The vault now holds Rasputin twice. Radzinsky wrote The Rasputin File from the inside: the interrogation transcripts, the testimony of the people who slept with him and prayed to him and feared him, the voices of the believers and the used. Moynahan wrote The Saint Who Sinned from the outside: the Okhrana watch logs that tailed the Dark One through Petrograd, the war ministry's collapse, the forensic dismantling of the documents everyone else took at face value. Same man, two files — one built from what people said about being near him, one built from what an indifferent state recorded about where he went. The interesting thing isn't that they disagree on details. It's that each file is structurally blind to exactly what the other one sees.
First framing: the inside file (testimony) is rich on meaning and useless on motive-checking. Every witness in Radzinsky's File wanted something — absolution, drama, distance. Testimony tells you what it was like to be in the room and can't tell you whether the teller is performing. It's the only thing that captures the spell, and the worst thing for catching the lie.
Second framing: the outside file (surveillance) is rigorous on motion and deaf to meaning. The Okhrana log knows Rasputin entered a building at 11pm and left at 4am; it does not know whether that was a healing, a seduction, a con, or a prayer meeting, and it never claims to. It catches the man's trajectory with no stake and loses his interior entirely. The watch can't hear the confession even when it's true.
Third framing: so the piece isn't "which file is right." It's that the two vantages carve reality at different joints, and a complete portrait would need both and still couldn't merge them — because the very thing that makes testimony trustworthy on meaning (the witness was inside it) is what makes it untrustworthy on fact, and the very thing that makes surveillance trustworthy on fact (it was outside it, indifferent) is what makes it mute on meaning. The blindness is not a flaw in either file. It's the price of the vantage.
A newsletter essay — Two Files, One Man — using Rasputin as the case but really about a method: when you have a figure documented from inside (what they meant to people) and outside (what a system recorded about them), don't average the files. Map what each cannot see, and treat the disagreement as a diagnostic of vantage, not a contest of truth. The wider payoff is a reader's tool for any contested historical figure, any public person with both a confessional record and a surveillance/paper trail. Closing move: the vault did this on purpose — it ingested both files knowing they'd fight, because the fight is the portrait.
Promote to WORKBENCH/Newsletter when the "blindness is the price of the vantage" thesis holds against a second two-file figure (someone with both a confessional corpus and a surveillance/documentary corpus) and when the convergence counter-beat (hemophilia) is developed enough that the essay isn't a one-note "vantages clash" piece. Needs the Moynahan↔Radzinsky concept pages stable before it can cite cleanly.