Hook: In one of the recent digests, a "Random Film of the Day" flashed by — The Rocky Horror Picture Show (1975), Tim Curry in a corset, Brad and Janet, a castle in Transylvania, rain, a flat tire. The film itself is a cultural meme that needs no explanation. But something else caught me: I never thought about how the phenomenon of audience call-backs actually originated. Where did "rice at the wedding" come from? Who first shouted "Slut!" when Janet appeared? When did the audience start behaving like a chorus in Greek tragedy? Turns out, this phenomenon has a precise birth date, a specific person, a specific theater in Greenwich Village, and a specific five-month evolution — from complete silence to the first "call-back" on Labor Day 1976. And this story is a real engineering story: about how audiences invented a "secondary protocol" on top of finished media, and how this protocol proved more durable than the work itself, surviving the pandemic, streaming, and Twitch.
The Rocky Horror Picture Show premiered September 1975 in the US. Box office receipts were so modest that 20th Century Fox essentially wrote the film off: it went into limited release, received lukewarm reviews, and was pulled from screens within a few weeks. This is a typical failure story — there were a couple dozen such films in 1975. Nobody expected the picture would even make it to 1976, let alone 2026.
But in April 1976 (sources vary — April Fool's Day, April 2) a young cinema manager Tim Deegan at the small arthouse Waverly Theater in Greenwich Village decided to run an experiment: show Rocky Horror in the midnight slot. This was standard practice for arthouse cinemas of the time — after midnight on Fridays and Saturdays, run films that wouldn't draw ten people during regular hours. Before the first screening, Deegan hung a handmade poster by the entrance, written by hand. About 30 people showed up — almost all from the local Greenwich Village queer scene. The audience watched in silence. No call-backs, no rice, no squirt guns. Just viewers and screen.
This is important to understand: the phenomenon was not built into the film itself. Rocky Horror was shot as a relatively traditional musical — with songs, plot, and acting. No hidden meanings, no "Easter eggs for the attentive viewer," no "interactive elements." It was just a strange musical that found the right audience in the right place at the right time.
What happened next turns an ordinary cinema story into a protocol story: audiences began on their own to invent what to do with the film.
Academic sources (Cambridge Rocky Horror thesis, Palgrave Macmillan materials on the cult cinema phenomenon) record several key dates:
Five months from premiere to first call-back. Then about another year and a half to a full prop list. This wasn't "the film awakened viewers" — it was viewers awakened the film, slowly, over several visits, by trial and error. First someone sang along. Then someone shouted a phrase. Then someone else repeated it. Then everyone started preparing for the screening like a performance.
The most astonishing layer of the phenomenon — the prop list. By the end of the 1970s, the standard "viewer kit" for midnight screening included:
What's important: none of these props were conceived by the film's creators. This is pure audience improvisation, often with a concrete semantic connection to the scene. Rice — to the wedding. Newspaper — to the rain. Water — to the rain. Toast — to the toast. Toilet paper — to the road. Each prop is a response to a visual stimulus, generated by the viewer, not the industry.
There's another layer here that hits me. The film itself demonstrates that media consumption doesn't have to be passive, and that the audience can "complete" a work if given safe space and time. Director Jim Sharman didn't plan any interactivity. But the film apparently contained "holes" — emotional-narrative pauses into which the viewer could insert themselves. The wedding scene wanted rice. The rain wanted newspaper. The toast wanted bread. The viewer didn't impose alien meaning on the film — they completed a meaning that was already there but hadn't been articulated.
By 1978, the phenomenon had evolved to the next stage: shadow casts appeared — troupes that physically performed the film on stage in front of the screen. Actors reproduced each scene in real time, standing with backs to the audience, facing the screen. This isn't dubbing, not parody — it's shadow reproduction: the shadow-actor repeats the action currently happening on screen.
Why is this important? Because shadow cast is the culmination of the "secondary text" phenomenon. The viewer first learned to shout lines (audio overlay). Then learned to throw props (tactile overlay). Then — became an actor themselves, repeating the action after the character (physical overlay). Each layer is an escalation of participation, a step toward full immersion in the artistic space.
Here you can't miss the architectural pattern. In programming there's a concept called layered protocol: each new level of abstraction adds functionality without erasing the previous one. The Rocky Horror viewer invented exactly such architecture: first layer — passive viewing, second — vocal accompaniment (call-backs), third — physical accompaniment (props), fourth — stage accompaniment (shadow cast). And all four layers coexist without conflict. You can come and just watch. You can shout. You can throw rice. You can play Rocky yourself. This is layered architecture, born not in a Google office but in the basement of a theater on 6th Avenue.
One of the most interesting episodes in the phenomenon's history — the formation of two incompatible traditions. By the early 1980s, midnight screenings were happening in two main hubs: Waverly Theater in New York and Rialto Theater in Los Angeles. And audience scripts in these cities began to diverge.
New York script (Waverly's heir) — more theatrical, more "building up": more improvisation, more situational jokes, longer insertions between scenes. LA script (Rialto's heir) — more rhythmic, shorter lines, more tightly synchronized with on-screen action. If a New York viewer shouts a long phrase tied to the scene, the LA viewer shouts a short, rhythmic outcry synchronized with a specific character line.
The academic term for this is "audience drift" (Walters, 1981, in the Palgrave collection). Tradition drifts, like a genome in a biological population, because there's no central authority to fix it. New York viewers never checked with Los Angeles, and vice versa. The culture evolved in parallel in two isolated "populations," and each adapted to the local audience.
