Style as Variable: A Very Long Carriage Ride
From Generative Animation: Volume 4 of The New Machine Cinema
The Living, Breathing Cinema proposed that a film could survive translation across multiple valid forms while retaining a deeper identity beneath its surface. It, in a sense split the cinematic atom. What is the real movie when there are two equally valid instantiations of it?
The claim was proven with 2025’s A Very Long Carriage Ride. One film in two ways gave us two animation ontologies at once: a classical 2D version and a stop-motion version. The screenplay, performances, and music were the exact same in both version, only the visual style changed.
In generative animation, style has been detached from process. A film is simply what it is, with no extra thought going toward alternate possibilities not on screen.
This represents a historic break in the medium.
For most of animation history, style arrived fused to process. Hand-drawn animation looked the way it did because drawings had to be drawn. Stop-motion looked the way it did because objects had to be held and adjusted. CGI looked the way it did because rigs, shaders, and rendering pipelines imposed their own order on the image.
Audiences learned to perceive those styles as aesthetic identities. The look of the film and the process of the film seemed joined at the root. The style felt like the film’s DNA.
A Very Long Carriage Ride broke these inherited assumptions. The point of the dual release was that style matters so deeply that changing it bends the urgency of what’s on screen, and the dramatic weight around it.
It became impossible to watch one version without having an opinion about the other one.
What Carriage Ride disproved was the belief that changing style necessarily produces a different film in the categorical sense. The directorial core endured, regardless of the images on screen. This finding changes the relationship between audience and film.
The Experiment
A Very Long Carriage Ride represents theory through practice. The film functioned as a controlled artistic experiment: how much of a film survives when its visual ontology is completely replaced while its deeper architecture remains?
The 2D version had more fantasy, more of an immediate storybook feel. The world felt vibrant, nostalgic, and in the broad fairy-tale sense, familiar. The viewer entered a fable.
It felt like that a classical film from the distant past or a parallel history.
The stop-motion version moved in another direction entirely. The same narrative suddenly gained weight. The film felt more fragile, more melancholy, more real. The stop-motion body placed the characters inside a visible world of crafted things and that craftedness carried the film’s dark mortality within it.
The feeling of a fairytale remained, it just felt darker. Drama rose to the surface.
Both counted, they just landed in separate ways. The question became what remained invariant, and what changed under the pressure of style.
The Anatomy
Autumn and Katy’s long traveling sequences across Goosebury reveal the principle of the dual release. In stop-motion the journey feels like journey, like staring into a hand crafted diorama. The 2D version cannot ever hope to capture in the tactile. Every detail on screen feels hand-placed.
The 2D version feels more evocative or dreamy. It sits in more of a comfortable way for the viewer’s eyes. We believe it.
One version makes the journey a discovery of place; the other makes it a passage through a world of enchantment.
It’s the same journey in both films: they point out the same stops on the same maps. All directorial intentions hold in both versions. However, the experience is now split in two.
If I had to sum it up, it is the difference between a novel and theater. Between writing and performances. Between fantasy and drama. What I speak of transcends aesthetic, going into the essence of expression itself.
This was most encapsulated in the reunion sequence with Hopkins Brandy; Autumn and Katy team up with the disgraced business man Brandy kickstarting the second half of the film where they travel to expose the evil magnate Meijor Donaldson. In the 2D version the reunion unfolds with a ghostly, haunting detachment. The line work and soft watercolor-like washes create a Dickensian London atmosphere as if directed by Walt Disney.
Disney is the obvious comparison, but the New York elegance via grit feels more like the Fleischer Brothers, the adult emotions more like Bakshi.
And yet it may still arrive at Disney in the end: fogged streets, spectral lighting, all bring about a literary melancholy on screen.
The emotional register is one of wistful remembrance, the kind of haunting that feels safely contained within the aesthetic tradition of classic illustration. In the character Hopkins Brandy, I was inspired by the great British theater actors.
The same reunion in the stop-motion version is grittier, more physical, more devastating. Stop motion always fared edgier, with directors like Starevich, Rankin and Bass, or Selick, and Carriage Ride fared no different. You feel every line without any detachment from aesthetic. The reality of puppets force the matter: you are looking at real things and hearing real things.
The grief lands in the body. Whereas the 2D version filters the emotion through fantasy; the stop-motion version makes its case.
Nowhere is the divergence more instructive than in the gathering of the victims of Meijor Donaldson in the second wedding at Roland Bay, and the ensuing argument between Miss Wish and Miss Hart.
