Save the Fragment Before Explaining It

A writer catches a line, an image, half an argument overheard on a train, and at once wants to know what it is for. Which book? What does it mean? The fragment can’t answer, because part of its charge is that nobody knows yet. Asked every time, the question turns catching into an exam, and the writer either stops catching or files each scrap under CHARACTER or SETTING, which nails it to one function. Something that might have connected to something next month is closed.

Chekhov’s notebooks, in the 1921 English translation by S. S. Koteliansky and Leonard Woolf, contain entries like these. “Every one has something to hide.” “O how nice it would be if bachelors or widowers were appointed Governors.” None says what it is for. According to the volume’s prefatory note, his papers included sheets in a special cover marked “Themes, thoughts, notes, and fragments,” and the note mentions no sorting into character, setting, or dialogue. The only status it names is used or not yet used. When he used a piece of material, he struck it out.

So catching and explaining stay apart, with time between them. Catching has to stay quick. Explaining is slow, and it needs context that doesn’t exist yet. The delay probably works because a fragment’s gaps are information, which an explanation on arrival swaps for whatever the writer can think of that day, while the person rereading the notebook a month later has read and written more and begins to see connections that were invisible at capture. Many fragments will never be used, and sorting is allowed, but the demand for meaning waits until there are enough fragments for meanings to appear between them. On the spot, add only a date and a place, because that context can’t be reconstructed later. Joan Didion’s “On Keeping a Notebook” opens with an entry logged that way (a hotel bar in Wilmington, 9:45 a.m., an August Monday), and she can still recall the scene it came from, though later in the essay she asks what a recipe for sauerkraut is doing among the entries.

Sei Shōnagon’s Pillow Book, composed at the Japanese court around the year 1000, shows another way to hold fragments. Arthur Waley, whose partial translation appeared in 1928, notes that the text is arranged not by date but under headings such as “Disagreeable Things” and “Amusing Things,” a scheme that in his words often breaks down. Under “Things that make one Happy” she lists picking up a torn-up letter that someone has thrown away and finding that one can “fit the pieces together well enough to make sense.” (Waley adds that the relative age and trustworthiness of the surviving manuscripts had not been fully investigated, so the arrangement may not be wholly hers.) A pile of fragments, the passage suggests, can be sorted by reaction without being sorted by use.

Collecting is pleasant, it feels like labor, and a pile never fails, so a bank of fragments can pass for writing. The strike-out is the check, since whatever is never struck out was only collected. Some fragments will stay mute for years without that counting against them.

Practice. Pick one container you will always have with you: a pocket notebook, a note on your phone, a plain text file. (If you already have a defined project with research obligations, source and organize that material from the start; this exercise is for material that has no project yet.) For seven days enter whatever catches you: overheard speech, an image, a question, an unfinished scene, a sentence from nowhere. Give each entry a date and a place and write it as it came, with no title, tag, explanation, or folder. Aim for three a day, and don’t judge them. On the eighth day read the whole run once, aloud, pen down. Pick the fragment that bothers you most and write for ten minutes from it without deciding what it is for; if those minutes yield something you keep, strike the entry through, as Chekhov did. By then you should have twenty or more dated entries, and perhaps one struck through.