The Thuvari Who Knows Too Much

The research bay smelled like paper and something faintly mineral, the way river stones smell when you pull them out of cold water. I noticed it the moment Vasht keyed open the door, before I saw anything else: four long work surfaces covered in physical documents alongside display screens, a rack of equipment along the far wall with components I didn't recognize, and Sonn standing at the center surface with their hands resting flat on a spread of materials as if they'd been waiting long enough to have practiced the pose.

Sonn was taller than they had looked on the comm screen. The Thuvari build ran lean and long, with a head proportioned slightly larger than human scale and eyes that were solid dark without a visible division between iris and pupil. The skin had a gray-blue undertone that changed with the angle of the light. Four fingers on each hand, the joints placed differently than a human's, which explained the particular quality of their hand gestures through the screen.

They looked at me the way I'd looked at the Solvari crew during the first supervised walk-through: cataloguing. I let them do it.

"Derek Caulfield," Sonn said. Not a greeting, exactly. More like reading a label out loud to confirm it matched the box.

"That's me."

Vasht stayed in the doorway. "I will leave you both for two hours. The secure link is active if you need me." She looked at Sonn. "He has read the full preliminary research queue."

"I know," Sonn said. "I can tell from his questions, which is to say I can tell from the fact that he will have them." They gestured toward a chair at the near end of the work surface, the kind of spare invitation that assumed I would take it or not based on my own assessment. "Sit, please, if you are comfortable."

I sat. Sonn pulled a stool to the opposite side of the surface and lowered onto it with the unhurried economy of someone who moved deliberately by habit rather than by effort.

The documents between us were Thuvari script mixed with diagrams, which I couldn't read but could observe. The diagrams showed curved lines and node clusters, overlapping circles with measurement annotations I couldn't interpret. One sheet had what looked like a cross-section of something biological, though whether it was Thuvari anatomy or Aeluun anatomy or something else entirely I had no way to tell.

"The files Vasht forwarded," I said. "Were those complete records."

Sonn's hands shifted on the surface. "They were the complete records the Compact holds. The Thuvari hold additional historical material that I have access to in my professional capacity." A slight pause before the next sentence. "The Compact records and the Thuvari records do not fully agree on several points of early integration history."

"Which points."

"The failure rate in early adoption, primarily." Sonn looked at me steadily. "The Compact's records indicate two documented cascade failures in the Thuvari adoption period. The Thuvari historical records indicate seven."

I held that for a moment. "Why does the Compact have the lower number."

"Because the Thuvari provided the records and selected what to provide." No apology in the statement, just the flat acknowledgment of how information works when one party controls its release. "I am telling you this because I believe you will not make useful decisions without accurate foundational data, and because I believe you will find the discrepancy eventually if you continue reading at the pace you have been."

"Seven failures." I looked at the diagrams. "What were the other five."

Sonn's hands rearranged. "Three additional cascade failures, outcomes similar to the two you read about. One integration that progressed to a capability level the adopting Thuvari then chose to discontinue, voluntarily, with my community's support. The process of discontinuation was successful and the subject lived a full lifespan." A longer pause. "And one integration that did not fail, did not cascade, and developed to a point where the Thuvari community made a collective decision that the subject needed to live separately from the community."

I watched Sonn's hands. They had gone very still. "Why."

"Because what the subject had become was incompatible with how our community functions." Sonn said it evenly, without heat. "We are, by nature and by choice, a non-confrontational species. Our social architecture depends on a degree of interpersonal transparency, on the ability to read one another accurately, that is foundational to how we live. What the integration had given this subject was the ability to read the community accurately while the community could no longer read them accurately in return. The asymmetry was not manageable."

I thought about that. "What happened to them."

"They chose to go." Sonn's voice didn't change, but something in their posture settled infinitesimally lower. "They lived another forty years on a separate world, doing work that was useful. They corresponded with the community regularly." A breath, not dramatic, just the ordinary kind. "I am describing this case to you not because I believe your situation will develop this way, but because I have committed to complete disclosure and this is part of what the historical record contains."

"Appreciated." I meant it. "Let's talk about the mechanics."