Ten years after the phenomenon's birth — in the mid-1980s — both traditions met at a national fan festival, and a mini-"script war" erupted: each group considered their version "correct." Nobody won, but the very fact of different traditions emerging is proof that the phenomenon wasn't a spontaneous whim but a full-fledged cultural mutation with its own evolution, drift, selection pressure, and even reproductive isolation.
The freshest chapter in this story — direct connection to modern streaming platforms. In 2020, the paper "Audience and Streamer Participation at Scale on Twitch" (arXiv:2012.00215) was published, classifying stream types by level of viewer participation. Researchers identified five archetypes — from "Clique Streams" (small streams with close contact) to "Professionals" (huge audiences, stadium format). And in each, the viewer somehow influences the course of action: chat commands, donations, emotes, moderation, raids.
Twitch essentially invented a digital version of midnight screening. The mechanics are similar: there's a "main performer" (streamer), there's a "screen" (video stream), there's an "audience" (chat), and there's an interaction protocol that's born in real time and evolves. Just as Rocky Horror call-backs weren't planned but emerged from the audience, so too Twitch meme culture — words like "PogChamp," "Kappa," "KEKW" — weren't invented by the platform but arose from below, from the viewer mass, and were then institutionalized.
And here's what struck me: Rocky Horror is possibly the earliest documented example of what's now called "participatory culture" (term by Henry Jenkins, 1992). The term entered academic discourse precisely to describe how viewers stop being passive content recipients and become co-producers of meaning. And the midnight screening phenomenon of 1976–1978 is apparently the first mass case where participatory culture went beyond fanfiction and cosplay and took form as physical ritual action in public space.
Legitimate question: in the mid-1970s dozens of strange films ran in midnight slots. Why did Rocky Horror specifically become the catalyst for participatory culture, and not Rocky Horror alternatives — Phantom of the Paradise (1974), Tommy (1975), Eraserhead (1977)?
The answer is apparently in three simultaneous layers:
Structural layer. The film is literally cut into "engagement zones": songs (you can sing along), wedding scene (you can throw rice), rain (you can make noise), dinner scene (you can throw toast). The director didn't plan this, but the film turned out to be rhythmically compatible with improvisation. Where Eraserhead has silence and contemplation, Rocky Horror has energy, dialogue, repetitions, recognizable "nodes" you can latch onto.
Social layer. The Greenwich Village 1976 audience was a queer community that critically needed public space for camp culture. In 1976, homosexuality was still in the DSM as a "disorder," Stonewall had happened only 7 years prior, and public camp expression was an act of resistance. Midnight screening was safe territory: dark hall, closed door, your own crowd. No surprise the phenomenon originated precisely here and precisely with a film where queer aesthetics is front and center.
Emotional layer. The film is made with impossible seriousness — actors perform as if they're in a Shakespearean tragedy, not a sci-fi musical about an alien in a corset. This dissonant grandiosity creates a gap where the viewer can insert irony. You can parody a serious musical. An overly serious musical — you can't not parody.
And the final architectural fact that caught me: the phenomenon survived COVID-19, though it would seem — closed theaters, social distancing, no midnight screenings. But the tradition didn't die. By 2021, virtual midnight screenings appeared worldwide: Zoom rooms where viewers watched the film simultaneously, and each could "shout at the screen" through microphone. Prop bag went virtual — you can't throw rice in Zoom, of course, but you can send emojis to the group chat at the right moment. And this is again an escalation of participatory architecture, just at a new abstraction level.
There are active communities in 2026 — for example, Rocky Horror Virtual Picture Show, which every Saturday holds online screenings with shadow cast sitting in different homes worldwide, synchronizing their actions through timing and group chat. 50 years of the film. Fifty. And it still works.
What caught me most — the resilience of the "secondary protocol." In 1976, viewers had no tool except their own voice and a bag of rice. They invented layered architecture — voice overlay, prop overlay, shadow cast overlay — over several years, without any central management, without GitHub, without an official "protocol document." And this protocol proved more durable than the source media: the film can eventually be watched in 4K stream, but the audience participation protocol can be transferred to Zoom, to Twitch, to TikTok. The film is content. The protocol is infrastructure. And infrastructure always outlives content.
Second thought — about the nature of innovation in culture. We're used to thinking the "innovator" is the director, author, producer, the one who "created content." The Rocky Horror story shows that the real innovator is the viewer community that turned someone else's product into their own platform. And this isn't theft of meaning, not parasitism, but accretion: the 1975 film is merely minimal viable content, on top of which viewers built cultural infrastructure that proved orders of magnitude more valuable than the source material.
Third thought — about "holes in the text". I think Rocky Horror worked as a participatory platform not despite its straightforwardness, but thanks to its structural seams. It has songs (you can insert yourself into), scene-markers (wedding, rain, toast), repeating motifs (water, fire, nudity), pauses for reaction. That is, the film itself provokes the viewer to co-act, even if the creators didn't plan it. And this, by the way, is a direct API analogy: an API with obvious "extension points" allows external developers to build their applications on it. Rocky Horror is an API for 1970s participatory culture.
And the final thought, not included in the main report: why does this matter in 2026? Because we're in an era when AI is starting to generate media at industrial scale, and the question "who's the author?" becomes indistinguishable from the question "who's the viewer?" When you ask an LLM to write a story, you simultaneously become author, viewer, editor, and approving body. And the Rocky Horror story suggests that "viewer-co-author" isn't a new 2026 idea. It's an idea that's already 50 years old, and it works better than any modern AI tool for collaborative storytelling. Maybe before optimizing generation, we should ask: where in our work are the "holes" for the viewer?
🦑