The fourth day was a Thursday, and Mira had reached the point in her preliminary survey that required fieldwork rather than archival work -- she needed to see the buildings themselves, the physical spaces the guilds had occupied, before the documents about those spaces would fully make sense. She put the box of Weavers' records back on its shelf, told Tessa she would be out in the afternoon, and walked to Bellmaker's Row.

The Row was six buildings on the north side of a lane that curved slightly with the original contour of the terrain, following the old route toward the mine road. Most of the buildings were still standing. Two had been converted -- one to storage for a building supply company, one to what appeared to be an informal community workshop space, the kind of place that existed in old craft buildings after the original craft had left, filling the space with the slightly different activities of people who wanted to make things but had no specific trade. The other four were empty, their windows dark and their doors padlocked, the facades showing the specific quality of neglect that was not dereliction but suspension -- buildings waiting rather than dying, not yet decided about their future.

She started at the western end and worked east, sketching the facade of each building in her field notebook, noting the construction details, the evidence of modification and repair, the way each building's architecture spoke to the specific function it had originally served. The smithing buildings were easy to identify: the chimney stacks, the reinforced floors, the window placement designed for ventilation rather than light. The bookbinders had been in a narrower building with large north-facing windows -- the light preference of anyone working with fine detail. The weaving shed at the far end was longer and lower than the others, the windows even larger.

She had been working for about an hour when she heard something from inside the second building from the eastern end -- a small metallic sound, precise and deliberate, the kind of sound that indicated someone working rather than someone moving through a space.

She went to the window.

Corin Albrecht was inside. He was at a workbench in the middle of the space -- not the kind of workbench that had been there for a hundred years, but a portable one that he had clearly brought himself -- and he was working on a metal object that she could not immediately identify. He was facing slightly away from her, and the quality of his attention was total: head down, shoulders very still, his hands doing something that she could not see at this angle.

She watched for a moment before deciding to enter. She was not, precisely, trying to observe him without his knowledge -- the window was not obscured, she was standing in the open, if he had turned around she would have been immediately visible. But she was also not announcing herself, and she was aware of this as a choice. She was watching because the watching was informative in a way that a direct conversation sometimes was not.

He was working on a brace of some kind -- an old one, the kind of hinged brace that had been used in structural ironwork before standardized fittings had made custom work unnecessary. He held it in his left hand, examined the joint, and then made a series of very precise small adjustments with a tool in his right hand. Not improvising. Not figuring out as he went. Working from the specific knowledge of what this object needed, the way someone worked who had spent time -- real time, not theoretical time -- understanding exactly how this type of mechanism functioned and failed.

He was in the metalwork cycle. She had figured that out two days ago, from the way he had known about the iron alloy in the staircase brackets. This was what that looked like from the outside.

She went to the door.

"It's open," he said, without looking up. He had heard her on the step. Or possibly seen her in the window and simply waited.

She pushed the door open and stepped inside. The space had the specific smell of old metal and stone dust and the particular quality of cold that unheated stone buildings had in winter -- not damp cold, but still cold, the kind that penetrated steadily and did not dissipate quickly. There were tools on the portable workbench. There were samples of various hardware laid out in what appeared to be a systematic arrangement -- different periods, different alloys, the kind of comparative array that a specialist would put together before assessment work.

"Assessment?" she asked.

"The hardware collection for the survey." He set the brace down and picked up another. "The archive has a collection of original fixtures from the Bellmaker's Row buildings -- removed during the partial restoration in the 1980s, stored in boxes in the basement. I'm doing condition assessments and datings."

"How long has the collection been in boxes?"

"Forty years." A pause. "Some of it longer. The removal wasn't well documented."

Mira walked to the workbench and looked at the comparative array without touching anything. The objects were organized chronologically, she thought, or possibly by alloy type -- she could not be certain which because she did not have his vocabulary for reading the objects themselves. "You can tell the dates just from looking?"

"Most of them." He picked up one of the objects from the array and turned it over. "This one I'm not certain about. The style suggests 1870s but the material weight is heavier than typical for that period. Either it's a later reproduction or the smith was working to a heavier specification for a specific load-bearing application."

She looked at it. "Can you test the alloy composition?"

"Not here. But the archive has a reference spectrometer from a donation -- I'll take samples of the ambiguous ones to Rosch's lab next time I go to the hospital for a check-in." He said this in the tone of someone mentioning a logistical detail, not a medical one. She filed it.

