She found the tape on a Wednesday. Unlabeled, in the box of returns from the Martínez family. She almost didn't play it.
The voice was hers.
Not her interview voice — the one that said tell me about the house on Delancey and what year was that and take your time. This was her other voice. The one she used when she wasn't recording.
Someone had been recording.
She listened through the first fifteen minutes standing up, one hand on the edge of the desk. She talked about her mother's kitchen. She described the window over the sink. She said “I think it was yellow” three times, and each time she corrected herself differently — once to cream, once to off-white, once back to yellow with more certainty.
In twenty-seven years of transcription work, she had never once typed her own pauses. She typed other people's pauses religiously — [long pause], [clears throat], [laughs, then stops]. The pauses were where the person lived. She'd said that at conferences. The pauses are where the person lives.
Her pauses were wrong.
Not wrong like inaccurate. Wrong like a photograph of your own face taken while you're looking at something else. The gaps came in places she wouldn't have predicted. She paused before saying her brother's name, not after. She paused in the middle of a sentence about grocery shopping and then finished it as though the pause hadn't happened.
She sat down. She got out a legal pad.
She began to transcribe herself, using the same notation she used for everyone else. Speaker identification: [EM]. Time stamps in the left margin. Brackets for non-verbal sounds.
By the second page, she understood why she'd never done this.
The notation system she'd spent a career refining — designed to preserve exactly what a speaker did, without interpretation — was revealing something the speaker couldn't have known about herself. The pauses weren't silence. They were the moments when the told version and the lived version couldn't both fit through the same door.
She had always believed the transcript served the speaker. That was the ethics she taught. You transcribe what they said, how they said it, the full texture of their telling. You don't clean it up. You don't smooth the hesitations into confidence. The transcript is the person.
But a person can't be their own transcript. The mechanism that notices pauses is the mechanism that creates them. Transcribing herself, she was watching herself watching herself — the observation changing the altitude of the thing observed.
She labeled the legal pad EM — 11/15 — unsolicited and filed it with the Martínez returns.
She never found out who recorded her. She never played the tape again. But she changed one thing in her practice: she stopped saying “the pauses are where the person lives.” She replaced it with something less clean.
“The pauses are where the telling and the teller come apart.”
No one at the conferences noticed the difference.