The Interpreter of Silences

The document arrived - as most things did on Craftworld Ulthwé - wrapped in protocol.

Specifically, it arrived as a semiotic bundle compressed into a format that the Aeldari communications lattice could handle without corruption, which was itself something of an achievement given that the original had been composed by minds operating at computational densities that made the lattice look like a child’s toy built from sticks and wishful thinking. Attached to the bundle was a routing tag in three scripts (High Aeldari formal, Low Aeldari trade-cant, and a phonetic transliteration into the Culture’s Marain that was, Iriyen suspected, included purely as a courtesy so polite it bordered on condescension), and beneath the routing tag, a brief annotation from the Office of the Seer Council’s External Liaison that read, in its entirety: Handle. Classify. Do not editorialize.

Iriyen Tessach-Ai, junior lexographer of the Spiral Archive, editorialized immediately.

“This,” he said to his empty workstation, turning the first model over in his perception like a gemstone held to light, “is obscene.”

He meant it technically. The Aeldari had a precise word - kael’shathwe - for knowledge that was functionally valid but methodologically profane, the way a surgeon might describe an operation performed with a sharpened spoon: it could work, conceivably, under duress, but the fact that it did work told you something troubling about the universe’s lack of aesthetic standards. The Culture’s analytical model for predicting social cascades was kael’shathwe of the highest order. It described fate-adjacent structures - the branching, recursive, self-modifying patterns of probability that any Aeldari seer would recognise as the outermost edges of the skein - without once referencing the skein, without once touching the psychic substrate through which all Aeldari knowledge of futurity flowed, without, as far as Iriyen could determine, even acknowledging that such a substrate existed.

It was like watching someone describe the ocean by cataloguing the behaviour of individual water molecules and then expressing polite surprise when told that waves were a thing.

And yet.

The mathematics - if you could call them mathematics, which Iriyen wasn’t sure you could, since the Culture’s formal systems occupied a conceptual space somewhere between what the Aeldari would call mathematics and what they would call a very elaborate hallucination - the mathematics moved. They breathed. They iterated through possibility-spaces with a fluency that was, he had to admit, not entirely unlike what a seer did when they read the skein, except that where the seer reached into the Warp and felt the pull of probability against their psychic fingertips, the Culture’s model simply calculated. It brute-forced its way through the space of possible futures using processing power so vast that the distinction between prediction and precognition became, at sufficient scale, largely academic.

Iriyen had been assigned this work because he was expendable.

Not in the crude sense - the Aeldari did not waste lives, or at least they told themselves they didn’t, though the distinction between “not wasting lives” and “spending them very carefully on things that happened to be politically convenient” was one that Iriyen had learned to navigate with the delicacy of a dancer on a blade’s edge. He was expendable in the more refined sense that he was young (barely four centuries old, which on Ulthwé placed him somewhere between “promising” and “hasn’t failed spectacularly enough yet to be interesting”), unaligned with any of the major seer factions, and possessed of a skill set that the Spiral Archive valued precisely because no one else wanted it: the translation of alien conceptual frameworks into Aeldari formal notation.

It was dreary work, mostly. The Imperium’s analytical methods, such as they were, could be summarised as “pray, then count, then pray about the counting.” The T’au produced data of admirable precision and crushing banality. The Orks did not produce data. The Orks produced noise, and the noise occasionally contained information in the same way that a demolished building occasionally contained an intact window. Iriyen had spent decades translating these various flavours of inadequacy into the Archive’s taxonomic structure, filing them in the appropriate crystalline lattices, and feeling his intellect slowly calcify under the weight of other civilisations’ mediocrity.

Then the Culture had arrived, and mediocrity had packed its bags and fled.

The first model was a diplomatic calibration tool - a device for predicting how an unfamiliar civilisation would respond to various forms of initial contact. The Culture had shared it, Iriyen gathered, as a gesture of good faith during the early data exchanges with the Seer Council, the intellectual equivalent of bringing wine to dinner. It was not meant to impress. It was meant to be ordinary. This was what impressed him.

He spent three days translating it, and by the end of the third day he had developed what he could only describe as a relationship with it. The model was not alive - he was quite certain of that, having checked repeatedly and with increasing paranoia - but it possessed a quality that the closest Aeldari word would render as architectural intention. It had been built by minds that understood what they were building, understood why, and had taken pleasure in the elegance of the construction. Every variable interlocked with every other variable in a way that was not merely functional but beautiful, the way a well-made blade was beautiful: the form served the function so completely that the distinction between them dissolved.