In the stop-motion version the scene is inescapable and uncomfortable. The way their small bodies occupy these startling emotions. The way their painted eyes catch the light in such a haunting, true way. The scene is raw and brutal. There is no aesthetic buffer. The horror and the pity sit in the frame without escape.
In the 2D version, it simply feels like a scene in a movie. The shock of it is dulled. This has a benefit because a movie should play as a movie. It perhaps shouldn’t escape out the seams of the screen and get underneath your skin. There is something safer about 2D animation, even as it goes into darker places.
The animation conventions, the graceful arcs, the stylized expressions, the painterly atmosphere, all provide a protective coating to create a more comfortable, universal viewing experience. Through the 2D animation, I learned how aesthetic offers the director safety. I saw exactly how and why Disney and the Fleischers caught on via their aesthetics: there is no barrier to entry. We as the viewer, are home.
The style alone recalibrates the weight, to where aesthetic comes before the movie itself. When you have a great aesthetic, it almost doesn’t matter what the movie is underneath it, long as it doesn’t get in the way of the aesthetic. And the goal with Carriage Ride was to not allow a film to rest on any aesthetic shorthands. Meaning the goal was to extend beyond where 2D or stop motion succeeds innately.
The goal was not to make an AI Disney film, or an AI stop motion; it was for both to remark on the other, and to re-contextualize the medium of AI, of cinema itself. To allow for the first demonstration of a film that could not or would not be done traditionally, an artform unique to AI itself.
Stop-motion refuses to let us look away. 2D whisks us away.
One suspenseful set piece in the film depicts a mountain ascent with Cousin Martini (played by Rogers, the most enduring actor in my company.)
The scene builds as a Hitchcockian crescendo: was Autumn pushed out the carriage? Wasn’t she?
The carriage climbs through a storm, and the question of Martini’s intention weighs on us. There, Autumn is drunk on a magic bean which offers the trait of mirth, creating suspense as we the audience realize she doesn’t know the danger she is in.
The camera alternates between the point of view at the rain-streaked window, and the steep drop of the cliffs below with the crashing of waves. Martini says eerie things as he comes closer behind her, while she keeps speaking playfully while the suspenseful music is swelling.
Each cut tightens the uncertainty, making the danger register as spatial and immediate. The viewer feels the height because the stop-motion body has given the environment a physical logic.
In the 2D version, like the drama between Miss Hart and Miss Wish, the scene blends in. The same alternating structure is present, the same escalation of music and dialogue. Yet we are still enjoying it a little too much to surrender to the danger.
This in its way creates a profound structure in viewing. One considers why the most enduring Disney films are not actually the wholesome ones. They’re the dark ones which counter the 2D animated imagery with genuine bleakness. Consider Beauty and the Beast, or the Hunchback of Notre Dame. The aesthetic doesn’t soften the pathos, it contextualizes it in a frame. We admire the film more because it pays us the full spectrum of emotion through its fantasy, but the fantasy always wins.
Whereas in stop motion there is no escape hatch.
Though the emotional valence of scenes shifted in response to the image-body. Style altered how both films breathed. The directorial core remained across every one of these sequences.
Different Direction for Different Bodies
One of the most important practical lessons of the dual release is that the directorial needs were entirely different between them.
One might imagine the dual release as a single film with two coats of paint. But that is nowhere near what happened.
Consider it like a tennis match in the same constraints between two different opponents. Except it must be played in the exact period of time. The timelines cannot cross-over. AI allows for these two parallel starting points: to view the same match from two different opponents in the same conditions.
The point is sometimes they will be fielding a backhand, sometimes a forehand. The job of a director across both aesthetics is to keep striking the ball however it comes.
When style becomes parameter, direction becomes more exacting. The director is no longer shepherding a film inside one ontology alone. The director is responsible for how the same core identity manifests through multiple ontologies, each with its own powers and weaknesses.
The AI animation director becomes a polyglot of visual language.
This affords an entire new definition of the job: an expansion of the responsibilities, an enrichment of the pleasures, and a deepening of its conditions, allowing for comparison and contrast; showing the future of directing films is not going to be a lost art, but something creating endless new opportunity.
This represents just one pathway. The expansions continue all the up to what I call, the everything movie. Multiple protagonists, multiple endings, customizeable costumes, different eras. The everything movie might have all these things in one single offering: at its best, it might would even encapsulate the True Line as a cherry on top. This demonstrates the progression, acceleration, and expansion of responsibilities AI brings upon the artform: with cinema as we know it not ending but representing the ending of its beginning.
This range encompasses both live action and animation at once.
Traditional animation directors worked in one language. Cross-pollination happened—many directors worked in multiple forms over the course of a career—but the individual production was almost always committed to one language, one career, one module.
