Mavis Gallant published her first short story in The New Yorker in 1951. She published her last in 1995. Across those forty-four years, she sent the magazine one hundred and sixteen stories. She wrote, across the same period, considerably more letters to her editors than she wrote stories.
A representative selection of these letters, focused on her thirty-four-year working relationship with the editor William Maxwell, is now available to researchers at the New York Public Library's Berg Collection. The folder is small, perhaps three hundred letters in total, but it is one of the most useful things in the archive for anyone interested in how a short story actually gets made.
Gallant lived most of her adult life in Paris. Maxwell lived in New York. They saw each other rarely. They wrote constantly.
The letters are not, mostly, personal. Maxwell was a friend, but he was also her editor, and Gallant kept the relationship professional in writing even when it was warmer in person. She addressed him, in almost every letter across thirty-four years, as Dear Bill.
What the letters preserve, in unusual detail, is the back-and-forth of a story going to press.
Gallant's drafts went to Maxwell in Paris by airmail. He read them carefully and wrote back, sometimes within a week, with queries. The queries were rarely structural. He almost never asked her to add or remove scenes. He almost always asked about sentences.
A typical Maxwell letter to Gallant, from 1968, runs to two single-spaced pages and concerns one story. It lists, in numbered order, perhaps thirty small questions. Most of them are about commas. A few are about specific words. One or two are about whether a piece of dialogue could end half a beat earlier.
Gallant's replies are roughly the same length. She accepts about half of his suggestions, defends about a third, and proposes alternatives for the rest. The defences are sometimes quite firm.
There is, for example, an exchange from October 1973 about the placement of a single word in a story called The Latehomecomer. Maxwell had suggested moving the word two sentences earlier. Gallant wrote back that the word belonged exactly where it was, and explained why for almost a page.
The word stayed.
What is moving about the archive is not the disagreement but the rhythm of agreement underneath it. Maxwell and Gallant agreed, fundamentally, on what a sentence was for. They agreed on what a story was for. They disagreed only at the level of execution, and even then they disagreed in a shared vocabulary.
This is rarer than it sounds. Most writers and editors are not, in fact, working from the same set of underlying assumptions. They are negotiating between two different aesthetics. The negotiation produces compromise sentences, which are usually worse than either party would have written alone.
Maxwell and Gallant did not produce compromise sentences. They produced Gallant sentences that were sometimes slightly improved by Maxwell's attention.
The letters also document the slow, grinding work of cuts. The New Yorker, especially under William Shawn's editorship, ran long. But it still ran shorter than Gallant tended to write. Her stories often arrived at sixteen or eighteen thousand words and had to come down to twelve or fourteen.
Gallant did not enjoy cutting. The letters show her arguing, sometimes paragraph by paragraph, for sentences Maxwell had marked. She lost some arguments and won others. The won ones are visible in the published stories; the lost ones survive only in the drafts in the archive.
Some of the lost sentences are very good. Gallant kept them in her files and, occasionally, reused them in later stories, sometimes a decade later, in slightly altered form. The archive lets a careful reader trace this migration.
Across the thirty-four years, the tone of the correspondence becomes warmer but not less rigorous. By the late 1980s, Maxwell, in his eighties, is writing slightly shorter letters, but the queries are still as exact. Gallant's replies remain firm. Their last working letter, from 1995, is about a comma.
Maxwell died in 2000. Gallant died in 2014. Her literary estate, which includes the carbon copies of her side of the correspondence, was placed at the Berg Collection in 2015. Maxwell's side had been there since 2001.
Reading the two halves together, in the order they were written, is an unusual experience. The letters do not form a narrative. They are, mostly, a long technical conversation about how the English sentence works.
But they are also, cumulatively, the record of a long professional friendship maintained almost entirely through paragraphs about other paragraphs. The two correspondents were rarely in the same room. They were, on the page, very nearly always in the same room.
Anyone interested in editing, in writing, or in the working life of a great short-story writer should spend a morning with the folder. It can be requested at the front desk of the Berg Collection on the third floor of the main library building on Fifth Avenue.
The reading room is quiet. The letters are unphotocopied. A pencil is permitted. A pen is not.
It is the kind of archive that does not, in the end, change what one thinks about Mavis Gallant's stories. It changes, instead, what one thinks about the labour that produced them.