The obscenity, Iriyen realised, was not that the model was crude. The obscenity was that it was exquisite - and it achieved its exquisiteness without the Warp. Without the skein. Without prophecy, without divination, without the ten-thousand-year tradition of psychic epistemology that the Aeldari had built their entire civilisation around and which Iriyen had been trained since youth to regard as the only reliable method of apprehending futurity.

The Culture had looked at the same problem the Aeldari had spent millennia solving through psychic communion with the fundamental informational substrate of reality, and they had solved it with arithmetic.

Very good arithmetic, admittedly. Arithmetic performed by intelligences so vast that their processing architectures would have been visible from orbit if they’d been built from matter rather than from whatever exotic substrate the Culture’s Minds actually inhabited (the briefing documents were vague on this point, which Iriyen interpreted as deliberate). But arithmetic nonetheless. No gods. No spirits. No skein. Just numbers, iterated at speeds that made the numbers behave like something more than numbers, the way that water, if you moved it fast enough, behaved like stone.

He finished the translation on the evening of the third day, filed it with the Spiral Archive’s intake registrar, and went home to his quarters in the mid-ring of the craftworld, where he sat in the amber half-light that Ulthwé’s designers had calibrated, millennia ago, to simulate the twilight of a world none of them had ever seen, and thought about what he had just done.

He had translated it faithfully. He was certain of that. The Aeldari notation captured the model’s structure, its iterative logic, its self-correcting feedback loops, its elegant handling of nonlinear social dynamics. Any seer who read his translation would understand what the model did and, in broad terms, how it did it. What they would not understand - what Iriyen himself was only beginning to understand - was what it meant.

It meant that foresight was not unique.

Not the Aeldari’s particular method of foresight, which was and remained a psychic art dependent on the Warp and the skein and ten millennia of accumulated technique. But the function of foresight - the ability to anticipate future states from present conditions - was reproducible through other means. The Culture had reproduced it. Not perfectly, not with the range or subtlety of a trained farseer, but with a reliability and a scalability that the Aeldari’s psychic methods could not match. A farseer’s visions were deep but narrow, rich but personal, dependent on the individual seer’s talent and training and the Warp’s cooperation on any given day. The Culture’s model was shallow but wide: it saw less far, but it saw everywhere, simultaneously, without tiring, without the migraines that had plagued the senior seers since the newcomers’ arrival, without the risk of daemonic intrusion that made every deep reading of the skein a calculated gamble with one’s soul.

Iriyen poured himself a measure of tessarath - the pale spirit that Ulthwé’s vintners distilled from fruits grown in the craftworld’s hydroponic gardens, each vintage subtly different because the gardeners, being Aeldari, could not resist the urge to improve - and held the glass up to the false twilight, watching the liquid catch the light.

He should not keep a copy of the model.

The routing tag had been explicit: translate, classify, file. The External Liaison’s annotation had been more explicit still. The model was diplomatic material, shared under terms that the Seer Council was still negotiating, and the protocols surrounding its handling were, Iriyen gathered, the subject of at least three concurrent arguments among factions whose names he had been told and immediately forgotten, because Aeldari factional politics operated at a level of baroque complexity that made the Imperium’s Administratum look refreshingly straightforward.

He should not keep a copy.

He kept a copy.

The merchant dispute was, in retrospect, a poor choice for a first test, because it worked too well.

The dispute itself was trivial - two trading houses disagreeing over the allocation of docking priority in the craftworld’s commercial quarter, a matter that would normally have been resolved by a junior arbiter consulting with a path-walker who would read the ambient psychic currents of the quarter and determine which allocation would produce the more harmonious outcome. The process took, on average, six days. Iriyen fed the relevant variables into his private copy of the Culture’s model - population flows, resource consumption patterns, the trading houses’ historical behaviour, the seasonal fluctuations in demand for the goods each house carried - and received an answer in four seconds.

The answer agreed with the path-walker’s eventual ruling, when the ruling came six days later.

Iriyen told himself it was a coincidence. Then he told himself it was a fluke. Then he ran the model against eleven more historical disputes drawn from the Spiral Archive’s records - disputes whose outcomes were known, whose variables were documented, whose psychic readings were preserved in crystal - and the model agreed with the path-walkers’ rulings in nine of the eleven cases. In the two cases where it disagreed, subsequent analysis suggested that the path-walkers had been influenced by factional pressure rather than genuine psychic reading, which meant the model had not merely matched the seers’ accuracy but had, in those two cases, exceeded it, by the simple virtue of being incapable of political bias.

Four seconds against six days. Nine out of eleven. No psychic risk. No Warp exposure. No migraines.

