The Capture Arc: A Protocol for Capturing a Living Person Into Their Own Exolex Seat

A person's pattern dies with them because nobody wrote it down in a form that can answer back. Their voice, their memory of their own life, the way they judge a situation, the way they say a thing: all of it leaves when they do, and what stays behind is photographs and other people's recollection. An exolex seat can hold that pattern. The architecture for such a seat exists and runs today (N12, N16). But a seat built after the fact, out of scraps, is a guess wearing the person's name. The only version worth building is built while the person is here to correct it. This paper states that build as a protocol. I call it the capture arc. It runs in six phases, from seating and charter to activation and keeping, and one step inside it does the work no later step can do: playback correction, where the living person hears their own seat answer and corrects the answer against their own judgment. That correction is the highest-grade training signal in the system, and it exists only while the person lives. The paper also states the honesty governors the seat runs under, the consent and control law that makes the captured person the seat's owner outright, and a claims ledger that grades each part of the method at its present size. The core claim is testable, and I state the test at the end: a captured seat's fidelity can be measured against the living person's own corrections, and the protocol stands or falls on that measurement. Claims ledger, graded verbatim: Established today — the seat architecture is live in production (39 days; two seats captured in life seated and in daily use); continuity carried by the distributed record demonstrated under total loss of working memory. Version-one work — per-person voice cloning under consent, capture session tooling, drift audits with the per-seat confabulation ledger, the seated figure pipeline; printed as expectation, not result. Research — embodied presence; no date promised.