I certainly agree with Matt Cheney:
. . .my problem is I don't particularly care what something is labelled, and I especially don't like issuing proclamations that all writers should or should not write in some specific way. What I desire as a reader and a writer has little to do with whether something fits anyone's definition of realist or not, and even less to do with how it fits into a marketing category. As a reader and writer, I find work compelling when it aims to push the possibilities of language and structure, of emotion and character, of ideas and effects. I grow bored with books that are a lot like other books and with stories that don't strive toward being something more than simple entertainment. These qualities can be found in works that fit just about any definition of realist or nonrealist. . . .
Although I'm not sure I would agree that the relevant opposition is between "realism" and "nonrealism" (Matt wants to speak up for the latter among those who would privilege the former) so much as between notions of the "well-made story" or of "craft" and, as Matt puts it, of fiction that pushes "the possibilities of language and structure." Historically, ideas about "the craft of fiction" certainly did arise along with the rise of realism, and even modernist experimentation tended to be along the lines of providing greater verisimilitude, conceived of as psychological depth, internal rather than external realism, rather than a rejection of realism as an ultimate goal. But writers who provide rather too much realism (e.g., Stephen Dixon or Harold Jaffe) without the accompanying comforts of conventional storytelling are as likely to be neglected by mainstream critics and readers of "literary fiction" as the most thoroughgoing fantasist.
It is also true that "realism" as a practice (if not necessarily the term itself) came to prominence as a more or less deliberate rejection of the kind of fiction Hawthorne called "romance" in describing his own work. According to Hawthorne, what we would now call realism "is presumed to aim at a very minute fidelity, not merely to the possible, but to the probable and ordinary course of man's experience," while romance, althought "it must rigidly subject itself to laws, and while it sins unpardonably so far as it may swerve aside from the truth of the human heart--has fairly a right to present that truth under circumstances, to a great extent, of the writer's own choosing or creation. If he think fit, also, he may so manage his atmospherical medium as to bring out or mellow the lights, and deepen and enrich the shadows, of the picture." Writers from James to Hemingway and beyond foreswore the managing of "atmospherical medium" in order to emphasize fidelity to the "probable" (partly, it must also be said, to bring fiction a perceived respectability that the Romantic dabbling in what Hawthorne also called the "marvellous" did not), and perhaps some of their influence does linger on.
One would think, however, that the work of writers like Beckett, Heller, Pynchon, and Sorrentino (among writers not classified as "genre" writers) would have demonstrated that "nonrealism" (I suppose that at this late date "romance" would sound too antique) is just as serious, just as "literary," as realism, psychological and otherwise. Even the relatively recent appearance of a variety of surrealism in the work of writers such as Aimee Bender and George Saunders finally suggests that "realism" as the attempt to depict "the probable and ordinary course of man's experience" is not really the sine qua non of acceptable literary judgment (although to some extent an indirect appeal to ordinary experience is still manifest in this fiction). As long as such fiction does not unsettle established conventions of craft and decorum too severely (as long as they still recognizably seem "like other books"), it can still be praised for its "imagination" and "invention." But Beckett and Sorrentino wrote books utterly singular in both content (the possibilities of "emotion and character" are strenuously pushed) and form, and it is this refusal to rely on mere formulaic craft, more than their rejection of realism, that keeps them on the margins of respectable opinion.