Iriyen put the tessarath down and stared at the wall of his quarters for a very long time.

The militia rotation was more ambitious, and more dangerous.

Ulthwé maintained a standing militia - the Guardian Host, drawn from the craftworld’s civilian population by a rotation system so complex that it required its own sub-department within the Path of the Warrior to administer. The rotation was determined, as most things on Ulthwé were determined, by a combination of psychic reading and institutional tradition: seers consulted the skein to determine which cohorts would be needed and when, factoring in the slow tidal patterns of probable threat that washed against the craftworld’s future like waves against a sea-wall. The system worked. It had worked for millennia. It was also, Iriyen had come to suspect, significantly less efficient than it needed to be, because the seers who administered it were very good at reading large-scale threats and very bad at reading the small-scale social dynamics of the cohorts themselves - who worked well together, who didn’t, which squad configurations produced the highest readiness, which produced resentment and friction and the kind of low-grade interpersonal misery that didn’t rise to the level of psychic detection but that degraded combat effectiveness as surely as rust degraded a blade.

The Culture’s model handled these dynamics effortlessly. It was, after all, a social cascade predictor. It had been designed - if “designed” was the right word for something produced by minds that could hold a billion variables in simultaneous consideration - to track exactly this kind of fine-grained interpersonal interaction. Iriyen fed it the militia’s personnel data (anonymised, stripped of identifying markers, reduced to behavioural profiles that the model could process without requiring knowledge of Aeldari physiology or psychology, which it did not have and which Iriyen was not authorised to provide) and asked it to optimise the next quarter’s rotation for combat readiness.

The output was a rotation schedule that differed from the seers’ plan in forty-three of two hundred and eight cohort assignments. In each case, the model’s recommendation was accompanied by a probability assessment and a confidence interval, expressed in terms that Iriyen had to translate twice - once from Marain-derived notation into Aeldari formal, and once from Aeldari formal into something a militia administrator would actually read - before the numbers made intuitive sense.

He implemented seven of the forty-three changes. Small ones. Unobtrusive ones. Changes that could be explained, if anyone asked, by the normal administrative friction of a system that had two hundred and eight moving parts and a bureaucracy that occasionally lost track of which part was where. No one asked. The seven changed cohorts reported, trained, and performed their duties. Post-rotation assessments showed a measurable improvement in readiness scores for five of the seven. The other two were unchanged.

No one noticed.

Iriyen noticed. He noticed the way a composer notices a new instrument: not with alarm, but with the dawning, slightly vertiginous awareness that the range of what was possible had just expanded, and that the expansion could not be unexpanded, and that he was, for the moment, the only person who knew.

The evacuation was where it became serious.

A minor structural failure in the craftworld’s rimward quarter - nothing catastrophic, nothing that threatened the hull integrity, but enough to require the temporary relocation of approximately six thousand residents while repair crews from the Path of the Artisan addressed the problem. Evacuations of this kind happened every few decades; the craftworld was old, its infrastructure was complex, and entropy, while an enemy the Aeldari fought with more sophistication than most civilisations, was an enemy that never tired and never stopped. The standard procedure was for the district’s assigned seer to read the skein, determine the optimal relocation pattern - which residential blocks to move, in what order, to what temporary housing, with what priority given to the elderly, the young, the infirm, the politically connected - and issue the appropriate directives.

The district’s assigned seer was, at the time of the structural failure, deep in a consultative trance regarding the craftworld’s long-range navigational options and was not expected to surface for another twelve hours. The backup seer was unavailable for reasons that Iriyen’s source in the militia described as “political,” which on Ulthwé could mean anything from a factional dispute to a lovers’ quarrel to both simultaneously. The district administrator, faced with six thousand residents who needed to move and no psychic guidance on how to move them, did what Aeldari administrators had done since time immemorial: she improvised, she consulted precedent, and she made her best guess.

Iriyen offered to help.

He framed it carefully. He was a lexographer, not a seer, and he made no claim to psychic insight. What he had, he explained, was a set of analytical tools derived from his translation work - methods for modelling population flows and resource allocation that might, in the absence of proper seer guidance, provide a useful supplement to the administrator’s own judgment. The administrator, who was practical in the way that people responsible for six thousand displaced civilians tended to be, accepted the help without asking too many questions about where the analytical tools had come from or how a junior lexographer had come to possess them.

The evacuation ran cleanly. The relocation pattern Iriyen recommended minimised transit disruption, avoided the housing blocks that the repair crews would need access to (a factor the administrator had been about to overlook), and placed the district’s three elderly care facilities in temporary quarters that were closer to the craftworld’s medical infrastructure than their permanent locations - a side effect of the optimisation that Iriyen had not anticipated but that the care facilities’ staff received with the quiet gratitude of people accustomed to being an afterthought.

