I SAT CROSS-LEGGED on my narrow bed, a stub of charcoal in one hand and a sheaf of scrap paper spread across my knees. The compass lay beside me, its needle still, pointing toward a book I could not reach.
The problem was simple enough in principle. I had a direction. If I took a second reading from a different position, I would have two lines. Where they crossed, my book would be. Triangulation. Child’s play for anyone who had studied under Master Brandeur.
But principle and practice are distant cousins at best.
The next morning, I descended into the Stacks with my compass and fresh determination. Two levels down, I found a quiet alcove near the eastern wall, tucked between shelves of Aturan histories that no one had touched in years. The needle swung, steadied, and pointed. I sighted along its length toward a distant shelf and estimated the angle against the corridor’s run.
Fifty-three degrees. Perhaps fifty-four. The needle was thin, my eye imperfect, and the margin was my uncertainty.
I paced off two hundred feet along the main corridor, counting each step. Then I took my second reading.
The needle pointed true. Sixty-one degrees off the corridor. Maybe sixty-two.
I sat down right there on the cold stone floor, heedless of the scrivs who might find me, and began to sketch.
The lines crossed behind the Four-Plate Door.
Not far behind it, perhaps ten or fifteen feet, if my angles were accurate. But it confirmed what I had discovered the day before, that my book waited in whatever space lay hidden from the world. I imagined it sitting on a shelf or a table, freshly processed, my dried blood pressed between its pages.
I spent the rest of the day in the deeper reaches of the Stacks, taking every reading I could manage. Angles, distances, positions, orientations. All imperfect alone, but with the hope that together they might tell the truth.
The next step was obvious. More readings from different positions would narrow my uncertainty. And if I took readings from different floors, I could triangulate in three dimensions, determining not just the horizontal location but the height as well.
I climbed the stairs toward the level above, already planning my approach.
Over the following span, I discreetly took readings everywhere. In the administrative sections on the sub-one level where students rarely ventured. On the main floor, with its high ceilings and reading tables. Down in sub-three, where the corridors narrowed and the air grew cold. I mapped the vertical tilt of my compass at each level, noted the horizontal bearing to the last degree I could estimate, and recorded everything in notebooks that grew fat with figures.
But nothing made sense.
A knock came at my door just as I was beginning my ninth attempt at reconciling the readings from sub-two with those from the main floor.
“Kvothe,” came Wil’s voice, patient but firm. “You’re late.”
I didn’t look up from my calculations. “For what?”
“Corners,” Simmon said from the doorway. “We’ve been waiting downstairs for the better part of an hour.”
“I can’t. I need to finish this.”
“I’ve barely seen you lately except for classes,” Sim said, stepping into the room. “Fela’s starting to think you’ve died.”
“I’m fine,” I said, gesturing helplessly at the papers scattered across every surface. “I’m close. I just need a little more time.”
“Close to what?” Wil had moved to stand beside Sim, his eyes scanning the figures and diagrams.
I pulled out my clearest sketch showing all of the readings overlaid on a rough map of the Archives.
Wil studied it for a long moment. “I don’t see the problem,” he said. “Four locations. Four objects.”
I gripped the edge of my desk. Four objects? My blood, divided and scattered? And each one a weapon waiting to be found.
“Kvothe?” Sim’s voice seemed distant. “You’ve gone pale.”
“Breathe,” Wil said, his hand steady on my shoulder.
I breathed. In. Out. The room steadied, though my hands still shook.
Sim and Wil shared a look.
“Whatever this is,” Sim said, waving a hand at my sketches, “you’re in over your head.”
I wanted to deny it. But my white-knuckled grip on the desk said otherwise.
“You’re good with numbers, Kvothe,” Sim continued. “But you’re not the best.” He paused. “I know what he did. But sometimes the best is what you need.”
The suggestion hung in the air. Brandeur. The man who had helped kill Herma. But I looked down at my diagram, and knew I had no other choice.
“Alright,” I said finally. “Tomorrow. I’ll go tomorrow.”
“Good.” Sim’s shoulders relaxed slightly. “Now will you please come downstairs?”
“One more night?” I asked. “I need to organize this first. If I’m going to Brandeur, I need to present it properly.”
They left, and I sat alone with my diagrams.
The next morning, I stood outside Master Brandeur’s office, my hand raised to knock for a long moment before I did.
“Enter,” came his voice through the door.
I took a breath and pushed through.
The office was buried in papers, weighted down with instruments, and smelled of chalk dust. The air was stale, as if his door rarely opened. Brandeur was alone. No students waiting for consultations, no colleagues stopping by to discuss research. Just a portly, balding man sitting behind a desk piled with work that no one would ever read.
I had played a part in that. A small part, perhaps, but a part nonetheless. And now I needed his help.
“Kvothe.” He looked up. Surprise, then relief, crossed his face. “I don’t remember summoning you.”
“You didn’t, Master Brandeur.” I kept my voice neutral, trying to ignore the tightness in my chest. “I came hoping you might help me with a problem.”
“A problem.” He leaned back in a chair that creaked beneath him. His eyes were sharp as ever, but they held a wariness I had not seen before. “What sort of problem brings a Re’lar to my door uninvited?”
