Thirteen Ways of Looking at a Film That Won't Sit Still

For half a century, the greatest film ever made was Citizen Kane. This was never a fact about the film; it was a fact about a ballot. Once a decade since 1952, the British magazine Sight and Sound has polled the world's critics, and from 1962 through 2002 the same title held the top of the result so reliably that its supremacy began to feel less like an average of opinions than a property of the universe. Then, in 2012, it was deposed by Hitchcock's Vertigo, a picture filed under minor on its release in 1958. And in 2022 the top slot went to Jeanne Dielman, 23 quai du commerce, 1080 Bruxelles — Chantal Akerman's three-and-a-half-hour study of a widow peeling potatoes and keeping an apartment, a film that had ranked thirty-sixth only ten years before, and that, on the morning of the announcement, several established critics declared an affront to the very idea of a canon. The director Paul Schrader suspected ballot-stuffing. The critic Armond White called the winner "a dull Marxist-feminist token." None of the films had changed. What had changed was the pressure around them.

This is the fact our cultural instruments are built to obscure. Rotten Tomatoes renders a film as a percentage of critics not actively displeased. Metacritic weights and averages. IMDb tallies a vote from one to ten; Letterboxd resolves a million private reckonings into a number with two decimal places; a streaming service promises you a "98 percent match" and means it as a kindness. Each of these instruments performs the same operation: it takes a living disagreement and reports it as a settled figure, engineered to look like the end of an argument rather than a snapshot of one still in progress. They answer the approval question efficiently, and then they treat the answer as a conclusion.

But approval is close to the least durable thing a work possesses, and close to the least revealing. A film can be widely approved and culturally inert — admired, shelved, and never thought of again. Another can be booed off the screen at Cannes and, three decades on, sit on university syllabi as the most sustained account of grief in American media. The score cannot tell these apart, because the score was never measuring the right quantity. It measures whether a verdict was reached. It does not measure what the work goes on doing after the verdict is filed — whether it lodges in the body, splits its audience into warring camps, returns to people in dreams, breeds imitators, or quietly rewrites its own reputation a decade late.

The Artifact Index proposes that this residue — call it cultural pressure — is both measurable and more worth measuring than approval ever was. It reads the open record: the three-in-the-morning forum thread, the Letterboxd diary kept across years, the marginalia of an academic essay, the half-life of a meme. None of that is truth. All of it is evidence of how an object is being lived with. From that evidence the Index derives thirteen axes. None is a score; none is a rank; there are no stars. Each, in the project's own phrase, "is a language before it is a number" — a claim stated in words, which you are invited to dispute. Wallace Stevens needed thirteen ways to look at a single blackbird. A film under genuine cultural pressure tends to require at least as many, and for the same reason: the object holds still for none of them.

What follows is a tour of the instrument, grouped into four movements — the geometry of a verdict, the hold a work takes on the people who let it in, the work considered as a system and as a frontier, and the long wake it leaves in the culture downstream.

Before a work can do anything stranger, it produces a verdict, and the verdict has a shape. The first three axes describe that shape — not whether the room liked the film, but how the room's opinion is distributed: whether it has converged, whether it is still at war, and whether the door was open to begin with.

01 · CNSConsensus

Consensus measures how aligned the public reading is across critics, audiences, and the diaristic record — how much a single, shared account of a work has been agreed upon. It is the axis closest to what conventional scores already track, and for that reason the easiest to misread. A high consensus reading is not an endorsement. It is a description of stability: the interpretive argument has concluded, and the culture has arrived at a resting temperature on the object. Casablanca reads high here; so does the cohort of films Roger Ebert gathered into his "Great Movies" essays, works whose greatness functions less as a claim than as a piece of shared furniture. The danger the axis is built to flag is that stability and exhaustion look identical from the outside. When everyone agrees on what a film is, it often means the work has been settled — filed, canonized, and quietly stopped being watched with any heat. The sociologist Pierre Bourdieu spent a career demonstrating that a shared verdict on a cultural object is frequently a shared performance of belonging rather than a shared experience of the thing. Consensus, in other words, can be the sound a work makes when it has been agreed upon prematurely.