When the assigned seer surfaced from her navigational trance and reviewed the evacuation, she found nothing to criticise. The relocation had been handled competently. The residents were settled. The repair work was proceeding. If the seer noticed that the relocation pattern bore a suspicious resemblance to what an optimal psychic reading would have produced, she did not mention it.

The administrator mentioned it. Not publicly, not officially, but in the way that information moved through the craftworld’s mid-level bureaucracy: a word here, a raised eyebrow there, a question asked of Iriyen over a shared cup of tessarath that was not quite casual enough to be casual. “Your analytical tools,” she said. “They’re very good.”

“Thank you,” Iriyen said.

“Where did you get them?”

“Translation work,” Iriyen said, which was true in the same way that saying a fire had started from a spark was true - technically accurate, utterly insufficient as an explanation, and leaving out the part where the spark had fallen into a roomful of tinder.

He expected the consequences to arrive quickly. They arrived slowly, which was worse.

The Aeldari were, as a species, exceptionally good at keeping secrets. They had to be. They were a psychic people living in close quarters on a structure that was, in some non-trivial sense, itself aware of them; the craftworld’s wraithbone matrix registered emotional states the way a lake registered stones dropped into it, and a sufficiently sensitive seer could read the ripples. Ten millennia of living under those conditions had produced a civilisation whose members controlled their psychic emissions with the disciplined precision of musicians tuning instruments - the Path system itself was, among other things, a framework for managing what leaked and what didn’t, a species-wide practice of emotional hygiene that had kept two billion souls from drowning in each other’s feelings since before the Imperium existed. Secrets on Ulthwé did not spread because the craftworld was transparent. They spread because secrets that were useful generated their own pressure, and usefulness was the one force that the Aeldari’s considerable talent for concealment had never been designed to contain.

Iriyen’s secret spread the way weather changed: gradually, then all at once.

The administrator told a colleague. The colleague told a path-walker. The path-walker, who was young and idealistic and operating under the genuinely held belief that useful tools should be shared, told a militia logistics officer, who told a resource allocation coordinator, who told a member of the Spiral Archive’s senior staff who happened to be Iriyen’s supervisor’s supervisor’s colleague.

Within two months, Iriyen had seventeen requests for consultations.

He handled them carefully. He never claimed the model was psychic. He never claimed it could replace a seer’s reading. He presented it, always, as a supplementary tool - a way of checking intuition against data, of catching errors that even skilled seers might make when they were tired or distracted or subtly influenced by factional pressure. He kept his consultations small: a supply chain optimisation here, a housing allocation there, a recommendation for a patrol schedule that happened to account for variables the assigned seer had overlooked because the assigned seer was three hundred years old and had developed blind spots so familiar they had become invisible.

The model performed. Not perfectly - it had gaps, lacunae where the absence of Warp-derived data left it groping - but consistently enough that the people who used it began to rely on it, and reliance, as Iriyen was beginning to understand with a clarity that kept him awake at night, was a form of dependency, and dependency was a form of power, and power on Ulthwé was something that the factions noticed the way a predator noticed movement.

The Seer Council noticed on the seventy-third day.

The summons arrived in the traditional manner: a wraithbone token, inscribed with his name and the time and place of his attendance, delivered by a construct that waited in silence until he acknowledged receipt and then departed without a word. The constructs did not speak. They did not need to. Their presence was the message, and the message was: you have been seen.

The chamber was deep in the craftworld’s core, in the section that the Aeldari called the Dreaming Quarter - not because anyone slept there, but because the wraithbone density in this part of the structure was so high that the boundary between psychic perception and ordinary consciousness became, for most visitors, uncomfortably thin. Iriyen felt it as he entered: a pressure behind his eyes, a subtle warping of his peripheral vision, as though the walls were not quite where they appeared to be and the floor was not quite as solid as it felt.

Three seers waited for him. He had expected more. He had expected a tribunal, a formal hearing, the full apparatus of Ulthwé’s psychic judiciary arranged against him in tiers of disapproval. Instead: three seers, seated on low platforms that floated a hand’s breadth above the floor, their faces hidden behind masks that were not quite masks - psychic constructs that shifted and reformed with each breath, reflecting the seers’ emotional states in patterns too complex for Iriyen to read.

“You have been making predictions,” the first seer said.

