What Wallace Said: The Author on Infinite Jest, 1996–2008

What this essay is

Wallace talked about Infinite Jest for twelve years, from the 1996 tour to the last interviews before his death in 2008, and he did not always tell the same story. This is a chronological account of every substantial claim he made about the book — what it was for, how it was built, how it ends, what it is and isn't about — with each claim set against the novel and against his other statements.

Two things organize the whole record. The first is that the 1996 statements are composition-fresh: he had just finished the book, and he speaks with the precision of someone still holding the blueprints. The 2000s statements are reconstruction: by 2003 he says on tape that he no longer remembers the book very well (paraphrased from audio), and even in 1999 he admits — in print — that the tidy accounts of what the novel is about were something he assembled afterward, in interviews, to have something to say. So the arc of this essay is also a warning about how to read it: the earlier he speaks, the more weight the statement carries, and he told us as much himself.

The second is that a novelist explaining his own book is a witness, not a judge. Where the archive has weighed a claim page by page, the verdict is recorded in caps (favored, strained, and so on). Wallace's intentions are one line of evidence among several; sometimes the pages back him, sometimes they don't, and a few times he is simply wrong about his own book.

Before the book: the 1993 program

The single most important thing Wallace said about Infinite Jest, he said before it existed. Interviewed by Larry McCaffery in 1993 while the novel was being drafted (the book is never named), Wallace laid out the entire machine in advance (mccaffery-1993-interview). Fiction's purpose: "Fiction's about what it is to be a fucking human being." Its mechanism against loneliness: serious fiction gives a reader access to another self, so that "we become less alone inside." His diagnosis of television, which the Entertainment would literalize: TV supplies "the facsimile of a relationship without the work of a real relationship." His account of irony, which the whole novel is pitched against: "Irony's gone from liberating to enslaving." He cited the image of "irony being the song of the prisoner who's come to love his cage."

Two 1993 statements bear directly on how the finished book behaves. He called good work a matter of the writer's willingness "To be willing to sort of die in order to move the reader, somehow" — the anti-seduction ethic behind a deliberately difficult book. And, glossing what post-structuralism left him, he said that once a text is done "the reader becomes God, for all textual purposes" — which, three years before publication, is already the theory of the ending: a book that stops and hands the reader the job of finishing it.

1996: the tour, and what he said the book was doing

"I wanted to do something sad"

Wallace's plainest statement of intent came to Laura Miller for Salon in March 1996 (miller-1996-salon-interview): "I wanted to do something sad." Not circumstantially sad: "It's more like a stomach-level sadness. I see it in myself and my friends in different ways. It manifests itself as a kind of lostness" — a condition, he said, of the privileged and well educated. He said the same thing to the Chicago Tribune's Mark Caro that spring: "I wanted to do a book that was sad. That was really the only idea that was in my head" (burn-conversations-with-dfw). The payoff he wanted was the one from the 1993 program, restated for a newspaper audience: great fiction is where "I feel human and unalone."

Not about drugs: the addictive continuum and the Failed Entertainment

Across the five-day Rolling Stone tour tape (published as David Lipsky's road-trip book, verbatim from the recordings — lipsky-becoming-yourself), Wallace insisted the book was not a drug novel. "drugs are kind of a metaphor for the sort of addictive continuum" the culture runs on; "if the book’s about anything, it’s about the question of why am I watching so much shit? It’s not about the shit; it’s about me." He named his own worst case: "I think my primary addiction in my entire life has been to television."

He also told Lipsky the working title, and it is the clearest thing he ever said about the book's form: "the original title was A Failed Entertainment. The idea is that the book is structured as an entertainment that doesn’t work." The novel is built to be the thing it warns against, and to fail at it on purpose. He extended the diagnosis into prophecy that has largely come true: "in ten or fifteen years, we’re gonna have virtual reality pornography." The images would be "given to us by people who do not love us but want our money" — and "We’re going to beg for it. We are literally gonna pay for it."

The one guardrail he put on all of this, and kept putting on it, is that the book is not a sermon against entertainment: "if the book comes off as some kind of indictment of entertainment, then it fails. It’s sort of about our relationship to it." Pressed on whether Marathe's argument for releasing the lethal cartridge was his own, he refused it flatly: "Steeply has the same argument. And it’s a terrible argument because Marathe is basically a fascist."