02 · FRCFriction

Friction is the inverse measurement: the width and durability of disagreement around a work. It is crucial to distinguish it from controversy at release, which almost every notable film generates and almost every film survives. Friction registers only the disputes that fail to close — where the gap between readings is still open years on, structural rather than superficial, and no interpretation has won. Aronofsky's mother! earned a rare F from CinemaScore polling while a wing of the critical establishment defended it as a masterpiece; that split has not healed. Kubrick and Spielberg's A.I. opened a fault line — was its ending a cruelty or a consolation? — that runs through its audience to this day. The Jeanne Dielman result is itself a friction event: a decade after the announcement the argument it started is still live, which is precisely the evidence the axis is looking for. The literary scholar Hans Robert Jauss called the frame a reader brings to a work its "horizon of expectations"; friction is what you measure when a work shatters that horizon and the pieces never reassemble into agreement. A high friction reading does not mean a work is divisive in the marketing sense. It means the work refuses to be assimilated.

09 · ACCAccessibility

Accessibility measures how available a work is to a viewer who arrives without context — no glossary, no prior initiation, no homework. It belongs in this movement because it shapes the verdict before the verdict is formed: a film that anyone can enter will be judged by everyone, while a film that demands preparation is judged mostly by the prepared. The axis is deliberately inverted in the Index's geometry, because it correlates negatively with nearly everything else the instrument values. Jeanne Dielman asks for two hundred minutes and a tolerance for real-time domestic tedium; Tarkovsky's Stalker asks for patience as a precondition of meaning. A Pixar feature or a Spielberg adventure, by contrast, hands the arriving stranger everything they need at the door. The Index is careful to insist that high accessibility is a feature, not a defect — an act of generosity, a refusal to treat the newcomer as an unprepared student rather than a full participant. But it is also the axis along which a work most often trades reach for depth, and the one that most reliably runs opposite to the readings that follow.

A verdict is what a culture says about a work in public. The hold is what the work does to a person in private — and then, sometimes, to a congregation. These four axes measure attachment: how long a work stays, how it returns, how hard it strikes, and whether it gathers a people around itself.

03 · OBSObsession

Obsession measures mention density across time — but with the recency curve deliberately inverted, so that an older work still being discussed scores higher than a new one briefly trending. This inversion is the methodological heart of the axis. It is built to tell the difference between fame and obsession, two states that volume alone cannot distinguish: a film can be enormously famous and entirely un-obsessed-over, consumed once and dropped. What the axis rewards is the work being lived with — the picture people return to for a fifth and tenth viewing, that accumulates rather than exhausts meaning. The theory threads that still spool out under Mulholland Dr.; the Letterboxd diaries that log In the Mood for Love on an annual cycle, like an anniversary. The media scholar Henry Jenkins described devoted audiences as "textual poachers," users who take up residence inside a work and refuse to leave. Obsession is the axis that registers the squatters — the readers for whom a film is not an event that happened but a place they keep going back to.

04 · HNTResidual Haunting

Residual Haunting is the most superstitiously named axis in the instrument, and, the Index notes, the most consistent across genres. It measures the frequency of language indicating that a work returns to viewers without invitation: that it surfaced in a dream, intruded during an unrelated moment, would not leave. Where obsession is a relationship the viewer chooses to maintain, haunting is one imposed on them — the work, in the project's phrase, has "installed something the viewer did not authorize." Hereditary sends people home with a dread that outlasts the plot; Don't Look Now leaves an image lodged behind the eye for years. The effect has a long theoretical pedigree — Freud's uncanny, the familiar made strange — but its most useful contemporary diagnostician is the critic Mark Fisher, who borrowed the term hauntology to describe culture stalked by the presence of what should be absent. Haunting is what the instrument detects when the discussion around a film stops sounding like criticism and starts sounding like a symptom: I can't stop seeing it. It comes back when I close my eyes.

08 · VLTEmotional Voltage

Emotional Voltage measures the sustained intensity of reaction to a work, regardless of polarity. Crying, panic, awe, nausea — the axis does not care which, only how much current is running and for how long. This indifference to sign is the point, and the most common misreading is to treat voltage as a proxy for quality. It is not. A bad film can be high-voltage; the instrument measures the current, not the source. Requiem for a Dream leaves viewers wrung out; Grave of the Fireflies devastates on a schedule; Uncut Gems induces a genuine, sustained anxiety in the body before the mind has constructed any thesis about it. The film theorist Linda Williams once grouped horror, melodrama, and pornography as "body genres" — forms whose success is measured not in interpretation but in whether the spectator's body does what the screen body does: flinches, weeps, recoils. Voltage is the axis for the body genres and for any work that earns the right to mean something by first moving the viewer who has not yet decided what to think.