Sonn nodded, and the hands came back to active arrangement on the surface, pulling a diagram toward us. "The Aeluun system is, to the extent the Thuvari understand it, a biological interfacing technology from a species that no longer exists. We do not know when it was created, only that the devices predate any space-faring civilization we have found historical evidence of. We do not know how many devices exist. We have found sixty-three across explored space. The Dominion holds at least forty-one of those, possibly more."

"The Aeluun themselves."

"Gone before any current civilization developed faster-than-light travel. We know they existed from the devices and from residual biological traces in the devices' design: the interface is calibrated for biological organisms, not mechanical or synthetic systems, which tells us the Aeluun were biological. Beyond that, we have inference and nothing more."

"The compatibility scores." I tapped the surface near the nearest diagram. "Those measure biological resonance with the interface architecture."

"Essentially yes," Sonn said. "The device reads neurological structure, metabolic processing patterns, and what I would describe loosely as the architecture of how an organism processes complex simultaneous information. Species with high scores have neural processing systems that are compatible with the interface's native communication format. Species with low scores do not, and forcing integration would do nothing useful beyond the first calibration stage."

"What makes my score high."

Sonn considered this with the particular pause I was learning meant they were choosing precision over speed. "Human neural architecture has a processing characteristic that the Thuvari have observed in other high-compatibility species: you do not compartmentalize incoming information before processing it. You run multiple information streams in parallel, allow them to interact, and produce outputs that integrate the interaction. Most species process information sequentially or in limited parallel tracks. Humans appear to run a wider simultaneous load before resolution." They looked at me. "The device's interface format operates the same way. You are, neurologically speaking, already running something close to the system's native language."

"And the Thuvari."

"We score between sixty and seventy-eight, historically. Enough for genuine integration. Not enough for the deeper compatibility levels." Another pause, and this one had weight to it. "When we first adopted the technology, we did not understand the score's implications. We believed sixty-eight was close enough to eighty that the difference was marginal. We were wrong."

I kept my eyes on the diagram. "You found out how wrong."

"Yes."

The word sat there, not asking for anything.

"The agricultural and architectural applications," I said. "What did those actually look like."

The quality of Sonn's posture changed slightly, the way a person's does when a conversation moves from difficult material to something they're more comfortable in. "In agricultural use, the integration enhanced spatial processing, pattern recognition of plant growth and terrain and water distribution, the ability to track a large number of environmental variables simultaneously and to sense imbalances before they became visible to unenhanced perception." They traced a finger along one of the curved diagram lines. "A Thuvari farmer with integration could manage a plot of land three to four times the size of what an unenhanced farmer could manage effectively, because they perceived the land's status continuously and in detail. Problems were identified early. Interventions were precise."

"And architecture."

"Load calculation, structural integrity reading, the ability to perceive the stresses in a building's material in the way you might perceive the difference between a solid surface and a hollow one by striking it with your hand. Greatly extended. A Thuvari architect with integration could walk through a completed structure and identify stress fractures that would not be visible to conventional detection for another year." Sonn's hands made a gesture that was probably explanatory in Thuvari body language, opening and then settling. "We built good things. Nothing collapsed that should not have held. The technology served us well inside those applications."

"Until it didn't."

"Until the seven cases I have described." Sonn folded their hands. "After the seventh, the community held a sustained consultation that lasted approximately one of your years, and the decision at the end of it was to cease adoption. Existing integration cases were supported through their natural lives. No new integrations were approved." A small, considered pause. "I was not alive for that consultation. But its outcome shaped everything about my understanding of this material."

Outside the ship's hull, there was only the drive's hum and the recycled air moving through the vents. Inside the research bay, the mechanical sounds were the same but the quality of the air was different, thicker with the smell of the documents and the faint mineral note I'd noticed on entry.

"The Dominion," I said. "They didn't make the same decision."

"The Dominion's founding species, the Vorath, made the opposite decision with full awareness of the risks. They understood what the Thuvari had found, because we shared our research history with the early Compact before the Compact and the Dominion divided. They took that information and determined that what the system could do under combat development conditions was worth the cost of what it did to the users over time." Sonn's voice was very steady and completely neutral, which told me more than any inflection would have. "The Dominion's gifted soldiers, at the early stages of their development, are using the same kind of enhanced spatial awareness and pattern recognition that the Thuvari used for agriculture. Under the stress conditions of sustained combat, those capabilities develop further and faster. The system responds to what the user requires of it. What a soldier requires of it is not what a farmer requires of it."