“A geometric one.” I produced the papers I had prepared the night before. My angles and distances stripped down to just shapes and numbers. “I’ve been trying to locate a point using angular measurements from multiple positions. But my readings don’t agree, and I can’t determine whether the inconsistency is error or something else.”
Brandeur’s eyebrows rose slightly. He reached for the papers, and I let him take them. For a long moment, he said nothing. His eyes moved across my notations, my diagrams, my work. I watched for dismissal, for refusal, for some sign of retribution.
Instead, his shoulders unknotted and his breathing eased. The wariness in his eyes gave way to the hunger of someone who had been starving for work that mattered.
“What do these groupings represent? You’ve labeled them one through four.”
“Different observation platforms,” I said, which was true enough. “At different elevations.”
“And the readings within each level are consistent, but the levels disagree with each other?”
“Yes. I can’t figure out why.”
Brandeur made a small sound, something between a grunt and acknowledgment. He pulled a fresh sheet of paper toward himself and began to work.
Minutes passed. Brandeur filled one sheet with calculations, then another. He cross-referenced my numbers, drew diagrams, scratched out figures and rewrote them. His face had entirely lost its guarded quality. He looked like a different man. Engaged. Animated. Alive in a way I had never seen during admissions or faculty meetings.
“Interesting,” he murmured, half to himself. “Very interesting.”
“What do you see?”
He held up one finger, asking for patience. He worked for several more minutes, filling a third sheet with figures. Then he set down his pen. “You have two problems,” he said. “Not one. Two distinct anomalies, and you’ve been treating them as a single error.”
“Two problems?”
He tapped the papers I had sorted by floor. “First, your readings from your primary observation level. This one.” He indicated my sub-two measurements. “These readings cluster into two distinct groups. Watch.”
He drew a diagram, plotting my angles as lines radiating from their observation points. “Your early readings intersect here.” He marked a point. “Your later readings intersect here.” He marked another point. “The gap between these clusters is too large to explain by measurement uncertainty alone.”
“How can you be sure?”
“Statistics.” He wrote a formula I did not recognize. “Given your stated precision and the number of observations in each cluster, I can calculate the probability that both clusters represent the same true location. That probability is very small. Less than one in a thousand. It just doesn’t look like it on paper because of the substantial vertical adjustment.”
“So it moved,” I said slowly.
“Yes.” Brandeur shuffled through the papers again, his enthusiasm undimmed. “The second problem is more interesting. Look at your data from each level separately.”
He spread out four sheets, each containing one level’s worth of measurements. “Level one. The readings cluster tightly here.” He marked a point. “Level two. Tight cluster here, or rather, two clusters, as we discussed. Level three, tight cluster here. Level four, here.”
“If your target is stationary, these points should coincide. They don’t. Each level gives you a completely different answer.” Brandeur pulled out a fresh sheet and began sketching rapidly. “But watch what happens when I plot the vertical position.”
He drew a simple graph, elevation on one axis, observation level on the other. Five points appeared, then he drew a horizontal line through four of them. “If we throw out the anomaly from level two, all four levels agree perfectly on the elevation. Your target is at the exact same height, regardless of which floor you measured from.”
I stared at the line.
“According to statistics,” Brandeur continued, “if these were truly four different objects scattered across space, the probability they would all share the exact same elevation is vanishingly small. This confirms what the movement data already suggested. You’re measuring a single object.”
The weight lifted from my chest. A single object.
“But now we come to the interesting part.” He returned to his diagram of the four horizontal positions. “Look at the disagreement here. If this were random error, these points would scatter in random directions. But they don’t.”
He drew lines connecting the four points. A rough spiral, each point rotated relative to the last.
“The disagreement is systematic,” he said. “Each level’s answer is rotated relative to the others. As if.”
He trailed off, waiting for me.
I stared at the spiral pattern, knowing he was trying to lead me somewhere but still not understanding.
Finally the pieces clicked into place. “The coordinate systems,” I said, my words coming faster. “What if the reference frame at level one is rotated relative to level two? And level two is rotated relative to level three? That would explain exactly this pattern.”
Brandeur looked up at me, and I saw genuine pleasure in his eyes. A teacher’s satisfaction when a student makes the leap. “Precisely. You’ve been assuming your coordinate systems align. That ’north’ at level one points the same direction as ’north’ at level two. But your data suggests otherwise. Each level seems to have its own orientation, rotated relative to the others.”
“But why would reference frames rotate between levels?”
“I have no idea.” Brandeur shrugged. “I don’t know what you’re measuring or where you’re measuring it. I only know what the numbers say. And the numbers say your coordinate systems don’t align.” He gathered my papers into a neat stack and held them out to me. “Whatever space you’re working in, it’s not oriented the way you think it is. Fix your coordinate transformations, and your data will converge.”
I took the papers. Rotating coordinate systems. Reference frames that shifted between levels. What could possibly cause that?
“How do I fix it?” I asked. “If I don’t know how much each level is rotated?”