06 · CLTCult Formation

Cult Formation measures the slope of devotion — the rate at which a committed community assembles around a work, and especially the steepness of that curve after an institutional rejection. The instrument does not distinguish between works born cult and works that became cult only after being abandoned; both produce the signal it tracks, which is collective rescue. The Rocky Horror Picture Show flopped, then mutated into the longest-running theatrical engagement in history, its audiences supplying a participatory liturgy the film itself never scripted. The Room converted its own incompetence into a midnight rite. Showgirls was reclaimed first as camp, then, by a serious minority of critics, in earnest. Blade Runner gathered its first true congregation on VHS, in the years before the official reappraisal caught up. Umberto Eco, writing on Casablanca, argued that a cult object must be "ramshackle" enough that its devotees can inhabit and quote it; Susan Sontag's notes on camp mapped the structure of affection for the failed-magnificent. What unites the readings high on this axis is that loving the work has become an act of membership — a way of belonging to the argument that the initial verdict was wrong.

The first two movements measured how a work is received and held. The next three turn to the object itself — to what it is made of and how far it was willing to go. They ask whether a film can be read as a system, whether its form took a real risk, and whether its content pressed against the limits of the permissible.

05 · SYMSymbolic Density

Symbolic Density measures the volume of interpretive activity a work sustains — theory threads, motif-tracking, disagreement about meaning, the long forum post that begins what if the whole thing is about. It registers whether a film is being read as a system rather than consumed as a story. The monolith in 2001: A Space Odyssey has generated more exegesis than most national literatures; the whole "ending explained" economy of contemporary criticism is symbolic density made into a content genre. Crucially, this axis is independent of consensus: a blockbuster can score very high agreement and near-zero symbolic density, because nothing in it asks to be decoded. Roland Barthes distinguished the "readerly" text, which the reader passively consumes, from the "writerly" one, which conscripts the reader into producing its meaning; symbolic density is the measure of the writerly. It is worth noting that Susan Sontag warned, in Against Interpretation, that the hermeneutic reflex can bury a work's sensory force under a heap of meanings. The axis takes no side in that quarrel. It simply counts whether the heap is forming.

07 · FRMFormal Risk

Formal Risk measures the risk a work took in form, structure, sound, or image — and, decisively, it measures the risk independent of whether that risk paid off. A film can be formally radical and emotionally inert; this axis says nothing about love, only about the refusal of a known shape. Jeanne Dielman stakes everything on duration, daring the viewer to watch domestic labor unfold in something near real time. Russian Ark compresses an entire feature into a single unbroken Steadicam take through the Hermitage. Eraserhead and Fire Walk with Me treat sound design as something closer to theology than to atmosphere. The axis is built to be careful here, because formal risk is the property most often confused with the others: with difficulty (which is accessibility inverted), with meaning (symbolic density), with danger (transgression). A film can rearrange the grammar of cinema and remain perfectly decent and legible while doing it. What this axis detects is only the moment a work could have taken the known shape and chose another — and the size of the bet it placed on that choice.

13 · TRXTransgression

Transgression measures the degree to which a work's content operated at or beyond the social, moral, or political limits of its moment — the extent to which it was considered not merely difficult but dangerous. The distinction from formal risk is the entire point. A formally conventional film, shot in plain language, can be the more transgressive object; a formally radical one can be perfectly safe. The instrument reads transgression in the institutional record of alarm: censorship, prosecution, the X rating, the moral panic, the content warning that becomes a news story. Kubrick withdrew A Clockwork Orange from British distribution himself, and it stayed effectively unavailable in the country until after his death. Pasolini's Salò remains a standing test of what a state will permit; Last Tango in Paris carried its scandal for decades, and then acquired a second, graver one. The director of Cannibal Holocaust was hauled into court to prove he had not actually filmed murder. These are not measures of quality, and not all of them have aged into respectability. The axis records only the historical fact that the culture, at some point, treated the work as a threat — and the size of the threat it perceived.