AA, religion, and the fear of sentiment

Wallace was consistent that the recovery material was not sociology. To Lipsky he said the most American thing now is "wanting to give yourself away to something. That I ended up thinking was kind of a distorted religious impulse. And a lot of the AA stuff in the book was mostly an excuse" — God-talk being hard to write after Dostoyevsky. On the radio with Michael Silverblatt that April he said much the same: the cultural sadness comes from a "loss of purpose or organizing principles, something you're willing to give yourself away to," and "the addictive impulse... is interesting and powerful only because it's a kind of obvious distortion of... a religious impulse or an impulse to be part of something bigger" (silverblatt-bookworm-1996). To Caro he put the research in one deadpan line — "If I hadn't gone to a bunch of AA meetings, I wouldn't have gotten rid of my TV" — research as self-diagnosis.

Miller drew out the corresponding claim about form: that banal, single-entendre truths — the AA clichés educated readers recoil from — "can actually be nourishing in a way that arch, meta, ironic, pomo stuff can't." That is the E-Unibus-Pluram program aimed at the reader's own defenses. To the WORD e-zine chatroom in May, he stated the enemy directly: "The worst thing about irony for me is that it attenuates emotion. 'IJ' is at least SUPPOSED to be both funny and very sad." And "the main function of contemporary irony is to protect the speaker from being interpreted as naive or sentimental" (word-ezine-1996-chat).

"Everything... on purpose": structure and the ending

Two 1996 statements underwrite every clue-hunting reading of the book. To Miller: "If it looks chaotic, good, but everything that's in there is in there on purpose." To Silverblatt, on the radio, he confirmed the draft was "actually structured like something called a Sierpinski Gasket" — after the cuts, "probably kind of a lopsided Sierpinski Gasket now" — and said flatly that "there's nothing in there by accident" (silverblatt-bookworm-1996). To Anne Marie Donahue at the Boston Phoenix he gave the musical version: "It may be a mess, but it's a very careful mess." And "It's really designed more like a piece of music than like a book" (burn-conversations-with-dfw).

On the ending itself, the bounding statement comes from the WORD chat, and it is the one the whole ending debate leans on: "there is an ending as far as I'm concerned. Certain kind of parallel lines are supposed to start converging in such a way that an 'end' can be projected by the reader somewhere beyond the right frame. If no such convergence or projection occured to you, then the book's failed for you" (his spelling, from the live chat). The last sentence is the sharp one: the projection is the reader's job, and its failure is a reading-failure. He gave Donahue the same idea as a metaphor of pitch: the book is "supposed to stop and then kind of hum and project. Musically and emotionally, it's a pitch that seemed right." And to Keats in the chatroom, asked which character is the book's "purely inactive" hero, he answered in one word: "Keats -- Hal."

The most precise thing he ever said about how much the ending resolves is not a public statement at all but a line from his correspondence with his editor Michael Pietsch, later printed in D.T. Max's reporting (dtmax-2009-the-unfinished). Defending the refusal to wrap things up, Wallace wrote: "We know exactly what's happening to Gately by end, about 50% of what's happened to Hal, and little but hints about Orin. I can give you 5000 words of theoretico-structural argument for this, but let's spare one another, shall we?" That is a designed asymmetry — Gately fully, Hal half, Orin barely — and it is the closest thing in the record to a specification of the book's endings.

The cuts

Wallace's own account of how much was cut from Infinite Jest is contemporaneous, specific, and does not reconcile with itself. To SPEAK magazine in 1996 he said the book "was cut by 400 pages or so. There were commercial cuts and literary ones. I listened, and cut out most of the literary ones" (speak-1996). To Lipsky the same year he did the arithmetic differently: "It was about five hundred pages longer. Of which four hundred unambiguously needed to go, and the other hundred was painful." By 1999 he told Lorin Stein for Publishers Weekly that Pietsch had cut "two or three hundred pages" (burn-conversations-with-dfw). The numbers wander with the retelling, a first sign of the memory problem this essay returns to.

He was firm about the endnotes as design, not decoration, telling Lipsky that "it’s a very different book depending on whether you read the endnotes or not." And, in a phrase that has annoyed genre-minded readers ever since, he insisted the thing wasn't quite a novel: "It’s not really a novel; it’s not supposed to be a novel."

1997–1999: consolidation

The 1997 tour for A Supposedly Fun Thing let Wallace restate the program at length. To David Wiley for the Minnesota Daily he gave the fullest oral gloss on "E Unibus Pluram": irony as a mode gone toxic — "The song of a bird who enjoys being in the cage" — while insisting the mode itself is not the enemy ("Irony itself is fantastic. It's one of the primary rhetorical modes") (wiley-minnesota-daily-1997). He guarded the reading the way he always did: "What I'm mostly trying to talk about is what it feels like emotionally to be thirty-four in this country" — a private citizen's report, not a manifesto.

