Marisol sends a voice message. She doesn't usually do this.
I play it twice. She says: I have been thinking about the sandwich. Not Claudio specifically — the sandwich as a category. I have been noticing my own small decisions differently since you sent the finding. This morning I chose a route to work. I did not know I had chosen it until I was two blocks in.
I sit with the phone on my desk in the Sleeve District studio, third floor, the one with the thermal-glass wall that faces the old rail line. Outside: a Felt Codec billboard switching between three process recordings on a 90-second loop. The billboard has been there four months. I stopped noticing it around month two.
The protocol has three participants: Claudio, Marisol, and me. All three of us consented. All three of us run scheduled sessions. Marisol did not run a session. She ran the protocol on her morning commute, on a route choice, without being asked, without documentation, without telling anyone until this message.
I write in the session log, under her name, in the column labeled session type: unsolicited.
Then I write the actual note: Day 1 of uncontrolled replication.
The protocol studies how people access the experience of deciding. I built it to be bounded — three subjects, controlled contexts, documented sessions. A route to work is not a controlled context. It is every morning, repeated, unobserved. The finding propagated into Marisol's daily life before I had finished writing it up as a finding.
I think about this for a while. Then I listen to the voice message a third time.
She says: I thought you would want to know.
She was right. I do want to know. I also do not know what to do with it yet. Those two things can both be true.