The castle library occupied three connected rooms on the upper east wing and held, by my rough count, approximately four thousand volumes. Orvel the librarian, a thin distracted man who seemed to regard the library primarily as a place to sleep, had organized everything by colour of binding. History sat beside philosophy sat beside technical manuals sat beside what appeared to be someone's personal correspondence that had been bound and shelved by accident.

  I spent two weeks walking through it all, building a mental map that was more useful than the shelving. The map confirmed what I'd suspected from the first volumes I'd read; the texts were rich in description, thin on explanation. How to cultivate, not why it worked. The specific movements. The techniques that had worked before and would presumably work again. The master knew because the master's master knew, and the chain of knowing reached back far enough that the original understanding had been lost and only the practice had remained.

  My old world's history had made this pattern familiar. Every technology had its dark age of "it worked, they knew it worked, therefore they did it." Alchemy to chemistry. Herb-craft to pharmacology. "This shape of hull has always sailed well" to fluid dynamics. At nine years old I was reasonably confident I was standing at one of those moments, in a field I didn't yet know how to name, which was the sort of thought that kept me reading long past when the tutors expected the lamps to be out.

  What struck me, though, was the volume of the collection compared to its depth. Four thousand texts sounded like a serious library. For a kingdom that had existed under continuous rule for four centuries, where the same king who had commissioned the library was still alive to use it, four thousand was thin. Sethara had roughly a hundred and twenty million residents and at least a dozen active guilds. The accumulated technical writing of all those practitioners across all those years should have filled rooms. Instead it filled shelves, and most of what filled them was history written by people more interested in kings than in the builders who made those kings' ambitions physical.

  I thought about the literacy rates I'd observed during my limited trips outside the castle. Guild documents, bills of sale, trade ledgers; the craftsmen's quarter ran on written text. But writing something down for commerce and writing it down for posterity were different acts. The merchant who kept a ledger was preserving records for his own use. The practitioner who might have written a treatise on formation design had no institutional reason to do so, no library requesting the work, no expectation that her knowledge mattered beyond the transaction at hand. The system was designed to preserve what the literate classes valued, and the literate classes valued politics.

  The enchanting texts Marvenn had helped me access were better than the open library's materials. More technical. Written by practitioners rather than historians. But even the best of them shared the fundamental limitation: they described how to inscribe without explaining why inscription worked.

  The formation patterns, the specific geometric arrangements that produced specific effects, were presented as given. This configuration heated; this one strengthened. Here was one that illuminated. No attempt to explain what property of the geometry produced the effect. The patterns were correct and the effects documented, and that was considered sufficient.

  I sat with the most technical text I'd found and worked backward through the logic. My own observational data gave me something to work with: years of holding mana in air, feeling it flow through patterns I'd created and through the stone channels of the castle's existing enchantments. How it moved and pooled and responded to shape. The understanding from age seven still held, that mana followed clear channels and clarity sustained structure. And there was the carpenter's unintentional lesson, the way wood grain ran and how working with it versus against it changed what was possible.

  And I had, freshly disturbing in my ninth year, a new observation I hadn't found anywhere in the texts.

  I'd graduated from observation to occasional participation at the castle smith's by now, allowed to pump the bellows or hand materials at direction. But I'd also negotiated time in the workshop of one of the city's smiths: an old woman named Garret who worked in the craftsmen's district near the castle walls. She'd agreed to let a curious prince-child spend time in her shop in exchange for the prince-child solving a mathematics problem she'd been struggling with related to alloy ratios. When I'd arrived at her door the first time, she'd looked at the two guards, looked at me, and said, "Palace job, is it," not as a question, and then held the door open.

  The negotiation had required a city expedition, which itself required a guard escort. Two guards, because the household protocols specified it for younger princes. Kever, a professional man who had been on the castle guard rotation for eleven years, walked me through the castle gate and into the city for the first time with anything resembling purpose.