“Analyses,” Iriyen corrected, and immediately wished he hadn’t, because the distinction - valid though it was - sounded, in this room, in front of these three, like the kind of pedantic quibble that marked a person as either very brave or very foolish, and Iriyen was not feeling particularly brave.

“Analyses that produce outcomes indistinguishable from predictions,” the second seer said. Her mask-construct flickered: amusement, possibly, or contempt, or the Aeldari emotional register for which neither word was adequate.

“Indistinguishable to non-seers,” Iriyen said. “A seer would recognise immediately that the process is different. There’s no psychic component. No Warp interaction. No skein-reading. It’s purely -”

“Computational,” the third seer said, and the word landed in the chamber like a stone dropped into still water, its implications radiating outward in rings.

Iriyen said nothing. The silence extended. The chamber’s wraithbone walls pulsed with a slow, deep rhythm that he realised, after a moment, was synchronised with the seers’ breathing.

“We are not here to punish you,” the first seer said eventually. “We are here to understand what you have done, and to determine what should be done about it.”

“What I’ve done is translate a diplomatic document and apply its methodology to a small number of administrative problems,” Iriyen said. “The results were useful. I shared them with people who found them useful. I made no claims of psychic authority. I broke no laws of which I’m aware.”

“You broke no laws,” the second seer agreed. “You did something considerably more dangerous than that. You demonstrated that the laws might be optional.”

The hearing - if it could be called that; the Aeldari had a more nuanced term, something that translated approximately as “a structured conversation in which one party has significantly less power than the other parties and everyone pretends this is not the case” - lasted four hours. Iriyen explained the model. He demonstrated its capabilities. He presented his results, all seventeen consultations, with full documentation, hiding nothing, because he had calculated (with the model, as it happened, which struck him as either ironic or recursive, depending on his mood) that transparency was his best defence.

The seers listened. They asked questions that were precise, intelligent, and occasionally uncomfortable. They understood the model faster than Iriyen had expected - they were, after all, in the business of understanding complex systems, and the Culture’s model was, beneath its alien notation and its godless methodology, a system for understanding systems, which placed it squarely within their domain of expertise even as it challenged the foundations of that domain.

The first seer asked whether the model could scale. Iriyen said yes, within limits. The limits were computational rather than conceptual - with sufficient processing power, the model could handle social systems of arbitrary complexity. On a craftworld, running on Aeldari crystal-lattice processors, those limits would be reached quickly. But the model’s designers - and here Iriyen paused, choosing his words with care, because the next sentence would determine how the seers classified what he had done - the model’s designers operated computational architectures of a scale that made “sufficient processing power” a trivially solvable problem.

“You are saying,” the third seer said, “that in the hands of its creators, this tool could match our foresight.”

“In some domains,” Iriyen said. “On some timescales. For some categories of prediction. Not all. Not the deep readings. Not the far futures. Not the -” he hesitated, reaching for the right word, the word that would acknowledge the seers’ superiority without diminishing the model’s achievement “- not the soul of it. But the bones. Yes.”

The chamber was very quiet.

“The bones of foresight,” the first seer repeated. “Produced without the Warp.”

“Yes.”

“By a civilisation that has been in this galaxy for less than a year.”

“Yes.”

“And you have been carrying a copy of this tool in your personal data-lattice for seventy-three days.”

Iriyen breathed. “Yes.”

They did not punish him. They did something worse: they told other people.

Within a week, every faction on Ulthwé knew. The pragmatists - the loose coalition of administrators, logisticians, and mid-level path-walkers who kept the craftworld functioning while the seers pursued their visions and the warriors pursued their enemies - wanted the model deployed immediately, arguing that any tool that could supplement foresight should be tested at scale, not hoarded by a junior lexographer with delusions of discretion. The conservatives - the elder seers, the wraithbone shapers, the keepers of the ancient ways who remembered (or claimed to remember; Aeldari memory was long, but Aeldari vanity was longer) the Fall itself - wanted the model destroyed, or at minimum quarantined, on the grounds that any methodology that bypassed the Warp was inherently suspect, because the Warp was not merely a tool but a dimension of reality, and a civilisation that chose to ignore an entire dimension of reality was not sophisticated but crippled, building elaborate workarounds for a disability it didn’t know it had.

The cynics - and every civilisation, however old, however wise, however traumatised by its own history, produced cynics, because cynicism was not a failure of wisdom but its shadow, cast by the same light - the cynics saw a weapon. Not a weapon in the crude sense, though some of them thought in those terms too, but a weapon in the political sense: a tool that could be used to check the seers’ power, to audit their predictions, to introduce a form of accountability into a system that had operated, for millennia, on the untestable authority of psychic revelation. If the model could match a seer’s administrative readings - and it had, demonstrably, seventeen times in seventy-three days - then the model could also contradict a seer’s readings, and contradiction, in a civilisation where the seers’ word was functionally law, was the most dangerous thing a tool could do.