Wiley also drew out two things new to the record. On the Entertainment's lethality, Wallace gave a concrete mechanism from science fiction — Larry Niven's wire-heads — and said the explanation had been cut: "wire-heads die. Because they don't eat or drink or anything. I mean it's like sustained orgasm." And "There's a lot of this stuff in [Infinite Jest], but a lot of it got cut out" (the bracket is Wiley's). And on the title, he did something the Hamlet-hunters tend to skip past: he deflated it. "it's supposed to be a long encomium to the dead father." And then: "So it's multivalent, but it's not particularly profound. I'm not very good at titles." The elaborate Hamlet mapping on that page is Wiley's own gloss, not Wallace's. On the wider Pynchon debt he was equally narrowing: the one Gravity's Rainbow borrowing he would certify is the word Brockengespenst in the Steeply/Marathe shadow scene ("that's an outright allusion"), while resisting the blanket comparison.

By 1999 the retrospective note has crept in. To Stein he described the AA material's real-world source — privileged people who "have ruined their lives and look like Dachau survivors. That is what Infinite Jest is about, in one sense" — but in the same breath admitted he only arrived at that account afterward, in interviews, needing a story about why he wrote what he wrote that was at least roughly true. He compressed the composition history to a formula — "a six-year novel that got written in three years" — and, having earlier called the book almost un-filmable, watched the film rights sell and hoped aloud the movie wouldn't get made.

The 2000s: distance and disclaimers

By the Oblivion years Wallace mostly declined to reopen Infinite Jest, and when he did he flagged that he was guessing. On Caleb Crain's 2003 tape he answered a close-reading question about Gately by confessing outright that he no longer remembered the book very well, and offered his reading as a guess (paraphrased from audio — crain-2003-tape). This is the statement that puts a standing caution over everything he said about the book's particulars after roughly 2000: the man explaining the novel is now, by his own account, a reader of it rather than its author.

Design intent survived the memory loss better than detail did. His clearest late statement on the ending came on Boston's WBUR in 2004, transcribed in print for Michael Goldfarb: "That book, actually, for me does resolve, but it resolves sort of outside the right-frame of the picture. You can get a pretty good idea, I think, of what happens" (burn-conversations-with-dfw). Asked about the novel's much-hated final sentence, he stood by it and enjoyed the reaction: "Yeah, I got hate mail about that," he said. The right-frame formula is the same one from 1996, worn smooth into a soundbite over eight years of being asked the same question — reassuring in its consistency, but by then a rehearsed answer, not fresh testimony.

The rest of the 2000s craft talk touches Infinite Jest only obliquely and should not be read as statements about it. In San Francisco in 2004 he described his taste for multi-strand, cross-cut structure — reaching, unprompted, for the word "figurant" for the extras a single-protagonist story reduces everyone to — and his recoil from the packaged commercial ending, but he was talking about Oblivion and about his practice in general; the interviewer named Infinite Jest only in passing (paraphrased from audio — city-arts-lectures-2004-kipen). The through-line of these years is the one he stated himself: his work was getting sadder, not funnier, and he had less and less to say about the big book behind him.

What holds up against the text

Weighed against the novel, Wallace's design claims fare well and his memory claims fare badly.

• "there is an ending... projected by the reader somewhere beyond the right frame" holds, in a calibrated form. The page-by-page weighing of the ending recovers the missing year's architecture in outline and grades that core favored, while ruling the year's endpoints and mechanisms undecidable by design — which is exactly the shape of Wallace's claim that an end exists and must be projected by the reader. He insisted an ending exists and refused to write it down; the text ratifies both halves.

• "Everything... on purpose" holds strongly. The constitutive-design claim — that the gaps are engineered, not accidental — is graded strongly-supported, with no page-cited counter-evidence anywhere in the archive. When Wallace said nothing was in there by accident, the pages agree.

• The Pietsch ration — Gately fully, Hal half, Orin barely — holds. Gately's survival-plus-relived-bottom is favored; Orin as the mailer of the Entertainment is favored (held inside the ending's undecidable-by-design flag); the Hal strand is the one left half-open, which is where Wallace put it. His private spec to his editor is the most accurate thing he ever said about his own endings.

• The Hamlet soft-pedal holds better than the reader-lore it corrects. Wallace's own downgrade of the title ("not particularly profound") sits closer to the text than the heavy Hamlet-structuralism readers layer on: the Hamlet frame itself is strongly-supported on the page (the skull is uncontroversially Yorick's), but whether the play is the organizing key or only the loudest frame among several stays open, with frame-among-several favored as the description of record — roughly where he placed it: a real frame, not a skeleton key. Readers argue this both ways — some treat the book as an American Hamlet, others reject the mapping outright — and the contradiction is itself part of the record; the one flat it-isn't-about-Hamlet denial on the record is a reader's, not his. But the author was, if anything, on the deflating side.