  The craftsmen's quarter announced itself by sound before sight. Metal on metal, first; the ring of a hammer on an anvil carried over the closer noises of foot traffic and cart wheels. Then the rhythmic scrape of stone-working, and under that the steady mechanical thump of a loom. The district occupied a whole section of the city, dense with workshops and shopfronts in a clustering pattern that made economic sense even to a nine-year-old: smiths near smiths, because proximity to competitors was also proximity to shared suppliers. Woodworkers adjacent, because they shared the same cart routes for raw materials. The more specialized trades branched off into side streets where the rents were presumably lower and the foot traffic thinner.

  The smell was coal smoke and hot metal and something sharp that turned out to be the lye vats of the leather workers two streets over. Not unpleasant. Industrial. It smelled like people making things, which was a category of smell I already loved.

  The foot traffic moved around us with the mild unconscious adjustment of people who have learned to read a livery. Not deference, not exactly; nobody bowed or cleared a path with ceremony. But there was a slight eddy in the flow, a few extra inches of clearance around the two palace guards, and a glance or two at the crest on their shoulder-plates that slid past me entirely. I was nine years old in an undyed linen tunic and brown wool trousers, castle-quality fabric but cut without any of the embroidery or dyed trim that would have marked me as someone worth noticing; I was not what they were looking at.

  What I noticed, because I noticed these things, was the gap between the castle and this district. The castle had formation-heated corridors, enchanted illumination on an evening schedule, stone that had been saturated with centuries of ambient mana until it practically hummed. The craftsmen's quarter had forge heat and hand-carried water. A few of the wealthier shopfronts displayed the soft glow of purchased enchanted lanterns in their windows; three weeks of a skilled craftsman's income, each of those lanterns, which told me something about the income of the shops that had them and the income of the shops that didn't.

  The guild pricing boards were posted at the quarter's main intersection. I stopped to read them while Kever waited with the particular patience of a guard who had learned to accommodate curious children; he'd already steered us to the outer edge of the intersection without comment, positioning us where we had clear sightlines and were not blocking cart traffic, which I suspected was habit rather than instruction. The boards listed minimum prices for common commissions alongside quality standards and the names of guild-certified practitioners. The Smiths' Guild board was the largest, with a pricing floor that Marvenn had already told me about: set several years earlier, ostensibly to prevent destructive competition but structured in a way that protected established smiths with existing client relationships at the expense of newer ones who might have offered lower prices for simpler work. The large buyers had seats on the guild council. The small buyers did not. The pricing floor had been set by the people who benefited from it.

  That was familiar. My old world had called it regulatory capture, and it happened wherever the regulated industry controlled the regulatory body. I didn't have the standing to say any of this to anyone, so I filed it and kept walking.

  Two streets later, the streetlamps changed. I noticed the mana first; the ambient density thinned over the space of half a block, from the faint residual hum of the castle district's formation network to something quieter and more ragged. Then the light itself shifted. The clean amber glow of enchanted fittings ended at a specific intersection, replaced past it by the warmer, unsteady orange of oil torches mounted on iron brackets.

  "Why do the formation lights stop here?" I asked Kever.

  He glanced at the last enchanted lamp, then forward at the torches. "Coverage boundary," he said. "The formation network was laid a hundred and twenty years ago. City's grown since then. Everything past Caulder Street is outside the original infrastructure plan."

  "Can it be extended?"

  "There's a petition process. Slow. Expensive." He shrugged, the kind of shrug that said he'd watched the process not happen for a long time. "The new districts make do."

  I looked back at the amber line behind us, then ahead at the orange. A hundred and twenty years ago, someone had drawn a boundary on a planning document. The city had grown past it. The boundary hadn't moved.

  Garret's workshop was different from the castle smith's. She worked with a wider range of materials, including metals I hadn't seen before; she also talked while she worked, steadily, as though she'd been alone with her craft long enough that any conversational partner would do, even a nine-year-old prince.

  Her shop was smaller than Kerren's, but denser. Tools hung from every available wall surface in an order that looked chaotic until you watched her reach for them; then you saw the logic. Frequently used tools at arm height, specialized pieces higher up, heavy stock along the floor. Sixty years of smithing had organized the space around the movements of her body until the workshop was less a room and more an extension of her hands. I'd seen the same phenomenon in Tova's carpentry shop and in the castle kitchen: the physical space shaped by decades of the same person moving through it until efficiency became architecture.