"And what does the system do with that."

"It amplifies the solutions the user generates under stress. This is the core mechanic and it is important that you understand it precisely." Sonn pulled a second diagram toward us, this one showing overlapping nodes with arrows indicating direction of influence. "The system does not impose capability. It does not have a library of abilities it distributes to compatible users. What it does is read the patterns of how you solve problems, identify the neurological circuits you use when you are under pressure and performing well, and amplify those circuits. It makes you more efficiently yourself."

I looked at the diagram. "Which means what I become depends on what I already do."

"Entirely." Sonn met my eyes. "A Dominion soldier who solved problems through aggression and speed before integration becomes faster and more aggressive. A Dominion soldier whose combat value was technical proficiency becomes more technically precise. The system does not change what you are. It enhances what you are, in the direction you apply it."

I sat back from the diagram. "No Thuvari ever developed combat capabilities."

"No Thuvari ever needed to survive a firefight," Sonn said, and the words landed with the specific weight of something that had been sitting in their chest for a while. "The stress conditions that feed the development are the conditions of mortal urgency. The system learns from how you solve the problem of being killed. Farmers are not in mortal danger during the harvest. The stress inputs that fed our integration were different in kind, not just in degree, and so the outputs were different."

"And nobody knows what happens when the stress inputs are different from anything the existing cases experienced."

"We know what happens with Dominion soldiers, because we have their combat data and their integration histories, in cases where we have captured relevant records." Sonn looked at the surface, not at me. "We know what happens with Thuvari adopters because we have full longitudinal records on everyone who integrated. What we do not know is what happens when the compatibility score is ninety-four point seven and the stress environment is active combat and the user's neurological processing architecture is what yours is." They looked back up. "You will be generating the data for that case."

The drive hummed. Somewhere in the ship's mechanical systems, the recycling interval clicked through its cycle.

"The failure modes," I said. "You mentioned cascade failure in the files. What else can go wrong that isn't in the Compact's records."

Sonn's hands went very still again, the tell I'd now identified for material they found costly to discuss. "The case I described, of the subject who could no longer live within the community. The failure mode there was not technical. The integration did not malfunction. The failure was social and relational." A deliberate breath. "When the pattern recognition develops to a high level, the user perceives the behavioral and physiological signals of others with a clarity that other people cannot reciprocate. You begin to read people accurately in ways they cannot read you. This creates an asymmetry." Their eyes stayed on mine. "In the agricultural context, this was manageable, because reading the land does not require the land to read you back. In a social context, particularly over years, the asymmetry changes relationships in ways the integration user may not fully control or want."

"People can't tell what you're thinking, but you can tell what they're thinking."

"In a progressively detailed way, yes."

"How far does that extend."

"At high development levels, in the historical cases with the highest Thuvari compatibility scores: intent, in some cases, before the subject themselves has consciously formed it. Reading the physical precursors of decision." Sonn's hands moved, not arranging anything, just moving. "A Thuvari who develops to that level and lives within a community that expects mutual transparency becomes something that community cannot accommodate. Not because of malice, but because the relationship has become structurally unequal."

I thought about that for a moment, practically. Reading intent before the person has formed it meant reading movement before the movement happened, reading the pre-signals that preceded action. In a farming community that was an unmanageable asymmetry. In a firefight it was a different calculation entirely.

"Sonn." I kept my voice flat and informational. "You're steering every hard capability I ask about toward what it does in a community, not what it does in a fight."

Sonn looked at me for a long moment without answering.

"Which tells me," I said, "that you know what it does in a fight and you're afraid of it."

The long fingers settled on the surface. "I am afraid," Sonn said, and the directness of it was unexpected enough that I held still, "of what happens to a person who develops those capabilities under sustained combat stress, specifically. The relational asymmetry in a community is painful. The relational asymmetry in a combat environment, between someone who can read intent before it forms and the people they are fighting against, is a different thing. It is not painful in the same way." A pause. "I do not know, precisely, what it does to a person to operate in that state continuously over months or years. I have the Thuvari records, which do not cover this territory. I have the Dominion's combat data, which covers outcomes but not internal states. I have inference."

