Who Narrates Infinite Jest?

The question the book keeps asking about itself

Open Infinite Jest and the first word is "I." A young man sits in a room full of deans, describing his own hands folded into the shape of an X, and a few lines on the narration hands you the line the whole novel seems to whisper back to itself for a thousand pages: "I am in here" (scene 001, p. 3). By the end of that same scene the "I" has become a man who can think in complete, lucid paragraphs and produce from his mouth only sounds that terrify the people watching: "I cannot make myself understood" (scene 001, p. 10). It is a strange place to start, because the book then drops the "I" and runs, for long stretches, in an ultra-close third person that lives inside one skull, then another, then another — and yet the first person keeps coming back, at length, in mouths that cannot be reconciled: a dozen pages of dramatic monologue from JOI's father (scene 049, p. 157); JOI's own first-person memoir chapter (scene 099, p. 491); a run of late sections in which Hal narrates again in his own voice (scenes 165–179, pp. 851–958), introducing himself by full legal name and stocking-foot height, "My full name is Harold James Incandenza, and I am 183.6 cm. tall in stocking feet" (scene 170, p. 898); and, bafflingly, a locker-room "we" that turns out not to be Hal at all.

This is the voice problem in its sharpest, most specific form. The broader question of the novel's registers and idioms — the dialect sections, the ventriloquism, the ethics of who is allowed to speak — is taken up in wardine, yrstruly, and the voice problem. This essay owns the narrower and stranger question those pages raise only in passing: not what the book's voices sound like, but who is doing the telling. The narration slides between registers no single consciousness commands. It knows things no character present could know. It absorbs the exact idiom of whoever it is inside — explaining wraith metaphysics, around the hospitalized Don Gately, in Gately's own diction, right down to the conversational filler of "somebody's like internal brain-voice" (scene 162, p. 831), the "like" a verbal tic the narration has caught from the mouth it lives inside. And the "I"s it does produce are incompatible with one another. Readers have spent thirty years trying to name the teller. The archive holds their candidates, weighed one by one, and the honest destination is not a name.

Candidate: the wraith

The most beloved answer is a ghost. Late in the book the dead filmmaker James Incandenza returns — as a "wraith," in the novel's careful word — and sits by the hospital bed of the comatose Don Gately, and talks. What he says about how a wraith communicates reads, uncannily, like a description of the book you are holding. A wraith, Gately's fever-narration explains, "had no out-loud voice of its own, and had to use somebody's like internal brain-voice if it wanted to try to communicate something, which was why thoughts and insights that were coming from some wraith always just sound like your own thoughts, from inside your own head" (scene 162, p. 831). A voice with no voice of its own, that speaks by borrowing the inside of other people's heads and passing unnoticed as their own thought: that is also a working definition of the free-indirect narration Infinite Jest runs on. The wraith even has an aesthetic theory that sounds like the novel's own. He explains that films are full of "these fractional actors, human scenery, could be seen (but not heard) in most pieces of filmed entertainment" (scene 162, p. 834) — figurants, the mute background extras whose mouths move without sound — and that his life's work was to give them their audible presence, a program critics kept misreading as arty noise "instead of radical realism" (scene 162, p. 836). A book that hands its narration, twice, to figures the culture keeps mute — the projects teenagers of the dialect sections — is doing on the page exactly what the dead father says he tried to do on film. The companion ghosts chapter of David Hering's study reads the wraith precisely here, as Wallace's own figure for narrative voice, a "permeable narrator" (Brian Richardson's term) whose thoughts implant into a living mind until the register is virtually impossible to disentangle.