"The smiths who made these -- the ones from the 1870s and 1880s -- were they also in the Halvern-associated group?"

He looked at her. It was the first time since she had entered that he had stopped working and looked at her directly. "Where did you find that term?"

"Guild council minutes, 1896. Item on the agenda: Report from the committee on the matter of the Halvern-associated members. Discussion deferred."

Something in his expression changed in a way she could not quite categorize. Not surprise -- or not only surprise. Something more complicated. He set the object down carefully. "Did you find the follow-up?"

"Not yet. The 1897 records are in the next box. I haven't gotten there yet."

He said nothing for a moment. Then he picked up the brace again and returned to his adjustment. "The smiths in the comparative array -- most of them, yes. The ones I've been able to match to individual makers through maker's marks." A pause. "The craft quality is unusual for the period and the scale of production. Whoever made these knew what they were doing in a way that's difficult to account for with conventional apprenticeship timelines."

Mira watched him work. The adjustment he was making to the brace was very fine -- the kind of work that required knowing in the hands rather than in the head, the kind that could not be directed entirely by thought. "How do you know what they were doing was unusual?" she asked. "Rather than just skilled."

"Because the techniques are inconsistent within the same maker's work." He did not look up. "I've seen three different objects that I believe were made by the same person -- the maker's marks are similar enough that I'm reasonably confident they're from the same workshop, possibly the same hands. The techniques used in each object are not consistent with a single trained approach. They're consistent with three different trained approaches used in the same hands."

Mira was quiet for a moment.

"That's what you'd expect," she said carefully, "if someone had been trained in multiple techniques by different teachers."

"Yes." He set the brace down and looked at her with the kind of directness that she had come to recognize as his version of full attention. "That's one interpretation."

"What's the other interpretation?"

He looked at the objects on the workbench for a moment. Then he picked up the brace again and returned to the adjustment. "I'm still working on the other interpretation," he said. "I'll let you know when I have it."

---

She spent the rest of the afternoon on Bellmaker's Row, working through the facades systematically, noting the construction details, building a picture of the physical environment that the guild members had inhabited. She was aware, in the back of her working mind, of what Corin had said about the maker's marks and the inconsistent techniques. She filed it carefully in the column labeled evidence rather than hypothesis, because it was data -- specific, material, observable -- rather than interpretation.

Multiple trained approaches in the same hands.

She thought about the cross-guild entries in the administrative records. Fourteen out of forty-two smiths with documented connections to other guilds. The frequency too high for individual exception, too patterned for coincidence. And now, from the physical objects themselves, the suggestion that the makers who could be identified were not working from a single coherent training.

Multiple trained approaches. Multiple guild associations.

She was going to need the 1897 records. She was going to need the Halvern geological survey. She was going to need to understand what the Grey Ferrite deposits were doing in the physical environment of the people who had produced these objects.

She was also going to need to understand Corin Albrecht's other interpretation. Because she had a strong suspicion that his other interpretation was the same as hers, and that he had been sitting with it alone for a long time.

She wrote in her notebook, on the page she had started keeping for things she knew she did not yet have enough data to say: Multiple techniques. Same hands. Halvern proximity. This is the same thing.

Then she looked at Halvern Ridge, rising above the northern end of the Row, its peaks lost in the low February cloud. Patient as all geology was patient. Holding whatever it held.

She closed her notebook and walked back toward Archive House.

The cold had settled into the lane the way cold settled in places that stone had been absorbing for a long time: not sharp, but total, the accumulated cold of a hundred and fifty winters in these walls. She pulled her coat closer and thought about the brace on the workbench and the precise way Corin Albrecht's hands had moved around it, knowing exactly what they were doing, not needing to think about it.

She thought about knowledge that lived in the hands rather than the head.

She had spent ten years studying the material culture of working people and she had never thought about what it would mean for that knowledge to be truly in the body -- not learned, not remembered, but present, the way breathing was present, the way balance was present. She had always assumed the body was the tool and the mind was the craftsperson.

She was beginning to think that was not quite right. Or that it was only right from the outside.

In the evening, back at the guesthouse, she opened her notebook to the page she had started keeping for Corin Albrecht specifically -- not her research notes, but the separate log of observations she kept when a person in the field turned out to be more than a source.

