From a brief 1982 interview with William Gaddis—
Q: The pervasive and distinctive authorial voice of The Recognitions gives way in J R to a self-effacing voice that seems to serve only functional purposes. Also in J R there is an increased dependence on dialogue. For verisimilitude in Lolita, Nabokov “travelled in school buses to listen to the talk of schoolgirls.” Did you take any special measures to hear and note the speech patterns of any of the wide variety of people who speak in “J R”?
William Gaddis: Style and content must match, must be complementary, accounting in part for a difference between the two books, though the lack of a conventional narrative style had already jarred a good many readers of The Recognitions when it appeared, as its hapless reviews show. J R was started as a story which quickly proved unsatisfactory, inspired- here’s the legitimate gossip—-by the postwar desecration of the Long Island village of Massapequa where my family had had property since around 1910, take a look at it now and you’ll see all the book’s worst hopes realized. In approaching J R as a novel, I was at pains to remove the author’s presence from the start as must be obvious. This was partly by way of what I mentioned earlier, obliging the thing to stand on own, take its own chances. But it was also by way of setting up a problem, a risk, in order to sustain my own interest, especially since the largely uninterrupted dialogue raised the further risk of presenting a convincing sense of real time without the conventional chapter breaks, white spaces, such narrative intrusions as “A week later . . .” How some of the writers I come across get through their books without dying of boredom is beyond me. As for what you call speech patterns, one is always listening and has got an ear or hasn’t, and without one, unless perhaps in dealing with an unfamiliar language and culture, no amount of your special measures like riding around on school buses will get you out of the swamp.