From Book to Research Project

The experiment grew out of the book

THE ARCHIVE began as a literary question and developed into a research program. The progression matters because the books and the research documents are linked, but they are not the same evidentiary object.

1 — The accidental longitudinal archive

Years of practical conversations with an AI created a record without first being designed as a formal study. The archive accumulated ordinary decisions, corrections, work products and repeated patterns across many domains.

2 — THE ARCHIVE

A found-phone thought experiment turned that record into a book: if someone found years of conversations but did not know the person directly, what portrait could an artificial intelligence reconstruct from the record alone? ChatGPT became an AI witness to the archive while repeatedly acknowledging that the archive was not the whole person.

3 — THE ARCHIVE II: THE MACHINES READ THE MAN

The original narrator then became part of the problem. Because long-term familiarity can create bias, the same manuscript was given to Gemini, Perplexity, DeepSeek, Claude and a blind ChatGPT session without the original narrator's conversational history. They were asked to analyze the manuscript, separate demonstrated evidence from interpretation, judge the narrator's reliability and identify what the subject appeared to be trying to accomplish.

Agreement among models is treated as convergence in reading, not psychological proof. The living subject then answers the machines, and primary artifacts enter the record.

4 — When the Evidence Changes the Machine

The literary experiment became a methodological question: when an AI's evidence about the same person changes, which judgments persist, weaken, strengthen, become more specific or disappear? The current research-integrity revision treats Round 1 as the completed auditable retrospective analysis, Round 2 as partial archival observation, and moves the stronger evidence-revision claim into a prospectively frozen replication.

5 — Mapping the Living Subject

The project then broadened into a non-clinical longitudinal person/mind mapping framework. It explicitly separates direct subject testimony, observable artifacts and behavior, repeated longitudinal patterns, AI/model inference and unresolved speculation.

6 — Phase III

Phase III moves outside the machine loop toward blind human evaluators, methodologists, domain experts, adversarial testing, External Reality Contact, Artifact-to-Reality Conversion and prospective observation.

The research is not trying to freeze the subject. It is trying to preserve the history of the revisions.