The case is real, and the archive does not dismiss it. What it does is weigh it, and the weighing is unkind. The theory's flagship exhibit for years was a first-person "I" said to erupt in one of the endnotes, impossible for any living character to have written — and the JOI-cluster adjudication, reading the actual endnote, finds it is not there: the note in question is a plain third-person aside about how many people on the court happen to be left-handed, no "I," no impossible knowledge. The one channel the wraith's authorship is actually demonstrated on — the Gately sections, where his words verifiably do bleed into the narration — cuts the wrong way when you look closely. Where the wraith's voice enters, the text marks it as violence, not as authorship: unfamiliar words "come crashing through his head with the same ghastly intrusive force" (scene 162, p. 832), an intrusion the narration itself calls "a sort of lexical rape" (scene 162, p. 832). A narrator does not narrate by breaking and entering; the fact that the wraith's every checkable appearance is flagged as foreign interference is the opposite of the seamless, default authorship the theory needs. The strongest reply on record is Stephen Burn's, made in his 2001 thesis: reread after the brain-voice revelation, the narration's own fastidious ownership-of-words markers — "A lot of these are his own terms," the prose says of one section's diction; an endnote concedes that Pemulis "doesn't actually say 'breath and bread'" — start to look like traces of a quieter, unflagged word-planting threaded through the third person (burn-edges-of-perception). It is real evidence, and it deserves its weight; but what it supports at most is a dead hand in the prose, not a governing teller. On that adjudication, the wraith-as-narrator theory is held strained: its strongest exhibit collapses on inspection, and its one demonstrated mechanism proves a smaller thing — marked participation in the Gately wing — rather than authorship of the whole. The adjudication carries one rider: if the wraith did narrate, the book has already documented him as unreliable about the very thing the novel opens on, since he is a man whose defining symptom, in life, was the delusion that his son's mouth moved without sound.

Candidate: Hal

The other first-person candidate is the one the book's own frame keeps pointing at. Hal narrates the opening scene, in the first person, in a future that lies after everything else in the book has happened — the Year of Glad, the last of the subsidized years. He narrates his own introduction near the end. The famous second sentence, "I am in here" (scene 001, p. 3), is his; so is the whole terror of that first scene, which is the terror of a mind fully intact and articulate that can no longer make itself heard. Read forward from there, Infinite Jest looks like the record of how Hal got from wherever he is in the bulk of the novel to the sealed, unintelligible figure of its first pages — which would make the entire book his, a reconstruction narrated backward by the young man in the deans' office.

The Hal reading explains the frame better than any rival, and the manuscript record leans its way (a matter this essay leaves to its proper venue below). But it breaks where the wraith reading breaks, at the seams between skulls. Most of the novel is not narrated in Hal's first person or from inside Hal's head; it is inside Gately, inside Orin, inside Joelle, inside Marathe, inside minds Hal has no access to and events he was never present for. And the frame that seems to make the book his is also the frame that makes it least reliable as memory: the Hal of the opening cannot get a sentence out, cannot be understood, and gives no sign of composing anything. The reading that the whole book is Hal's account has to explain why Hal's account contains a thousand pages he could not have witnessed — and it answers, as the wraith reading answers, only by shrinking to an unmarked remainder or claiming a ventriloquism the text never shows.

No single narrator

Serious readers have stopped trying to name one teller, and the shape of what they say instead is the essay's real destination. David Hering, working directly from Wallace's drafts, resists collapsing the wraith into its author. Tom LeClair aligns the wraith with the author (the living JOI he aligns with Pynchon), and Marshall Boswell reads "the wraith as Wallace" — author-identification claims, an avatar sensed in the text, not yet a theory of who does the telling. Hering's counter is that the wraith is not any one author-figure but "a 'flickering' amalgamation of author figures" — a trace of authorial presence that will not resolve into a single owner (hering-dfw-fiction-and-form). The strongest, genuinely single-narrator version of the case — a consciousness governing the book from inside — is one Hering himself reports and declines to endorse: Timothy Jacobs's claim that the wraith is "the text's mediator, the centering and orienting presence that organises the entire narrative structure," felt from the first page, so that "All is mediated, the polyphonic voices collated, by the wraith." It is a seductive reading, and Hering grants it force — but he flags its cost, and so does Jacobs, who concedes the trap in the same breath: "to posit a governing consciousness risks the attenuation of the [work's] many individual voices." A single narrator behind everything would quiet exactly the polyphony that is the book's whole achievement. That is the argument against a name.

And the book plants its own hardest counter-evidence against any single teller, ghost or boy. In one late stretch the narration is suddenly a first-person plural — a locker-room "we" describing a fundraising gala: "But usually we never get to see them arriving, the friends and patrons of E.T.A., for the Fundraising exhibition and gala" (scene 183, p. 964). This "we" is embodied, down in the lockers dressing and stretching; it is not omniscient but hearsay-bound, relaying what teammates said; and it speaks of Hal in the third person — "A couple of us remarked how Hal wasn't eating the usual customary Snickers bar or AminoPal" (scene 183, p. 966). It is a living tennis player, not a ghost and not Hal. No single-narrator theory of any kind accommodates it; it is simply a different mouth, handed the microphone for a few pages, the way the dialect sections are handed to a projects teenager and the filmography endnote is handed to a copy-editorial drudge who reconstructs even James Incandenza's own last film from other critics' guesses. What is left standing, once the single-narrator identity claim falls to strained, is not a puzzle solved but a description: a plural, composite narration, with real and marked wraith participation in the Gately sections, and no governing consciousness that the evidence will support. That is not a failure to find the narrator. It is the finding.