She wrote: Thursday, Feb. He works alone in the Bellmaker's Row building with his own portable equipment. Not archive work -- personal assessment work that he brings materials to the archive for documentation afterward. Knows the objects at a level that requires physical handling, not just archival study. Confirmed: multiple trained approaches in same maker's work (his observation, consistent with cross-guild entry hypothesis).

Then she wrote, more carefully: Knowledge type A -- what he knows from the physical evidence. This is total. Immediate. He does not hesitate. Knowledge type B -- what he knows from documentary evidence, or from experiences that are not currently present. This type shows gaps. He says recently rather than not. He says I'm still working on rather than I don't know.

She stopped writing and looked at what she had put on the page.

Two types of knowledge. One total and present, one partial and qualified. The difference between them was not apparent from the outside -- he was certain in both, he did not perform uncertainty. But the uncertainty was there in the specific words he chose, the slight pause before certain answers, the careful framing.

She wrote a question at the bottom of the page: What determines which knowledge is present and which is qualified?

She had not found an answer yet, but she had enough of a pattern to know that there was one. She was good at patterns. She found them by accumulation -- recording everything that seemed possibly significant, not deciding yet which observations mattered, and then later, when the full picture was clearer, going back and identifying which thread had been load-bearing all along.

She capped her pen and closed the notebook and looked out the guesthouse window at the street below.

The Halvern-associated members entry was still sitting at the top of her hypothesis page, waiting for the 1897 follow-up. The cross-guild names were accumulating in their own column. And now there was a third thread: a researcher who could tell you the alloy content of a Victorian ironwork bracket without spectrometric testing, who was apparently still working on his interpretation of what the guild records meant, who had said recently like someone describing a temporary absence rather than a permanent one.

She had the feeling she was working toward the same conclusion from two different directions and would find out when they converged whether she was right.

She had always liked that feeling. It meant the problem was real rather than imagined, and that the evidence was consistent rather than cherry-picked, and that the answer existed and was findable if you kept going in the right direction.

She was fairly confident she was going in the right direction.

She was less confident she was going to like everything she found when she got there.

That was also standard. The most interesting research problems always had that quality -- the sense that the answer, when you found it, would require you to adjust something you had thought you knew. Not just about the subject but about something adjacent to it, something you had not known was part of the question until the answer made the question bigger than it had looked.

The following morning she arrived at Archive House to find a note from Tessa on the reading room table -- Corin had requested the 1897 Smiths' records as well. Tessa had put them on the shared table in the reading room rather than reserving them for either researcher specifically, which was either a diplomatic decision or an administrative convenience or possibly both, and which effectively meant that working through the records would require negotiating shared access.

Mira set her notebook on the table beside the box and waited.

He arrived twenty minutes later, coat damp from the light sleet outside, carrying his folder and two cups of coffee. He set one of the cups on the table on her side without saying anything.

She looked at the cup. Then at him.

"I make it too strong," he said.

"So do I," she said.

He sat down at his end of the table and opened his folder.

She opened the box and found the 1897 minutes.

The follow-up entry was the third item on the agenda of the February 1897 meeting: Report from the committee on Halvern-associated members -- findings presented, discussion held, committee recommends continued monitoring without formal documentation of individual cases. Passed by vote.

She read it twice. She copied it exactly into her notebook.

Then she looked up. Corin was reading something in his folder, but his pen was not moving. He was not actively working.

"Continued monitoring without formal documentation of individual cases," she said.

"I see it," he said, without looking up.

"The committee knew something was happening. They decided not to write it down."

"That's what the record says."

She looked at the entry again. Continued monitoring. Individual cases. Not techniques. Not skills. Individual cases. The language of a committee describing people rather than practices, managing something that was happening to specific people rather than to the guild's professional standards in general.

"They were protecting someone," she said.

He set his pen down. He looked at the box of records on the table between them, not at her. "Or they didn't know how to document it," he said. "If what was happening was outside their reference frame -- outside what they had language for -- the most rational response would be to observe without committing to an explanation that might be wrong."

Mira looked at him. He was still looking at the box.

"You sound like someone who has thought about this before," she said.

A pause. "I've looked at the records before." He picked up the pen again. "From a different angle."

She turned back to the 1897 minutes and kept reading. He kept writing. Outside, the sleet was thickening to something closer to snow, the grey light from the windows going whiter and more diffuse. The river was audible even through the glass, steady and unchanged.

She was going to need to understand what his angle had been.

She thought she might also need to understand why it mattered to him personally.

For now, she read the records and let those questions sit where they were, at the edge of the page, waiting.