, settled comfortably while rain began tapping against the roof again.

Gray sat beside him and let the silence last for a while.

Bastiodon shifted closer after a few minutes, enough that the edge of his shield rested against Gray’s shoulder. Gray put one hand against the old armor and ran his thumb over a scar he had known for years.

“I’m trying very hard not to use you to justify this.”

Bastiodon rumbled quietly.

Gray looked at him. “That is not permission.”

Another rumble came back, slightly deeper.

“You are biased.”

Bastiodon nudged him hard enough to move him several centimeters along the bench.

Gray sighed. “That is also not evidence.”

Bastiodon settled again as though the matter had been resolved to his satisfaction. Gray leaned against his side and watched the rain beyond the shelter.

The comparison was impossible to avoid.

Shieldon had come from a far more conventional fossil. Restoration had still involved uncertainty because every restoration did, but nobody had seriously questioned whether Shieldon represented a coherent historical species. Gray’s father had found the fossil, researchers had restored it, and Shieldon had emerged furious at the light and very much alive.

Dracovish was different. Gray would knowingly authorize a process whose historical fidelity remained unresolved, and he had spent most of the last several days treating that uncertainty as though it might be disqualifying by itself.

The more he learned, the less certain he was that it should be.

The machine would not fabricate a Pokémon out of fantasy. The fossil material was real, the biological information embedded within it was real, and whatever emerged from the restoration chamber would be real too. The uncertainty lay somewhere between those facts, in the question of whether the final arrangement belonged to ancient history or to the reconstruction process itself.

That did not reduce Gray’s responsibility.

If anything, it sharpened it.

Bell found him there about twenty minutes later. She carried no files, which Gray considered immediately suspicious.

“You’re done for today,” she said.

“I can decide that.”

“No. Lyra decided it.”

Gray looked toward the building. Through the glass doors, Lyra stood with her tablet in hand and was doing an unconvincing job of pretending not to watch them.

“Mutiny.”

Bell sat on the edge of the covered platform, far enough from Bastiodon to respect his space. “How bad is the list now?”

“Thirty-two questions.”

“You arrived with nineteen.”

“Progress.”

“That is not what that word means.”

Gray looked out at the rain for a while before answering anything else. Bell did not rush him, which he had come to appreciate.

“Mara helped.”

“I thought she might.”

“The field site did not.”

“I thought it wouldn’t.”

“Priya made everything worse.”

Bell smiled. “She usually does.”

Gray rubbed one hand slowly over Bastiodon’s shield. “I’m more confident that a Dracovish can have a good life.”

Bell nodded.

“I’m less confident anyone can honestly tell me what the hell one is.”

“That also seems accurate.”

Gray glanced toward her. “That doesn’t bother you?”

“It used to.”

“And now?”

Bell looked across the enclosure toward the habitat buildings. “Now I think we spent too long assuming uncertainty about origin had to mean uncertainty about value.”

Gray considered that carefully.

Bell continued before he could respond. “I still want the evolutionary answer. I want a complete skeleton that makes every argument disappear. I want to know whether the ancient world ever contained something exactly like Bramble, because that would settle questions people have been fighting over for decades.”

“But?”

“But Bramble does not become less alive because I don’t know.”

Gray looked away.

That thought followed him into the evening.

After dinner, he returned briefly to the observation wing. He expected Bramble to be swimming again, trying to dismantle enrichment equipment, or harassing whichever handler had been assigned to him.

Instead, Bramble was asleep.

He had settled partly in the water near one of the shallow shelves, his head resting against the edge while the rest of his body remained supported by the pool. His breathing was slow and steady, and nothing about the scene demanded attention.

He looked ordinary.

Not anatomically. Gray doubted Dracovish would ever look ordinary to him in that sense.

Behaviorally, though, there was nothing strange about it. Bramble was simply a tired Pokémon sleeping at the end of a day whose scientific importance mattered almost entirely to humans.

Gray stood at the glass for several minutes.

The day before, Bramble’s energy had been persuasive. Watching him swim, investigate Bastiodon, and try to steal Gray’s bag had made it impossible to reduce him to a reconstruction problem.

The stillness was more persuasive.

A life could not justify itself through spectacle forever. It had to remain worth living when nothing interesting was happening.

Back at the visitor residence, Lyra had already gone through the next day’s schedule by the time Gray opened his notebook.

“You’re making another list.”

“Yes.”

“Should I be worried?”

“Probably not.”

“That sounds reassuring.”

Gray ignored the comment and stared at the blank heading. The first version sounded too much like a research protocol, so he deleted it. The second sounded like he had already decided to restore the fossils, which irritated him enough to delete that one too.

Eventually he wrote:

Conditions Required Before Restoration

Lyra noticed the heading from across the room. For once, she did not say anything.

Gray appreciated that.

The list beneath it was much shorter than the one he had made in Sinnoh. He needed adequate aquatic habitat, long-term veterinary support, documented pain monitoring, enough behavioral freedom that the Pokémon could choose how it lived, no evidence of severe predictable developmental instability, access to an experienced restoration facility, emergency support during early development, and a care plan that extended long beyond the point where Dracovish stopped being an interesting research question.

Gray read the list twice.

Yesterday he had wanted answers to everything.

Tonight he needed enough certainty to accept responsibility.

That was different.

After several minutes, he added one final condition beneath the others.

Wanting the Pokémon cannot be the only reason. It also does not have to be irrelevant.

Gray stared at the sentence longer than any of the practical requirements.

Lyra finally looked over. “That one took you a while.”

“Yes.”

“Do you believe it?”

Gray turned toward the dark window. The institute’s habitat buildings were visible in the distance, their lights blurred slightly by rain.

“I think I’m starting to.”

Lyra nodded and returned to her tablet.

Gray looked back at the notebook.

He still had not said yes. Nothing had been authorized, no machine had been turned on, and the Fish and Drake remained exactly as they had been when Elias found them in Sandgem.

But somewhere during the day, Gray had stopped trying to construct an argument for why restoration should not happen.

He had begun working out what would have to be true before he could responsibly allow it.