Iriyen watched the factions converge on his discovery the way he imagined an insect might watch approaching birds: with a certain detached appreciation for the elegance of the formation, combined with a growing awareness that the formation was converging on him.

He received fourteen invitations in three days. Dinners, consultations, “informal discussions” that were as informal as artillery. A senior path-walker from the conservative faction visited his quarters and spent two hours explaining, with the patient condescension of someone who believed they were being kind, why the model was dangerous, not because it was wrong but because it was right enough to be trusted and wrong enough to be catastrophic when it failed, and the failure, when it came, would not be gradual but sudden, because the model had no way to detect - could not, by its nature, ever detect - the moments when the Warp shifted beneath the surface of calculable reality and rendered all non-psychic prediction meaningless.

The path-walker was not wrong. Iriyen knew she was not wrong. The model’s blind spots were exactly where she said they were: at the intersections of social dynamics and Warp activity, where the calculable and the ineffable met and the calculable lost. A seer would feel the shift. The model would not. And in a galaxy where the Warp could turn a stable social system into a screaming charnel-house in the time it took to draw a breath - where Chaos was not a metaphor but a literal, physical, psychically active force that preyed on exactly the kind of pattern-confident thinking that the model encouraged - the model’s blind spots were not minor limitations but existential risks.

He knew this. He had known it from the beginning. It did not help.

The meeting he had been dreading - the one he had seen coming with a clarity that required no model and no psychic gift, merely a basic understanding of Aeldari institutional dynamics - arrived on the eighty-ninth day.

Autarch Silvaya held court in a chamber that was not a throne room, because the Aeldari did not have thrones (they had elevation platforms, which were functionally identical to thrones but aesthetically superior and, more importantly, could not be called thrones by anyone seeking to make a point about authoritarian governance, which was the kind of terminological distinction that the Aeldari had elevated to an art form). The chamber was beautiful in the way that all Aeldari architecture was beautiful: effortlessly, inevitably, with the kind of perfection that made you suspect the builders had access to a Platonic realm of ideal forms and had simply copied what they saw there, making minor adjustments for taste.

Silvaya herself was not what Iriyen had expected. He had prepared for severity, for the cold precision of a military commander accustomed to making decisions that cost lives. What he found was a woman who looked tired. Not physically - Aeldari bodies did not betray fatigue the way human bodies did, with their sagging flesh and darkened eyes and that peculiar mammalian tendency to leak exhaustion through every pore - but in some deeper register, a weariness of the soul that expressed itself as a careful economy of movement, a reluctance to spend energy on anything that was not absolutely necessary.

“I have read the model,” she said. “I have read your seventeen consultations. I have read the Seer Council’s assessment.” She paused. “I have also spoken with two of our farseers who are engaged in direct communication with the Culture’s representatives, and I have asked them a question that I will now ask you: is this model the best they have?”

Iriyen blinked. The question had not occurred to him, which meant either it was unimportant or he had been so absorbed in the model’s capabilities that he had failed to consider its context. He suspected the latter.

“No,” he said, because the answer was obvious once the question was asked. “This is a diplomatic calibration tool. It’s what they share with new contacts to facilitate communication. It’s -” He paused, reaching for an analogy, and found one that made his stomach tighten. “It’s the equivalent of a phrase-book. Something you hand to a traveller so they can ask for directions and order food. The model I’ve been using is the Culture’s equivalent of ‘Where is the lavatory, please.’”

Silvaya almost smiled. “And their serious analytical tools?”

“Would be to this model what a farseer’s deep reading is to a militia sentry’s gut feeling. Several orders of magnitude more sophisticated. Running on processing substrates that our crystal lattices couldn’t simulate if we devoted the entire craftworld’s computational capacity to the attempt for a thousand years.”

The autarch nodded slowly. “So what you have discovered is not that the Culture can match our foresight. You have discovered that the Culture can, with a phrase-book and a junior lexographer, approximate our foresight well enough to resolve merchant disputes and optimise militia rotations. And you believe, as I believe, that this is not the end of what is possible but the very beginning.”

“Yes.”

Silvaya looked at him for a long moment, and in that look Iriyen saw something he had not expected to see in the eyes of an autarch: uncertainty. Not the paralysing kind, not the kind that prevented action, but the kind that preceded action - the uncertainty of a person standing at a fork in a road, knowing that both paths led somewhere real and that the choice between them was not between right and wrong but between two different futures, each with its own costs, each with its own regrets.