  I was working the bellows while she drew a bar of iron-silver alloy from the fire when she said: "You can feel it, can't you. When it's right."

  "The metal," she said. "When it's ready. There's a moment when it's telling you to go. You learn it over years."

  I watched her work. She struck at the precise moment I felt the mana in the bar shift, a minute change in its flow pattern. "I can feel the mana change," I said. "In the hot metal. It opens. Just before it's right."

  She paused. Set the bar back in the fire. Looked at me with eyes that had been around a long time. "Mana sensitive," she said. "More than most." Not a question. "What do you feel?"

  "In the metal?" I thought about how to describe it. "When the metal gets hot enough, the mana in it becomes less fixed. Like it has more room to move. And at a certain temperature, just before you'd usually strike, it reaches a point where it seems to be waiting for direction. Like it's asking for a pattern to follow. And then it cools again and that opportunity passes."

  A long silence during which Garret watched me with an expression I couldn't interpret.

  "I've been a smith for sixty years," she said. "I had never heard anyone describe that who wasn't at least Anchoring stage."

  Sixty years. At Awakening stage, which gave her a lifespan of roughly a hundred and twenty, she was past the midpoint of her career. In my old world, sixty years of craft would have made her a master approaching retirement. Here, she was established but not remarkable for her tenure; there were Breathing-stage smiths in the city who'd been at the anvil for a hundred and fifty years, and the guild histories mentioned Anchoring-stage masters who'd practised for centuries. The last Shaping-stage smith in Sethara had left three hundred years ago, according to the guild records, and nobody had replaced that level of capability since. The whole craft had been slowly losing altitude for centuries, the ceiling of what was possible dropping each time a high-stage practitioner died or departed without passing on what they knew.

  "I'm Awakening," I said. "Not even Breathing yet." Technically the truth in every formal sense.

  "Hm," said Garret.

  She didn't teach me anything directly. She just worked and talked, and by the end of the afternoon I understood: the moment I'd described, when hot metal's mana became receptive and open, was not just a window for striking. It was a window for channeling mana into the metal while it was still open; for guiding mana into the material's structure rather than carving channels on its surface afterward.

  "Infusion," Garret said. "That's what the old masters called it. Working the mana in with the metal, not after." She paused. "Nobody does it anymore. Not since the last Shaping-stage smith left the city three centuries ago."

  "Why does it require Shaping?" I asked. "If the mana is already open and receptive."

  "Because the kind of mana you need to guide in must be refined," she said. "Breathing-stage mana is too rough. It would contaminate the material rather than integrate. You need something clean enough to blend." She tapped her chest. "Shaping stage is when your mana becomes clean enough."

  I turned that over. The economics of it were bleak; if infusion required Shaping stage, and Shaping-stage cultivators were the rarest practitioners in the kingdom, then infused goods were functionally extinct. Any surviving infused items were irreplaceable artifacts, their value infinite in the most literal sense. You could not buy what no one alive could make.

  "But the window is there even for basic metals," I said. "Could you prime the material with something: a base inscription, something that refined the incoming mana as it entered? Use the inscription as a filter?"

  Garret's hammer, which had been moving toward her anvil, stopped.

  She looked at me for a long time.

  "That's been tried," she said finally. "Pre-inscribed base metals. It was one of the last things the Shaping-stage smiths worked on before they left. Couldn't get the filter clean enough. The inscription introduced its own mana signature and contaminated the infusion."

  My stomach dropped. "The inscription's own signature interferes with the incoming mana?"

  "That's the problem." She picked the hammer up again. "You'd need an inscription that was somehow neutral. That didn't impose its own pattern. Nobody figured that out in three hundred years of trying."

  She started hammering, and she didn't speak again for an hour, and I pumped the bellows and turned the filter contamination problem over and over, looking for an angle the old masters might have missed.