“Measure it.” Brandeur’s tone was practical now, a teacher assigning homework. “Find some way to determine the angular offset between your reference frames. Then transform all your observations into a single unified system. The mathematics is straightforward. I can show you the rotation formulas if you need them.”
He did then, spending another quarter hour walking me through the process of transforming coordinates from one rotated frame to another. He was a good teacher. Patient and clear, building each concept on the foundation of the last. Nothing like the silent antagonist I had imagined from years of watching him vote against me at Hemme’s side.
I wondered, for the first time, how much of that had been his choice. Hemme had a talent for finding people’s weaknesses and pressing on them, and Brandeur was not a man who made friends easily. If Hemme had been one of the only people to show him kindness, to include him, to make him feel like he belonged.
It would not excuse what had happened. But it might explain it.
“Thank you,” I said, gathering my papers. “This has been more helpful than I had hoped.”
Brandeur nodded, and I saw gratitude in his eyes. “It’s been a fascinating problem,” he said.
I grabbed some more paper and headed back to the Archives. Wanting to avoid the many eyes that might question what I was about, I bypassed the main floor and descended to sub-one, then to the stairwell that led down to sub-two. At each turn of the stairwell, I traced the angle onto a paper wedge. When I reached the bottom, I laid my collection of pieces edge to edge. Quarter-turn. Half-turn. Three-quarters. Full rotation. Another quarter-turn.
I continued downward, tracing each step with growing certainty. By the time I reached sub-three, the pattern was unmistakable. Each descent turned the world around me.
I stood in the corridor outside Puppet’s room and pictured it. Sub-one, turned on its axis relative to the main floor. Sub-two, turned again relative to sub-one. Sub-three, turned once more. Each level connected to the next by a single stairwell, but that connection was a pivot point, not an anchor. The floors did not stack like plates in a cupboard. They spiraled like a nautilus shell, each chamber turned from the one above.
Puppet had lived on sub-three for years. He claimed his room was directly beneath the Four-Plate Door, one level above. Had spoken of listening for sounds through the stone, of pressing his ear to the ceiling and imagining what secrets might be kept just overhead.
But he was nowhere near the door. The geometry of the Archives had fooled him as thoroughly as it had fooled me.
I left the Archives, my verification complete. But what I needed was precise measurements I could trust.
I returned to the Archives in the early afternoon and spent the next several hours with the surveyor’s compass and measuring cord that I had borrowed from the Fishery. The main floor. Sub-one. Sub-two. Sub-three. Each level measured with exquisite care, each bearing recorded with precision that made my earlier attempts look like guesswork.
The numbers were better than I had hoped. The main floor was rotated twenty-one degrees from the entrance. Sub-one, thirty-four degrees from the main floor. Sub-two, fifty-five from sub-one. Sub-three, eighty-nine. Each level turning further than the last, spiraling like a nautilus shell.
By evening, I had applied all the transformations and the clusters converged. All four floors. All agreeing. All pointing to the same location thirty feet east of the Archives’ exterior wall.
I sat back and stared. I had assumed sub-two lay deep underground. But the rotated geometry of the Archives and the grade of the hill meant it extended not downward but outward, just below the surface. Just beneath the oldest building on campus.
Just beneath Mains.
And Puppet, poor Puppet, had spent years listening for secrets through a ceiling that was nowhere near the room he sought, never thinking to question what seemed so obviously true.
I checked the time. Evening had settled in, and I remembered I had promised to meet my friends. If I left now, I’d have just enough time to make it to Corners.
Even so, when I arrived at Corners, Sim, Wil, and Fela had already claimed our table with a half-empty pitcher in the middle. In the first hand, I missed a crucial trick, overlooking a sequence I should have seen coming three plays earlier. Wil raised an eyebrow at my distraction but said nothing. A tankard of ale helped. Then, a few more hands, and gradually the game pulled me in. For a few hours the world became just cards, friends, and the comfortable rhythm of play.
When we finally dispersed near midnight, I headed toward Mains. Most of its windows were dark. I circled to the side closest to the Archives, climbed the old drainage pipe, and within minutes pulled myself onto the lower rooftop. From there I picked my way across the maze of peaks and valleys.
A few compass readings from different positions gave me what I needed. The lines converged on a section of roof between two taller ridgelines. I moved there and checked my compass one final time. The needle pulled hard toward the rooftop beneath my feet, tilted at such a steep angle that I knew I was nearly centered above my book.
It seemed so unlikely that I moved to the edge and peered down. Just to verify.
And sure enough, below was Auri’s abandoned courtyard and apple tree.
I sat back and pulled out my blueprints, sketching in the courtyard’s position. Thirty feet across. The blank wall formed its northern border. Whatever lay behind the Four-Plate Door would be just beyond that wall, and the entire space should fit within Mains’ foundation.
I looked down at the blank wall again, at the peaceful courtyard in the moonlight. How many times had I practiced here, never questioning why one wall lacked windows? How many times had I met with Auri, whatever secrets lay just beyond that blank stone?
The geometry was clear now. The Four-Plate Door sat both in the Archives and beneath Mains. I gathered my compass and diagrams, and headed back toward my room. But I already knew where my next steps would take me.
| Chapter 8 | Contents | Chapter 10 |