"What does the inference say."

"It says that the system amplifies what you require of it, and if what you require of it is the continuous reading of people as threats or non-threats, as vectors of intent to be neutralized, it will become very good at that specifically." Sonn's voice was as steady as it had been throughout, but the words were the most direct they had been. "I am not telling you this will happen. I am telling you I do not know that it will not."

The air moved through the vents in its measured interval.

"What can I do to stay out of that territory," I said. It was the same question I'd asked through the comm screen, in the chapter Vasht had read over my shoulder. I asked it again because the answer from Sonn in person was different from the answer Sonn gave through a screen.

"Be deliberate," Sonn said, as before. And then, unlike before: "Use the pattern recognition for what you need in the moment, and when the moment ends, let it rest. The system responds to demand. If you run it at full load continuously, it will calibrate toward continuous full-load operation as the baseline. If you allow it periods of genuine low-demand, it will retain more flexibility." They looked at their hands. "This is easy to advise and I understand it will be difficult to implement in active field conditions. But it is the most direct answer to your question that I have."

"A tool you keep in the sheath when you're not using it."

"Yes." Sonn looked up. "Though I would add: the sheath becomes harder to close the longer the tool stays out."

I looked at the diagrams spread between us. Agricultural work and load-bearing calculations, patterns in the land and stress in stone. People who had used this carefully and still found five more failures than they'd admitted to anyone outside their community. A seventh case that had ended not in death or breakdown, but in a quiet goodbye and forty years of correspondence with people who could no longer quite see you clearly.

"One more thing," I said. "You keep telling me what it cannot do in terms of what the Thuvari never pushed it to do. What it cannot give me because no Thuvari ever needed that capability." I put my hand flat on the surface near the diagram, not touching the documents. "What does your inference tell you it can do, that nobody has documentation on yet, given my score and my particular neural architecture."

Sonn was quiet for what felt like thirty seconds. The drive hummed.

"The predictive extension," they said finally. "Pattern recognition that has developed far enough, in a compatibility case with your neurological characteristics, may extend past current-moment positional awareness into short-range intent modeling. Reading not just where someone is but where they are going, before they go there." The long fingers pressed flat on the surface. "It is theoretically possible. The Thuvari system models predict it at approximately the eighty to eighty-five compatibility band. You are above that band. Whether the theoretical holds in practice, and what the cost of operating it is, I cannot tell you. It has not been observed."

"But if it develops."

"If it develops, you will perceive the immediate future of anyone in your pattern recognition range, in physical terms, by reading the pre-movement signals their body generates before conscious intent resolves into action." Sonn held my eyes. "Two to four seconds. Perhaps more. Depending on the development level."

The research bay was very quiet.

"That's a significant capability," I said.

"Yes."

"And the cost."

"Unknown. The cost for the basic pattern recognition, documented, is the system running continuously even when you want rest. The cost for the extended capability, at theoretical stages: the records stop before they get there."

"You're telling me the maps run out before we get where we're going," I said, which was what I'd said through the comm screen two days ago.

Sonn looked at me steadily. "And I am telling you we will build the new ones together, and I will not hide what they show. Whatever they show."

It was the most personal thing they'd said in two hours. I recognized it for what it was: not a promise of good outcomes, but a promise of honest ones. In the specific circumstances I was in, that was worth more.

"Then let's start building," I said.

Sonn nodded, a slow deliberate incline, and reached to the side of the work surface for a covered container I hadn't paid attention to yet. They set it between us and did not open it.

"The device is in here," they said. "I will not show it to you today. I want you to sit with what we have discussed first, ask me anything else that has come up, and come back in the morning. The integration should begin when you have a full night's rest and no unresolved questions." A pause. "Not because unresolved questions are a problem, but because the questions that remain after sleep are the important ones. The others resolve themselves."

I looked at the covered container, which was approximately the size of a shoebox, made of a material that was neither metal nor plastic but something between them, sealed with a latch that had no visible mechanism. Not a safe or a case, more like a carrying vessel. Something built for the thing inside to travel in safely, not to be locked away from.