  The cultivation level I'd need was years off. Garret didn't answer when I asked anything more; she just hammered, and I pumped the bellows, and I was still working through the implications when Kever collected me at dusk and walked me back to the castle.

  Just before he'd arrived, Garret had paused her hammering long enough to say, eyes still on the iron: "Most apprentices stop asking questions once they can make clean welds. They think the asking was the apprentice part and the knowing is the master part." She'd gone back to the fire. I'd sat with that for the rest of the afternoon, turning it alongside the filter idea, equally unable to let either one go.

  If you discover this tale on Amazon, be aware that it has been stolen. Please report the violation.

  The walk back was its own education. The craftsmen's quarter gave way to wider streets as we approached the castle district, and the change was visible in the infrastructure. Enchanted streetlamps appeared, their glow steadier and cleaner than the forge light and candle smoke we'd left behind. The buildings grew more spacious; the foot traffic shifted from workers in practical clothing to officials and minor nobles in fabrics that signalled status through implication rather than declaration. The castle's warming inscription at the main gate, three hundred years old and still functional, wrapped around us as we passed through, and the temperature difference between the city's autumn chill and the castle's regulated comfort was the kind of thing you stopped noticing if you'd never left. I had left. I'd noticed.

  Back in my room that evening, I thought about Garret's workshop. The tools, the heat, the organized density of sixty years. And the loose papers on the bench near the forge.

  She'd shown me her working notes on alloy ratios during my visit, the same ratios I'd helped her calculate. They were written on loose sheets, no cover, no binding, stacked on the corner of her workbench where the forge heat reached. Two of the pages had scorch marks at the edges. A third had a dark spot where a spark or a hot filing had landed and burned through a number.

  I had cloth, leather scraps from the castle stores, and a basic understanding of heat-shedding inscriptions. Not sophisticated work. First-year theory applied to practical materials.

  It took three evenings. A folded cover of heavy linen, doubled for stiffness, with a soft leather strip across the spine to hold the fold. On the inner face of the cover, a simple inscription: a three-node heat-dispersal formation that would shed temperature passively, keeping the surface below the flash point of paper. Crude by any serious enchanter's standards. Functional by Garret's.

  I brought it on my next visit. She opened it, turned it over, ran her thumb along the inscription the way she would have checked a weld seam.

  "Practical," she said. Same tone she said everything.

  She put her notes inside it and set it on the bench, three inches from where the loose pages had been. No ceremony. No thanks beyond the word. She went back to work.

  The cover was still there on my third visit, and my fourth. The pages inside were unscorched.

  The question I asked in tutoring that created a problem came in a theory session with a substitute instructor. Feldren was ill. The substitute was a court official's relative who taught cultivation as a supplement to his primary duties and who, I could tell within minutes of his first explanation, did not particularly understand what he was teaching.

  He was covering Breathing stage: the transition from Awakening, the establishment of a mana cycle. His description was accurate enough but left out something critical: why the cycle mattered. He described the physical process (mana flowing in, circulating, flowing out) without mentioning that the purpose of circulation was refinement. The mana you pulled in was ambient and mixed; what you breathed out was somewhat purified; what built up in your body over time was increasingly dense and clean. The cycle was a purification system.

  I sat through this for twenty minutes and then I raised my hand.

  "The mana circulation," I said. "Is the cycle's purpose primarily to purify the mana, or is there another mechanism?"

  He looked at me blankly.

  "Like," I said, trying to clarify, "if you just held mana without cycling it, you'd accumulate ambient mana but it would stay ambient: mixed quality, various densities, not integrated. The cycle seems to be specifically for purification and integration. Like a filter, almost. But then at Anchoring stage you form a core which seems to be a concentrated reservoir of the already-purified mana, and the core changes how the cycle works because now it's the core that refines rather than the cycle itself."

  The substitute held up a hand. "Where did you read about this?"

  "I inferred it," I said. "From the cultivators I've observed and the cultivation texts in the library. There are descriptions of the Anchoring core formation that imply."

  "This is outside the scope of today's lesson," he said firmly. "We're covering basic Breathing technique."

  "Right, but understanding why the cycle works might help the other students learn to do it more effectively."