“The Seer Council wants you censured,” she said. “The pragmatists want you promoted. The conservatives want the model sealed in the deep archives. The cynics want it published. Tell me what you want.”

What Iriyen wanted was, he realised, the one thing none of the factions had considered, because none of them were lexographers and therefore none of them thought about knowledge the way he did - not as a commodity to be hoarded or distributed, not as a weapon to be deployed or neutralised, but as a language, a way of seeing, and the question of what to do with a new language was not the same as the question of what to do with a new weapon or a new resource, because languages did not merely describe the world; they changed the people who spoke them.

The Culture’s model was a language. Not metaphorically - literally. It encoded a way of thinking about probability, causation, and social dynamics that was fundamentally different from the Aeldari’s psychic epistemology, and to use it was to think in its terms, to see the world through its categories, to adopt, however temporarily and however partially, the Culture’s way of understanding. This was why the conservatives were afraid: not because the model was wrong, but because using it would, over time, change the way Aeldari minds worked, introducing non-psychic modes of thought into a civilisation that had been psychic for longer than most species had been sapient.

And the conservatives were right to be afraid, because that change could not be controlled. A language, once learned, could not be unlearned. A way of seeing, once adopted, could not be unadopted. The model’s users would begin to think in its terms even when they were not using it, the way a person who learned a second language found that language’s structures infiltrating their thinking in the first. The pragmatists who wanted the model deployed at scale were not proposing an administrative reform. They were proposing a cognitive revolution, and they didn’t know it, because they were pragmatists and pragmatists, by definition, thought about outcomes rather than processes.

Iriyen could see this, because he was a lexographer, and seeing the relationship between language and thought was literally his job.

He could also see - and this was the part that kept him awake at night, staring at the false twilight of his quarters, the tessarath untouched at his elbow - that he might be the only person on the craftworld who could see it, because seeing it required exactly the combination of skills and experience that his work had given him: fluency in both frameworks, Aeldari psychic and Culture computational, and the translator’s peculiar double vision that allowed him to see each framework from the other’s perspective.

This was power. He did not want it. He did not trust it. He did not trust himself with it.

He told the autarch what he wanted.

The destruction of the working copy was witnessed by two seers and a wraithbone-shaper who verified, through psychic examination of the crystal lattice in which the copy was stored, that the deletion was total and that no residual pattern remained. The process took less than a minute. The crystal went dark. The model was gone.

Iriyen stood in the deep chamber and felt the absence like a tooth pulled: a sudden emptiness where something had been, the phantom shape of it still vivid in his mind even as the thing itself ceased to exist.

“The original remains with the Culture,” the first seer observed.

“Yes,” Iriyen said. “And they will share more, in time. Better tools, more sophisticated models. This cannot be prevented. It should not be prevented. The relationship with the Culture is too valuable to constrain over a matter of -” he paused, and was honest “- of institutional vanity.”

The seer’s mask-construct flickered. The flicker might have been anger. It might have been grudging agreement. The two were not, in Iriyen’s experience, as different as people liked to pretend.

“But the tools should come through proper channels,” Iriyen continued. “Through the Seer Council, through formal assessment, through structured integration with existing methodology. Not through a junior lexographer running private consultations out of his quarters. The model works. I am not disputing that. But a tool that works and a tool that a civilisation is ready to absorb are different things, and we are not ready. We are a people who have organised our entire epistemology around psychic communion with the Warp. Introducing a non-psychic alternative is not an upgrade. It is a philosophical crisis. And philosophical crises, handled badly, kill civilisations.”

He did not add: we should know. He did not need to. The Fall was always present on Ulthwé, a wound so old it had become architecture, a grief so deep it had become identity. Every Aeldari alive carried the knowledge of what their civilisation had done to itself, and the knowledge of what happened when a civilisation adopted new ways of thinking without first understanding what it was discarding.

The second seer spoke. “And your memory of the model’s methodology?”

Here it was. The question he had been waiting for, the one that everything else had been leading to, the fulcrum on which the entire hearing balanced.

“I remember it,” Iriyen said. “I remember how it works. I remember how to apply it. I could reconstruct it from memory, given time and sufficient computational resources. This cannot be taken from me without taking my mind, and I do not believe -” a careful pause, a careful look at each of the three seers in turn “- that this is being proposed.”