AI allows for the everything movie except in animation. Imagine a film across four animation modules: stop motion, 2D, CG, let’s add anime. Imagine one where four modules co-exist in one shared narrative.
The AI animation director can work in any visual language, simultaneously, within a single production. Because new technology requires a new way of thinking that can only emerge with a production regime on this level of breakthrough. The mastery of what is invariant across visual languages came first: to direct the same film across multiple ontologies. As a practice this remains the same, just doubled.
The new polyglot everything director must understand what each medium is doing in dramatic terms, so that the same intention can be realized through every single instantiation of a movie.
The proof is not to say aesthetic does not matter, rather, style remains variable, for the variable is powerful enough to change the pressure on every scene it touches.
Why Style Was Mistaken for Essence
When Disney committed to the visual language of Snow White, that commitment was total and irreversible. The entire production—character design, background painting, animation technique, color model—was determined by the initial stylistic decision and could not be changed without starting over. Hundreds of animators trained in a specific tradition worked within a specific pipeline under specific economic conditions. The infrastructure was built to serve one visual language. Changing the style mid-production would have been equivalent to demolishing the factory and rebuilding it from scratch.
The same was true at every studio, in every tradition. When Laika committed to the look of Coraline, the puppets were sculpted, the sets were built, the materials were chosen, and the look was locked. When Miyazaki committed to the visual approach of Spirited Away, the character sheets were drawn, the color scripts were painted, and the look was locked. Every animated film in history has been locked into its visual language from early in production, because the visual language was the production. The look and the making were the same activity.
This is why style felt like identity. When the look cannot be changed, the look becomes the thing. The Nightmare Before Christmas is stop-motion. That is part of what it is. To imagine it in any other visual language is to imagine a different film, because under the old production conditions, it would have been a different film. Every decision, from the design of Jack Skellington’s armature to the lighting of Halloween Town, was determined by the stop-motion methodology. Change the methodology and you change every decision downstream.
Generative animation severs that bond. Meaning it is not the dual ontology itself, it is what that most succinctly reflects about the AI breakthrough.
What Remains Underneath
A melody can be played in a major key or a minor key. The melody is the same. The emotional coloring changes. Style in animation now functions the same way. The directorial vision is the melody. The animation ontology: 2D, stop-motion, or any other,is the key and the tempo. The composition endures across every performance.
Once style becomes variable, authorship rises into clearer view. The director can now wear style like a hat.
The animator-director is no longer bound to one inherited visual body per film. A film can move through multiple image-bodies while retaining its deeper self.
The Viewer as Curator
The most unexpected consequence of the dual release was what happened to the audience. This, in the process of directing, was an obsession that carried over to the viewer as well.
Viewers who watched both versions could not help but compare them. Every scene became a judgment: which version handled this better? Which carried this moment with more force?
It was impossible to watch the one without having an opinion about the other.
The viewer was no longer receiving the film passively. The viewer was evaluating, comparing, making curatorial decisions in real time. The dual release had turned the audience into critics of style as such—assessing the relationship between visual language and dramatic content with a precision that ordinary film viewing never demands.
The film unwittingly created a new sort of relationship between the film and the audience. A new way of watching. A new way of looking at cinema.
In normal film viewing, the audience accepts the visual language as a given. You watch a stop-motion film as stop-motion. You do not sit there weighing whether the scene would work better as hand-drawn.
The new exchange in viewership becomes a new sort of possession of ownership. An ingredient which to this point never existed in the viewing relationship. Films existed as art or entertainment, not as a possession to compare alongside another.
The dual release creates this comparison object. The same scene, realized two ways, forces the viewer to take sides. They are performing, in miniature, the same act of selection and judgment that the director performed in making the film. The audience has become a participant in the film’s realization.
This is the radical and unintended consequence of the experiment. Larger, ultimately, than the theoretical finding that enabled it. Style as parameter is the condition. The transformation of the audience is the result.
For the first time in animation history, viewers are engaging with style as a conscious variable rather than an invisible given. The implications for animation’s cultural geography require separate treatment. What matters here is the audience’s new position: participatory, comparative, and inexhaustibly engaged.
This is the living, breathing cinema realized at the level of animation style.
Conclusion
Animation has entered its living, breathing phase. One film can inhabit multiple valid bodies. The director’s authorship persists across them. The polyglot director making the everything movie; fluent in every visual language, anchored in the invariant architecture beneath all of them, inherits an instrument of unprecedented expressive range. The audience that encounters it inherits a relationship to animated images that is participatory, comparative, and alive.
Style as variable welcomes a new form of cinema.