  "Sorrim." His voice was flat. "We will not be diverging from the curriculum."

  I stopped. Thessa, next to me, was making a small movement with her hand that I'd learned to interpret as please stop talking, you're doing the thing again.

  I stopped talking.

  The content of the question was correct, I was fairly sure. The problem was the delivery. I'd posed it as a correction, in front of a classroom, to a court instructor who hadn't asked for one.

  Nine years old, no formal training past basic level, and I'd just implied the man was teaching an incomplete version of his own subject. The fact that he was a court official's relative rather than a professional cultivation instructor made it worse, not better; his position depended on connections, and being corrected by a child in front of other children was the kind of thing that got remembered. Household politics ran on exactly these slights, small humiliations that accumulated into resentments that could last decades or centuries depending on the parties' lifespans.

  The kind of thing that made people remember you. And I was trying very hard to be forgettable.

  I went to Marvenn after class. He was forty-one, with three decades of navigating the castle's political currents, and he'd tell me the truth.

  "Too much?" I asked.

  "The substitute instructor is going to mention it to Levet," he said. "Levet will file it. And Levet files everything."

  "Is that a problem?"

  Marvenn thought about it. "For now, no," he said. "The seventeenth son asking clever questions is an interesting note in a file, not a threat to anyone. But you should be careful that it stays that way."

  "How do I make sure it stays that way?"

  "Keep the questions genuine," he said. "You're asking because you actually want to know, right? Not to show off?"

  "Yes."

  "That'll show through eventually. The show-offs get predictable. You get..." He paused, looking for the right word. "Interesting. In a different way." He paused again. "Just maybe phrase things as questions rather than corrections."

  "Right," I said. "That's fair."

  "And Sorrim?" He said my name with a particular deliberateness that meant he was about to say something careful. "Whatever you're actually doing, the cultivation and self-study and workshop time, be thoughtful about what level you perform at. There's a wide space between 'invisible' and 'noticeable' that's actually quite comfortable. You don't have to choose the extremes."

  I looked at him. "What level do you think I'm actually at?"

  He met my eyes. "I think you're doing something deliberate, and I think you're good enough at it that most people haven't noticed." A pause, and something shifted in his expression; not warmth exactly, but the careful consideration of someone calculating costs. "I'm not going to ask you to explain it. But I want you to know that if I've noticed, eventually someone less friendly will too."

  I was quiet for a moment. "Is that a warning?"

  "It's an observation," he said, in the same tone I'd used with him four years ago in that window seat, and I thought the echo was deliberate. He remembered that conversation.

  "Then no problem," I said.

  We walked back to the main corridor and went our separate ways. I thought about it all evening. Marvenn had seen through the performance and wasn't going to press it. That was more than I'd expected from anyone in this household.

  That night I lay awake longer than usual. The corridor outside my room creaked in the particular way it did when the temperature dropped and the old wood contracted. Marvenn had seen through the performance, and he wasn't going to press it. I kept turning that over, the way you turn a coin in your pocket without deciding to spend it.

  The conversation with Garret stayed in my mind for weeks, and the longer it stayed the more I found in it.

  I kept turning the infusion concept over. Working mana into material during manufacture, not carving it on the surface afterward. The difference nagged at me the way a crooked joint nagged a carpenter.

  Every enchantment I'd seen in the castle worked the same way: make the object, then inscribe channels on it. The mana lived in those channels. If the material degraded, the channels went with it, and the enchantment died. Material and mana, side by side but never truly joined.

  What Garret had described was something else. Mana worked into the grain while the metal was still open and receptive. To separate the mana from the material afterward, you'd have to destroy both. They were one thing. The difference was like paint on wood versus dye in cloth. One sat on the surface. The other was part of the material itself.

