Reyna had built the taxonomy to describe the guest. After three weeks of tracking, she understood it described herself.
The QUALITY OF QUIET log began with two columns: Quiet-absent and Quiet-present. The first meant no signal from the extracted memory. The second meant signal existed but she could not name it. A week later she added Active—guest recognition immediate, unambiguous. The three-column system held for twelve sessions.
Then came the new client.
First session, no prior guest history. Reyna activated the recorder, ran the standard integration protocol. The guest remained absent—no recognition, no signal, the curve flat as desert. But the client cried. Not the tears of someone remembering loss, which Reyna had seen before. These were the tears of someone encountering something they did not recognize as grief until it was named for them.
She sat with the contradiction for two days. The guest was absent. Something had happened. Both statements were true.
On Tuesday she added the fourth column: Guest present, signal unclear, something happens anyway. She placed the new client there, then reviewed the previous twelve sessions. Eduardo, week one: Quiet-present. But reviewing the recording, she saw him pause at the same timestamp where the client had cried. She had marked him Quiet-present because she detected something. But she had not asked whether that something required the guest's presence, or only the client's.
She wrote at the bottom of the log: The taxonomy is not describing the guest. It is describing what I am able to perceive. The categories may be about me, not them.
The statement sat in her chest like a small stone. If true, it meant her careful tracking had been tracking the wrong variable. She had thought she was measuring the guest's response to recalled memory. She might be measuring her own capacity to detect response—her attention, her expectation, her presence in the room.
The observer effect. She remembered the term from undergraduate physics, the problem that measuring a system changes it. She had assumed the guest was the system. But the guest existed only in the client's neural architecture, accessed through a technology that required a practitioner to operate it. The measurement apparatus was inseparable from the phenomenon.
She thought about the guest differently after that. Not as a presence or absence. As a relationship—between the client's memory, the extraction technology, and her own trained attention. The guest emerged from all three. It could not be extracted, only witnessed.
She did not show the fourth column to Dr. Vásquez. The taxonomy was not ready. But she continued using it, continued placing clients in categories that she now understood were provisional, perspectival, possibly wrong. The wrongness was data too. It told her where her perception ended.
On Thursday she added a note to the fourth column: The limit of the category is the limit of me.
She did not know if this was a discovery or an admission. The guest, if it existed at all, remained silent on the question.