Indeed, there are times when the license of fiction, sparingly employed in the service of nonfiction, results in a great book with no negative effects on the lives of those involved. Old Mr. Flood seems precisely such a book. For one thing, no one real person is defamed in Mitchell's composite of Flood (unlike Stephen Glass' untruths about Vernon Jordan). Nor does Mitchell's use of a composite detract from the realism of Old Mr. Flood's compassionate, elegant, reportorial portrait of the Fulton Fish Market. Like a novelist, Mitchell takes license with dialogue in order to dispense with some of the ancillary randomness that is part of everyday life and arrive at a more highly stylized portrait. The quotes in the Old Mr. Flood are models of eloquent compression, such as you rarely find in real life and usually find in fiction. The point? To create a work that provides more aesthetic pleasure than a less highly wrought one, a distillation that makes us feel something essential about the world described, and thus has a greater chance of being remembered, read, used.
After all, unlike newspaper stories, literary journalism seeks to make or "conjure up" a broader reality—to bring us into a world. This isn't news of the who-what-when-how-why variety, but news of the kind that V.S. Naipaul said only the novel can deliver—news that resonates with the potency of its presentation. Strictly segregating fact from fiction hobbles literary journalists unnecessarily. Where fiction is an inclusive genre, one that allows for its conventions to be violated, journalism relies on a system of conventions intended to guarantee objectivity. But clearly even these conventions don't make for pure objectivity, which from the start compromises the sanctity of the fact/fiction opposition. (Consider recent studies suggesting that firsthand witnesses often remember events differently, calling into question the reliability of some journalistic witnesses.)
Fine, you might say; the use of storytelling technique is important to journalism. But why extend that to inventing dialogue or deceptively conflating characters? Because everyday speech and the limitations of reporting a story don't lend themselves to the kind of compression of narrative, image, insight, event, and character that Mitchell, Liebling, and, say, Ryszard Kapuscinski have been able to get down on the page.
This does lead to complicated, messy questions. Take Kapuscinski's The Emperor.A chronicle of the last years of Haile Selassie's dictatorship in Ethiopia, The Emperor is a deeply memorable mythopoetic vision of absolute power; novelists like Rick Moody gush about it. The Emperor is also full of inaccuracies about Ethiopia. Some appear to be intentional, surreal exaggerations; others seem simply like sloppy reporting. After the book's publication, many readers were understandably indignant that these inaccuracies might be accepted as historical fact. Nevertheless, The Emperor succeeds in depicting Selassie's corruption with a rich concentration that transforms it to near-allegory—as Kapuscinski now belatedly suggests he intended it to be read.
So, the salient question is: Why didn't Kapuscinski simply publish The Emperor as a fact-fiction hybrid, with a prefatory note? Why didn't Mitchell append a note to the Old Mr. Flood stories in The New Yorker? It's possible that our fixation on fact as highest good and on maintaining the antipodes of journalism and fiction has created a problematically rigid division of genres—one that may encourage writers to lie (and then later come out of the closet, sheepishly, as genre-benders). Understandably, we think we value magazine pieces and nonfiction for their factual truth. But surely the impact of literary journalism derives in part from aesthetic intelligence and authoritative vision. Mitchell's Old Mr. Flood is a world you want to read about not because it's utter fantasy but because it seems real—in fact, it's a world that seems more real, more pressing in its moral accounting than those you find in many well-documented but dull examples of magazine journalism.
So, perhaps the problem is partly that our culture has no label for this kind of work, and that, systematizing creatures that we are, we need labels. Maybe we even need a new magazine genre, somewhere between fact and fiction. As for how and when it ought to be used, the only way to determine the answer would be on a case by case basis; in large part it depends on how worthwhile the result is. A system that asks writers to evaluate their own self-worth (in advance) is not a simple one; take the fact that Truman Capote's rigorous notion of a factually accurate nonfiction novel has quickly given way to a less well-enforced sub-genre, one example of which is Maria Flook's new book about Christa Worthington, a journalist murdered on Cape Cod in 2002. But such a system is theoretically feasible: Fiction writers pillage the lives of friends all the time; we tend to shrug off the negative consequences when the result is Saul Bellow's Herzog or a Robert Lowell poem. Certainly when in doubt, a journalist should assume it's not OK to take licenses like those described here; they're tools to be used rarely. But let's not take Mitchell off the syllabi because other writers lack his judiciousness and talent.