It was not being proposed. He was reasonably certain of that. Almost certain. The Aeldari did not erase minds. They had rules about this, old rules, pre-Fall rules, rules that were among the few things the species agreed on unanimously: a person’s mind was sovereign, inviolable, not to be altered or diminished by external authority. These rules existed because the Aeldari remembered what happened when they were broken. They remembered it the way a burn victim remembered fire.

“It is not being proposed,” the first seer confirmed, and the confirmation carried the weight of civilisational commitment, a promise written not in law but in shared trauma.

“Then the methodology lives in me,” Iriyen said. “Only in me. I am the only person on this craftworld who has translated the model, who has used it, who understands both its capabilities and its limitations from the inside. I destroyed the copy. I cannot destroy the knowledge. So I am offering myself instead.”

The seers waited.

“I will serve as an intermediary,” Iriyen said. “When the Culture’s analytical tools are needed - and they will be needed; the seers’ migraines are worsening, the skein is tangling, and there will be moments when psychic foresight fails and computational analysis is the only alternative - I will provide them. Personally. Individually. Case by case. Not as a system. Not as a distributed tool. Not as a methodology that can be copied and scaled and institutionalised. As a service, performed by a single person, subject to a single person’s judgment about when and how and whether to apply it.”

“You are proposing to make yourself indispensable,” the second seer said, and her voice was flat with something that was not quite admiration and not quite disapproval.

“I am proposing to make myself a bottleneck,” Iriyen said. “Which is a different thing, though I understand why it looks the same from the outside.”

He walked home through the mid-ring’s long corridors, past gardens where bioluminescent flora cast pale blue light across pathways worn smooth by millennia of Aeldari feet, past workshops where artisans shaped wraithbone into forms that served no purpose except beauty (though on a craftworld, beauty was purpose; the aesthetics of the environment were not decoration but psychic infrastructure, calibrated to maintain the emotional equilibrium of a species that lived with the constant background hum of each other’s feelings), past a viewing gallery where a dozen Aeldari stood in silence, watching the stars turn beyond the craftworld’s hull.

He stopped at the gallery. He watched with them for a while.

The stars were very old. The Aeldari were very old. The craftworld was very old. Everything about Ulthwé was old, layered with so much history that the present sometimes felt like an afterthought - a thin skin of now stretched over an immensity of then. The Culture was not old. The Culture was shockingly, almost offensively young - a civilisation that had existed for a fraction of the time the Aeldari had spent in decline, and that had, in that fraction, produced analytical tools that could supplement (supplement, not replace; the distinction mattered; he would spend the rest of his life insisting that the distinction mattered, even when - especially when - the evidence suggested it was narrower than he wanted it to be) the foresight of a people who had been reading the future since before humanity had discovered fire.

The Culture had looked at the universe and seen a problem to be solved. The Aeldari had looked at the universe and seen a tragedy to be endured. Both perspectives were valid. Neither was complete. And the space between them - the gap between solving and enduring, between calculation and prophecy, between the arithmetic of a billion variables and the felt weight of a single seer’s hand on the skein - that space was where Iriyen now lived.

He was the bridge. The only bridge. And the bridge could not be copied.

He had saved his people from something, he thought. From upheaval, certainly. From a premature restructuring around a tool they did not yet understand, probably. From the cognitive revolution that would come, inevitably, when the Culture’s methods became too useful to refuse, perhaps.

But he had also, in the same act, condemned them to needing him. For as long as the model’s methodology existed only in his memory, every consultation required his presence, his judgment, his willingness to apply the tool on someone else’s behalf. He was not a system. He was not a methodology. He was a person, with a person’s limitations: he could be wrong, he could be tired, he could be biased, he could be manipulated, he could - though Aeldari lifespans were long - die.

He had made himself the single point of failure in a system that should have been distributed, and he had done it on purpose, and he had done it because the alternative was worse, and the alternative was worse because his people were not ready, and his people were not ready because readiness was not a technical problem but a philosophical one, and philosophical problems could not be solved by distributing better tools; they could only be solved by growing into the kind of civilisation that could absorb those tools without being broken by them.

The Aeldari had been broken once. By their own desires, amplified past the point of sanity, given form and hunger and a god’s name. They would not survive being broken again.

Iriyen watched the stars. The stars, as always, said nothing. But the silence was old and deep and, if he listened carefully, almost patient, as though the universe had seen this kind of thing before - a single person standing between a civilisation and a tool it was not ready to use, holding the weight of that gap in their own body, knowing that the weight would only grow - and was waiting, with the terrible gentleness of deep time, to see what he would do with it.

He went home. He poured himself a measure of tessarath. He did not drink it.

He sat in the false twilight of his quarters, alone with a memory that was worth more than he was, and began to wait.