The final movement measures what outlives the encounter — what detaches from the work and travels. A film's wake spreads in three directions: outward, into a culture that need never have seen it; forward, into the works it makes possible; and through time, as its own verdict is revised. These are the slowest readings the instrument takes, and often the most revealing.

None of the films had changed. What had changed was the pressure around them.

10 · RCHReach

Reach measures how far a work's imagery, references, and language have spread into the general cultural record, beyond any audience that has actually seen it. Its governing distinction is that reach is not depth. A work can achieve enormous reach on a single image while remaining, as a whole, largely unwatched. Almost everyone can mime the Psycho shower stabs or hum the two notes of Jaws; "Here's Johnny" and the elevator of blood circulate among people who have never sat through The Shining; the silver robot of Metropolis is a fixture of the visual culture decades after the film itself receded from view. Richard Dawkins coined "meme" in 1976 for exactly this phenomenon — a unit of culture that replicates from mind to mind, shedding its origin as it goes. Reach is the axis that tracks the memes a work sheds: the fragments that escape the medium and become part of a shared language available even to those who never encountered the source.

11 · PRGProgeny

Progeny measures documented generative influence — how many later works explicitly cite a film, how many aesthetics it defined, how often it surfaces as an originating reference. It is the forward-looking complement to reach: where reach tracks fragments escaping into the general culture, progeny tracks lineage running into other art. Blade Runner is the clearest case in modern cinema; for forty years, nearly every rain-slicked, neon-drenched future on screen has been in conversation with it, acknowledged or not. The Odessa Steps sequence of Battleship Potemkin has been restaged so often — most famously in The Untouchables — that the homage now travels further than the original. Seven Samurai seeded a genealogy of remakes and reworkings that has not stopped branching. Harold Bloom argued that strong art is made under the "anxiety of influence," every new work wrestling with the ancestors it cannot escape; the art historian Aby Warburg built an entire atlas to chart the Nachleben, the afterlife, of images as they migrated across centuries. Progeny is the deliberately conservative measure of that migration: it counts acknowledged inheritance, not silent borrowing, because influence is most often uncredited and the instrument prefers to under-claim.

12 · ARCCultural Arc

Cultural Arc is the axis that the Sight and Sound story illustrates best, and the one most easily confused with the others. It measures neither current esteem nor current volume of discussion, but the distance between a work's original verdict and its present one — the degree to which time has revised the reading. A film beloved at release and beloved ever since scores low here, however great it is, because its reputation never had to move. The high scores belong to the works whose reception had to be corrected. Blade Runner earned roughly thirty-three million dollars against its budget and a shrug from critics in 1982, then was rebuilt by a leaked director's cut and a 2007 restoration into a foundational text. It's a Wonderful Life underperformed in 1946 and lapsed into copyright limbo, only to be resurrected by decades of public-domain television broadcasts into a national ritual. John Carpenter's The Thing was savaged on arrival and is now studied as a masterpiece. Walter Benjamin wrote of the "afterlife" of a work — the Fortleben in which its meaning continues to be produced long after its maker has finished. Cultural Arc is the instrument's measure of afterlife: not whether a work is currently admired, but how far it has traveled from the place it started, and therefore how unfinished the verdict on it always was.

Set the thirteen readings side by side and a work stops being a point on a scale and becomes a shape — a profile of pressures, spiky in some registers and flat in others. Twin Peaks: Fire Walk with Me, booed at Cannes in 1992, reads low on consensus and accessibility and very near the ceiling on haunting, cult, voltage, and arc: the signature of a work the culture rejected and then could not stop returning to. A competent, beloved, four-quadrant hit produces nearly the opposite profile — high consensus, high accessibility, high reach, and almost nothing on friction, haunting, or arc, because nothing about it remains unsettled. Neither shape is a grade. They are different kinds of life.

That is the whole wager of the instrument, and the reason it refuses to issue a number you could put on a poster. A score asks is it good? and returns an answer that was stale the moment the argument it summarized moved on. The thirteen axes ask a harder and more honest set of questions — what is this doing, to whom, and for how long? — and accept that the answers keep changing, because the changing is the phenomenon. The map does not tell you what to watch. It tells you how a work is being held: as a settled classic or an open wound, a private haunting or a public language, a finished verdict or one still, decades on, being written.