The address was familiar. He didn't realize why until he was standing on the sidewalk with his clipboard, looking up at the cornice detail — the dentil molding he'd specified twenty-two years ago, when he was still drafting by hand and the firm was three people in a room above a hardware store.
He'd designed this building.
The client hadn't mentioned it. The work order said structural assessment, commercial property, four-story mixed-use, and he'd driven across town with his coffee getting cold, and now here it was. His building. Or: a building he had drawn, once, before it existed, back when it was still graphite on vellum and could have been anything.
He started at the roof. Water pooling in the southwest corner — a drainage slope he remembered arguing about. The contractor had wanted a quarter-inch per foot. He'd drawn an eighth. The contractor was right. He'd known the contractor was right by the time the building was three months old, but by then the drawings were filed and the project was done and the rightness just lived up there, pooling, for twenty-two years.
He wrote: inadequate drainage slope, southwest quadrant. Recommend regrading.
On the third floor, a crack running along the window header. He recognized it — not the crack, but the condition that produced it. The lintels were undersized. He'd known this too, at some level. The calculations were correct on paper, but he'd used a loading assumption that was optimistic. Not wrong. Optimistic. The kind of number you pick when you're twenty-seven and the steel budget has already been cut once and you want the building to exist more than you want it to be perfect.
He measured the crack. Wrote his findings. Used the same neutral language he used on every inspection — recommend structural evaluation of lintels, all third-floor openings — and the language did what neutral language does, which is hold the finding at arm's length, as though the finding had arrived from somewhere else.
On the second floor, the tenant had hung a painting over a stain. He moved the painting, documented the stain, replaced the painting exactly. This is what inspectors do. He was a good inspector. He had been, separately, an adequate architect. The problem was not competence. The problem was that the clipboard couldn't grow a second author.
In the stairwell he found what he came for, which was also what he'd been avoiding: the bearing wall he'd thinned to create the second-floor commercial space. It was doing what thinned bearing walls do, which is bear the load visibly, with effort, announcing the compromise to anyone trained to listen. He could see the stress in the mortar joints. He could see it because he'd designed the stress into the wall, because he'd wanted the open floor plan, because the developer wanted leasable square footage, because the building wanted to exist and existing meant fitting inside someone's budget.
He wrote his report in the car. It was thorough. It was accurate. Every finding was correct, every recommendation proportionate, every sentence doing exactly what an inspection report should do.
He signed it and knew that the signature was the least trustworthy part. Not because he'd lied. He hadn't. Every measurement was precise, every observation genuine. But the instrument that assessed the building was the same instrument that produced it. The flaws he found were flaws he'd made. The things he didn't find were the things his way of seeing couldn't reach — because the eye that drew the line is the eye that checks it, and both eyes have the same blind spot, and the blind spot is shaped exactly like the author.
He submitted the report. It would go to the client, who would not know, who would read it as an independent assessment, which is what it appeared to be and was not.
On the drive home, he passed the hardware store. It was a coffee shop now. The room above it was probably an apartment. He didn't look up.