  The historical texts had a word for this: "mana-forged." It showed up in descriptions of legendary artifacts, always alongside words like "unbreakable" and "impossible to replicate." The technique had been lost when the last Shaping-stage smiths died out, centuries ago. The guild histories were frustratingly vague about the transition; one generation, Shaping-stage smiths were rare but present in major cities. The next, they were gone, and nobody had recorded where they went or why. Knowledge preserved for guild members, the texts said, as though that explained anything. But the guild that had preserved it was gone too, dissolved roughly thirty years before my birth in a dispute about licensing standards that had left both factions with nothing and the rest of the world with less.

  But the threshold fascinated me. Below a certain temperature, the metal's mana was fixed, unyielding. Above it, everything opened at once, not gradually but all at once, the way ice became water at a specific point. The mana's relationship to the metal reorganized at that critical temperature, and the window existed only during the transition.

  The basic principle? Garret had shown me the window of receptivity in ordinary iron-silver alloy. The full technique was years beyond me. A minimal version might not be.

  Garret was still in the craftsmen's district, still talking while she worked. I'd go back.

  The library reorganization project I described to myself as "in my head" began to leak into reality around age nine and a half.

  A mental reorganization only. I'd abandoned the idea of rearranging the physical shelves early as too visible. But I'd started writing a finding guide, in plain text, in one of the blank journals I'd accumulated. I already had the mental map; this was for anyone else. For anyone else who might want to use the library and was confused by the colour-based shelving system.

  Orvel had no interest in it; I hadn't told him. The library could have been far more useful than it was, and that gap bothered me the way waste always did.

  The guide took six weeks to complete, evenings and free periods, and ran to forty-three pages. It listed every book in the library by subject, cross-referenced by author where known, with brief descriptive notes about content and quality. The notes were occasionally honest in ways that a formal catalog would not be: "This text claims to cover advanced formation theory but the author does not appear to have practical experience and several techniques are described incorrectly" was one such note. And: "The historical survey of the northern territories is useful for geography but the author's political analysis should be treated skeptically."

  I shelved the guide in the library itself, at the end of the reference section, under a neutral title that indicated its purpose. I didn't tell anyone about it.

  Three months later, I noticed that the guide had clearly been used. The spine was slightly more worn, the pages showed the fingerprints of someone who had worked through it with focus. I didn't know who.

  The anonymity didn't bother me. Someone had found the right book faster because of forty-three pages I'd written; whether they knew who wrote it was beside the point. I thought about the servant who'd managed firewood routing for twenty years, retired, and left no record of how any of it worked; six months of quiet chaos had followed. There was always knowledge like that, sitting in one person's head, waiting to be lost.

  The question of cultivation theory kept compounding.

  By age nine I had worked through everything in the restricted library section and most of the relevant material in the open library. I'd supplemented with the three papers and a monograph from Tutor Giselle, and I'd had two more conversations with her that had given me additional threads to pull. The picture of what was known was filling in. What that meant was a sharper view of what wasn't.

  Two notebooks filled with questions nobody had thought to ask. The one that kept me up at night: why did mana lose coherence crossing node boundaries? Practitioner H had the best partial theory on this, and it was sitting in our restricted section, gathering dust, probably unknown to everyone who could have used it. There were others, a whole web of gaps where the formal knowledge just stopped and practitioners shrugged and said "it worked." Garret's window-of-receptivity, the filter idea: completely absent from every text I'd read.

  I was nine and couldn't answer any of it. The notebooks kept filling.

  Marvenn's access to the restricted section had expanded over time, partly through his own persistence and partly through the sheer number of tutors he'd cultivated relationships with. By our ninth year I had, through him, read substantially more technical material than anyone would have expected from a seventeenth son at Awakening stage.

  The technical material confirmed some things I'd arrived at independently and surprised me on others.

  What surprised me most was the existing theory of mana grades. I'd been developing my own understanding of mana quality: clean versus rough, organized versus ambient. The technical texts had an existing framework, not the same as mine but related, with different vocabulary and a different emphasis.

  The official framework classified mana by stage. Mine classified by quality: the specific texture and organization, regardless of stage. You could have low-quality mana at high stage or high-quality mana at low stage. The combination determined what you could do with it.