"Fair enough," I said.

"Is there anything else you need to discuss today."

I thought about the seven failures versus two. I thought about the subject who had spent forty years in voluntary separation, corresponding with people who couldn't quite see them anymore. I thought about what Sonn had said about the sheath becoming harder to close the longer the tool stayed out.

"The Dominion soldiers," I said. "The ones who have been using this for years under sustained combat conditions. What do they look like to live alongside, for the people around them."

Sonn's hands settled flat. They did not look away. "The accounts the Compact has from prisoners of war and from Dominion defectors describe the long-term integration cases as effective soldiers who have become, over time, difficult to be close to." A measured breath. "Not because they are cruel. Because the asymmetry of perception that I described makes ordinary relationship difficult. They read people accurately. They are not read accurately in return. After years, this tends to produce either withdrawal into isolation or an overcorrection toward control." Sonn's voice stayed even. "The Dominion manages this by keeping their high-development cases in specialist roles with limited interpersonal contact during non-operational periods. The integration becomes, for those individuals, the primary relational experience they can access without asymmetry. Because when they are fighting, the people they are reading are also trying to read them, and the mismatch is the point."

The combat environment as the place where the asymmetry finally made sense. Where being un-readable was an advantage rather than an isolation.

I stood up from the chair, working out the same familiar knot in my right shoulder that had been there since the sedative field. "I'll be back in the morning."

"I will be here." Sonn began to rearrange the documents on the surface, not hurriedly, just settling things into their proper order. "Sleep well, if you can."

"I'll manage," I said, which was not agreement.

At the bay door I stopped. "Sonn. The seventh case. The one who chose to go. Did they know what they were choosing before the integration was complete, or did they find out after."

Sonn looked up from the documents. The dark eyes held mine across the length of the research bay.

"After," they said.

I nodded and went out into the corridor.

The ship's hum was the same as it had been all morning, the same recycled air and subsonic drive frequency. Everything around me was identical to what it had been two hours ago. I walked back toward the bunk compartment and passed the common area, where two Solvari crew members were eating something from sealed containers and talking in their own language, and they tracked me with peripheral attention without stopping their conversation. Tenvar was not among them. I noted the absence.

Back in the bunk compartment I sat down on the pad and looked at the flat gray wall across from me and did not pick up the reading device.

The system would make me more efficiently myself. That was the honest version of it, stripped of the comfort Vasht had layered onto it and the careful hedging the Compact's records used. Whatever circuits I ran most effectively under pressure, those were what it was going to amplify. All the pattern recognition and spatial awareness that I'd spent twenty-two years building on forty acres and in Cocke County's broken terrain and in the hours of stillness that hunting requires, all of that was input, and the device in Sonn's covered container was going to read it and build on it and the shape of what I became after that was going to be recognizably mine.

I was either comforted or threatened by that, and I hadn't decided which.

What I had decided: Sonn was afraid of the material, as I'd told Vasht. But they weren't afraid of it carelessly. They had looked at seven case files instead of two and had carried that discrepancy for long enough that disclosing it to me required something. They had steered every capability toward its social cost because the social cost was what they understood most clearly, and they had told me the truth about the limits of their own maps with the particular honesty of someone who has seen what happens when those limits aren't disclosed.

I trusted precision. I distrusted warmth. Sonn had given me precision without warmth for two hours and had managed, in the last three sentences, to give me something that was both.

I lay back on the pad and looked at the ceiling and thought, practically, about the covered container in the research bay and what it held. An artifact from a species that didn't exist anymore, warm to the touch according to Sonn's advance description, responsive to pressure in ways metal shouldn't be. Something built to interface with biology. Something that would, tomorrow morning, start asking questions of mine.

Whatever I answered, it would work with.

My shoulder ached. The drive hummed. The recycling system counted intervals I couldn't hear but could feel in the air's quality.

The seventh case had chosen to go. But they had also corresponded for forty years, and the records noted they found useful work. That was something. Not enough to build a decision on, but something.

I was still turning it over when sleep came, and the last thing I thought before it took me was: in the morning, ask Sonn what useful work looked like when the maps had run out.