The Page Is a Poor Recording Device
The sentence that mattered most in a testimony I recorded in 1994 is not on the tape. It is in a margin note I made ten minutes after switching the machine off.
What the room is reading this week. Long-form work by people who write for a living and people who write to stay sane.
The sentence that mattered most in a testimony I recorded in 1994 is not on the tape. It is in a margin note I made ten minutes after switching the machine off.
One sentence, one man, three punctuations, three different accounts of who is to blame. Spoken language has no commas. Every mark a transcriber adds is a claim.
A river became a flood, a house became a residence, pots became possessions. Nothing was factually wrong, and the person who said it had vanished from her own sentence.
A man told me about his father's arrest four times in one sitting, circling back and adding a layer each pass. An editor asked me to put it in order. There was no order to put it in.
A woman said the word for alone three times in one paragraph. An editor circled all three as redundant. The three repetitions were not doing the same job.
A liver gone black with worry read as the sharpest line in the chapter in Darija and a dead cliche the moment I moved it into French, where the same image wore out generations ago.
A copy editor removed what she called Quechua interference from a testimony, meaning the word order the speaker actually used. Ask which grammar you correct toward without noticing.
Most people who will ever read one of my books will read it in a language I did not write it in, which means a sentence that only works as a sound is a bill charged to almost nobody who can pay it.
An editor I studied never let a testimony leave her archive until she had read the edited version aloud to the speaker and watched their face. A signature cannot catch what that catches.
Remove every adjective from a soft paragraph and you usually find either a sentence that needed nothing added or an empty one the adjectives were hiding.
Four seconds of nothing on a tape can be a search for a fact, a refusal, or the cost of saying something already fully remembered. A waveform cannot tell you which.
One word, correctly spelled and grammatically placed, told my editor the ending of a subplot I had not yet decided, because I meant it the French way and wrote it the English one.
An editor upgraded a man's plain account of his brother's funeral into something more literary. It read better and kept two of three facts. I asked her to put it back.
A witness named Delia nodded through my draft of the night a hillside gave way. Her cousin, standing forty meters up the same slope, stopped me at the third sentence.
Two independent counters on the same transect never agree exactly, and the discipline is not averaging them. It is investigating only where they disagree.
A footnote reading only translated loosely sat in my own anthology for years before I noticed it apologized for nothing in particular. An apology needs to say what it is sorry for.
A reader circled the scene where a grandmother closes a child's fingers over a coin and asked for a footnote. I did not add one, because the plot proves what the gesture means three scenes later, faster than a definition could.
A publisher sent a six page, thirty one term glossary for approval. Twenty six of the terms were already explained by the sentences around them. The glossary was not informing readers, it was confessing which sentences had failed.
My first published essay ends on a sunset over a fjord. I would delete that paragraph today, because the sentence was never actually about that fjord.
A woman signed a form about a project and later found the exact sum she borrowed from a cousin printed in a draft. The detail that made the sentence good was the detail that exposed her.