  Neither framework was complete. The synthesis didn't exist yet. And I'd stumbled into an odd advantage: I'd noticed quality effects before learning the official system, which meant my understanding was empirical rather than inherited. Most practitioners learned the framework first and then tried to fit their observations into it. I'd gone the other direction. Observation first, framework second. I'd learned in another life that starting from what you could actually see, rather than from what the theory said you should see, sometimes let you notice things the theory had been built to overlook.

  The cipher journal was filling up. I couldn't present any of this at nine years old. But somewhere between the observation notes and the material categories, a shape was forming that I couldn't quite see yet.

  I kept not transitioning to Breathing. This required justification, even to myself, so I laid out both sides: the wisdom argument was that my Awakening foundation was substantially better than standard, and transitioning would end that phase and modify the specific sensitivities I'd been developing. Some would persist; some would be overlaid by the new cycle.

  The cowardice argument was that Breathing was where cultivation ability began to compound, and staying in Awakening was just avoiding the risk of finding out I'd wasted my time.

  I tested this against my observations. The mana-material sensitivity I'd developed had practical value confirmed through years of use. It would inform every inscription I would eventually make.

  Would that sensitivity survive the Breathing transition? In principle yes. In practice, the Breathing cycle shifted how practitioners sensed ambient mana. The texts described the shift as making Breathing-stage practitioners less aware of fine ambient details and more aware of their own internal mana states.

  I wasn't sure I wanted to make the trade yet.

  This wasn't indefinite delay. I had an ending condition: when the sensitivity work produced the result I was aiming for, a complete high-quality map of how mana behaved in each major material category, I would transition. Not before, because Awakening-stage sensitivity was the tool I needed for the mapping work. After, because Breathing would open capabilities I couldn't access in Awakening.

  I gave myself two more years. Whether that was discipline or just a way to avoid finding out I'd been wrong, I still couldn't tell.

  The restricted section had one more text I hadn't found until my third systematic search: a small bound volume, hand-lettered, tucked behind a larger reference manual as if someone had placed it there and forgotten it.

  The title, written in ink that had faded significantly, was: Notes on Mana Grade Classification: Working Draft.

  Sitting on the cold floor of the restricted section closet, I read it from beginning to end.

  Incomplete. The author, identified only by initials, had documented seventeen case studies, each describing a practitioner's mana quality and work outcomes. The correlation was visible: high quality mana at lower stages consistently produced better work than lower quality mana at higher stages.

  The classification system, though, was never finished. Quality had been described in subjective terms, "refined," "clean," "organized," without operational definitions. The subjectivity made the framework useful to anyone who already had the sensitivity and useless to everyone else.

  Which was my problem too. I could feel mana quality, distinguish high from low, but had never defined the classification in a way someone else could use. Someone without my sensitivity would read the working draft and learn nothing they could apply.

  This was the second time I'd found someone else working on the same questions. The resonance work. Now the quality classification. In both cases the previous work was incomplete, abandoned before the question was resolved. In both cases the researcher had been working alone, with no one to push back on their subjectivities or force them toward rigor.

  I put the volume back in its hidden spot and sat there for a while, cross-legged on the cold floor, listening to the restricted section's silence. Two researchers, maybe more, working alone on the same questions. Both stopping before the answers came.

  What had made them stop?

  The question sat with me as I walked back through the library's main rooms, past Orvel dozing in his chair, past the shelves of colour-sorted knowledge. Had the researchers run out of interest, or out of institutional support? Had they been working in the margins, the way I was, and eventually found the margins too narrow to sustain the work? In a kingdom where resources followed importance and importance followed position, a researcher without position was a researcher without materials, without collaborators, without anyone to tell them their methodology was flawed. The working draft's author had needed someone to say "define your terms." Nobody had.

  I was nine years old, building my own framework in cipher, in the same margins. The parallel was not comfortable.

  The working draft had been in this room the whole time. Anyone with a borrowed key and the right questions could have found it. The knowledge hadn't been destroyed or hoarded or lost to a guild collapse; it had been sitting two rooms from people who could have used it, in a section that nobody visited because nobody knew to visit it. The distribution problem wasn't always dramatic. Sometimes it was just a door that nobody opened.