One further claim of his is worth marking against the pages the other way. He let a chat questioner's phrase stand and named Hal the book's "purely inactive" hero, a one-word answer in a jokey chat. The novel's actual verdict on what Hal is — damaged or newly, unbearably sincere — is left open, and the mechanism of his collapse is favored for withdrawal, not the drug plot fans reconstruct; "purely inactive" is a mood, not a diagnosis, and shouldn't be over-read.

What he misremembered or waved off

• The cut arithmetic never reconciles. "400 pages or so" (SPEAK, 1996), "five hundred pages longer... four hundred unambiguously needed to go" (Lipsky, 1996), "two or three hundred pages" (Stein, 1999): the numbers contradict each other and mix manuscript-page and book-page accounting. Pietsch later called the 500-page figure "a great exaggeration" and put his own suggestion nearer 250. Wallace was an unreliable narrator of his own editing.

• "Not about drugs" and the autobiography dance. He worked both sides at once — granting that Hal's age and appetites were his, that Kate Gompert's ward was "a quarter mile farther down" his own road, then warning readers off treating the book as a coded memoir. The grants and the denials are equally on the record; neither cancels the other.

• The self-account of his own breakdown. In 1996 he framed the McLean episode as "almost more of... an artistic and a religious crisis" than a breakdown, insisted "this wasn’t a chemical imbalance," and said antidepressants were "never an option." The posthumous record has him on Nardil from 1989 — the same year — onward. Both stand as spoken; they do not reconcile, and the tour-era Wallace is on record minimizing exactly what the later record centers. His heroin testimony is a smaller version of the same doubleness: a volunteered admission and a flat denial in the same conversation.

• The memory itself. By 2003 he said plainly he didn't remember the book well (paraphrased from audio). This is the disclaimer that should temper every late gloss: design-intent statements (the right frame, the failed entertainment) survive it, because those were decisions he made; close readings of particular scenes do not, because those he was reconstructing like anyone else. And the honest key to the whole archive of his talk is the thing he told Stein in 1999 — that the accounts of what the book is about were assembled after the fact, for interviews. He was, at his most candid, warning us not to trust him as the book's final authority. The pages are the authority; he is the best-informed witness who ever testified about them, and an inconsistent one.

Sources

• mccaffery-1993-interview — the pre-publication program: fiction-as-human, loneliness, the TV and irony diagnoses, the reader-as-God line. Everything the book does is stated here first.
• miller-1996-salon-interview — the plainest 1996 statement of intent ("I wanted to do something sad"; "everything... on purpose").
• lipsky-becoming-yourself — the deepest single well: the "Failed Entertainment" title, the addictive continuum, the cut figures, the endnote design, the Marathe-is-a-fascist refusal, and the 1996-vs-later biographical divergences.
• word-ezine-1996-chat — Wallace's own typed words: the canonical "beyond the right frame" statement with its failure clause, the irony diagnosis, "Hal" as the inactive hero.
• speak-1996 — the earliest primary figure on the cuts ("400 pages or so") and the E-Unibus-Pluram program claimed for the book by name.
• wiley-minnesota-daily-1997 — the fullest oral gloss on the irony thesis; the wire-head lethality mechanism; the title deflation; the bounded Pynchon accounting.
• burn-conversations-with-dfw — the scholarly anthology: Donahue's "careful mess" and hum-and-project pitch, Caro's book-that-was-sad line, Stein's retrospective-construction admission, Goldfarb's 2004 right-frame answer.
• silverblatt-bookworm-1996 — the Sierpinski-Gasket structure, the nothing-by-accident assurance, and the addiction-as-distorted-worship gloss (radio, quoted from the fan transcript).
• crain-2003-tape — the 2003 tape where Wallace says he no longer remembers Infinite Jest well; the source of this essay's standing caution on late statements (paraphrased).
• city-arts-lectures-2004-kipen — Oblivion-era craft talk (figurants, cross-cutting, anti-commercial endings), general rather than IJ-specific (paraphrased).
• dtmax-2009-the-unfinished — the Pietsch correspondence, with Wallace's private specification of how much each strand resolves.
• mcnally-making-of-ij-nadell-pietsch — Pietsch and Nadell on the making of the book: Pietsch's "great exaggeration" correction of the 500-page cut figure and his own nearer-250 estimate.
• the-ending, structure-and-hamlet, hal-collapse, addiction-and-recovery — the theme pages where the claims above are weighed against the pages.