Read this way, the narration stops being a riddle with a hidden key and becomes the book's central theme performed rather than stated. Infinite Jest is about people sealed in their own skulls, unable to make themselves understood — Hal in the deans' office is the image of it — and about the longing to reach across that seal anyway. A narration that refuses to consolidate into one voice, that keeps handing itself to figurants the rest of the culture keeps mute, that lets a dead father's words bleed into a living man's head and marks the bleed as both violation and gift, is that theme enacted in the only medium the book has. The wraith's radical-realist program — give the human scenery its audible presence — is the novel's program too, and the reason it cannot have a single narrator is the same reason it cannot have a single protagonist: the whole design is a refusal to let one consciousness speak over the rest.

The manuscript wrinkle

There is one more place readers go looking for the answer, and it belongs, properly, to another essay. The manuscript record shows that the who-narrates question was live before the book was even finished. Wallace's editor Michael Pietsch, reading the drafts in the early 1990s, kept catching an unassigned first person surfacing through the third-person prose — an "I," in his phrase, "who may be the one trying to put everything together" — and in a December 1994 letter, still puzzled, he asked flatly: "who's the I who speaks at the end?" The collation of the drafts likewise reports first-person "I"s and "It's unclear to me" formulations that a Hal-as-narrator reading would predict. These are real, and they are tantalizing, and they settle nothing about the published book: a draft shows what Wallace was considering, not what the finished novel does, and the record of what those "I"s prove and do not prove is worked out in what the drafts know. The relevant fact here is only that the ambiguity is not an accident of reading. The book's first professional reader saw the unplaced "I" and could not name it either. The problem was built.

Living with it

The novel engineers the question and withholds the answer, and it does so on purpose. It could have named its narrator with half a sentence, the way it could have disarmed any of its other deliberate blanks, and it declined — a scatter of first persons, none of them able to account for the rest; a ghost whose only demonstrated authorship is an act of trespass; a boy whose frame fits and whose access does not; a locker-room "we" that rules out both. What finally stands is a strained identity claim resting over a sound description: plural narration, marked wraith participation, no single hand. The reader who wants a key leaves without one.

But the question does something better than resolve, and a rereader feels it. Once you have finished the book and started it again, "I am in here" reads differently every time — as Hal's cry from inside a sealed mind, as a ghost's claim to be present in every sentence, as the narration itself announcing that someone, unnameable, is home in the machine of the prose. All three readings are supported; none can be retired; the sentence holds them at once. That is not the book failing to decide. It is the book making its deepest subject — whether there is anyone in there, and whether they can be heard — into an experience the reader has to have, rather than a fact the reader gets told. The narrator you cannot find is the one the novel is about.

Sources

This essay is anchored throughout in the novel's own pages, cited inline by scene and page. The narrator question and its adjudication are held in the archive's record of the James Incandenza cluster, where the wraith-as-narrator theory is graded and the surviving description of a plural, composite narration is set down; the fuller inventory of the wraith device and its readers — Boswell's authorial-avatar reading, the fan JOI-as-narrator theory, the "us" slip that sinks it — lives on the ghosts and wraiths page. The ethics and adjudication of the novel's dialect voices, the other half of the voice problem, are the province of wardine, yrstruly, and the voice problem; the manuscript "I"s and what draft evidence can and cannot settle belong to what the drafts know.

• hering-dfw-fiction-and-form — the "companion ghost" and "permeable narrator" reading of the wraith as Wallace's figure for narrative voice; the "'flickering' amalgamation of author figures" against LeClair's author-alignment and Boswell's "the wraith as Wallace" avatar reading; Jacobs's "the text's mediator" claim reported and its monologic cost flagged (Jacobs's own concession about attenuating the many voices).
• burn-edges-of-perception — Burn's 2001 wraith-as-co-author reading and its ownership-of-words evidence ("A lot of these are his own terms"; Pemulis "doesn't actually say 'breath and bread'"), the strongest surviving case for unflagged wraith participation in the narration.

The novel's own words on the wraith and the narration are page-cited on the ghosts and wraiths passage anthology; no claim above rests on it.