War and Peace, Part II: Natasha Rostova · essays & theory

A reading · through the lens of theory

Twice on the road to Otradnoye, Andrei passes the same oak. Going out it is bare. Coming back it is in full leaf. Bondarchuk photographs both plainly, unhurried, with no music telling you this is the moment and no dissolve stitching one state into the other. He shows you a tree. Later he shows you the tree. Everything that turned the first into the second has been left out of the film.

That omission is the method. What is on screen is not a tree changing; it is the gap between two states, and the gap is time itself, handed over raw. Deleuze has a name for an image built that way — a chronosign, an image where time appears directly instead of being measured out by somebody's movement through it. Most films give you time as a by-product: a man crosses a room, so time passed. Here nothing crosses anything. Two shots, months apart, and what you feel between them is duration with no job to do. And the change registered isn't in the tree at all. It's in Andrei, who has no action available to express it, so the film gives it to a plant.

Hold that, because it explains the strange thing about this movie. Natasha Rostova is the second part of the largest production in Soviet film history, a state project mounted to take the national novel back from King Vidor. Sovscope 70. Six-track stereo. Army cavalry on standby, museums emptying their vitrines. And Part II spends all of it on a girl in rooms between roughly 1809 and 1811: a first ball, a wolf hunt, an evening of folk music at a country uncle's, an opera, an elopement that fails. No Austerlitz. No Borodino. Bondarchuk had seen something structural in Tolstoy, that the war chapters and the peace chapters run the same method at two scales, and Part II is where he proves it by removing the spectacle and leaving the technique standing there uncovered.

You can watch this as a sumptuous literary adaptation and be entirely right about the porcelain and wrong about the film. The Deleuzian lens earns its keep by explaining two things that lavishness cannot: why the ball feels the way it does, and why the two oaks land without a word of commentary.

So go to the ball. The camera does not take a good seat and watch. It comes in the door when Natasha does, gets down among the dancers, lifts, circles. You arrive the way she arrives — provincial, half a beat behind, dazzled by the sheer quantity of it. Then Andrei asks her to dance, and what changes first isn't her face. It's the room's physics. Deleuze noticed this in musicals and called it the movement of world: normally a character moves while the world holds still, but once the dance takes over, the relation flips. The dancer stops being the source of the motion. The world takes it up and carries the body. Natasha does not dance her way into society; society revolves, and she is carried inside it. Bondarchuk cast Lyudmila Savelyeva, a Kirov-trained ballerina with almost no screen experience, and the choice stops looking eccentric the moment you see that the part doesn't need someone who can act a room. It needs someone a room can move.

The craft debt is exact. Ophüls had already run a crane around a waltzing couple in unbroken circular motion in The Earrings of Madame de..., using the dance itself as the ellipsis that carries the emotional change; I Am Cuba, two years before, had shown how to pass a camera through crowds and doorways as though nothing supported it. Bondarchuk's invention is doing this at 70mm, with cameras built for landscapes and too heavy for it, on rigs made for the purpose. Scale and intimacy in the same gesture. And the camera never stops being felt as a camera — Deleuze's dicisign, where the film perceives a character who is perceiving, so that you sense a frame set around her framing.

Then the face. At 70mm a face is architecture. The Thaw had already found the technique — Urusevsky in The Cranes Are Flying, the wide-angle shoved in until the room bends around a head — and Petritsky rescales it to large format for Natasha's flushed passages through crowded rooms, and harder still at the opera and at her collapse after Anatole: canted frames, whip pans, eyes going straight to the lens. Deleuze splits the close-up in two. One kind reflects a single settled quality, wonder or grief, complete in itself. The other, the potisign, is a face where small movements gather toward a change of state that hasn't happened yet. Natasha's is nearly always the second. She is a threshold, filmed.

The other inheritance is Gance. Napoléon layered remembered and imagined images over a character's face, and Bondarchuk takes superimposition and dissolve out of the transition business entirely and uses them as thinking — images sitting on faces to carry memory and anticipation at once. That last part matters: you frequently cannot tell which you are being given. This is the crystal-image, where the actual present and its virtual double become impossible to prise apart. Superimposition is simply the literal instrument for it. Two images in one emulsion, neither one subordinate.

Tatyana Likhachyova's cutting is the least-discussed thing here and arguably the boldest, because it holds two rhythms that normally cannot share a film. Scenes run at conversational length with nothing advancing — Deleuze's dead-time, the everyday held open, time released from its duty to the plot. Against that, the hunt and the opera pile fragments up faster than you can parse them, accumulation crossing into shock, which is the pathetic montage Eisenstein built and which Bondarchuk keeps in the toolkit. Meanwhile the interior-monologue voiceover is cut against the picture freely, so speech, thought and vision run on separate tracks. That is a mild audiovisual disjunction, nowhere near Duras — but a state epic in 1966 running three tracks that decline to agree is not a small thing.

There is a theoretical oddity worth naming. The staging is Bazinian: rooms populated to the edges, characters minding their own business in depth, the ballroom disclosed as continuous inhabited space instead of assembled out of coverage. Bazin argued that this hands the viewer the work of selecting, and preserves the ambiguity of the real. But Bondarchuk's camera selects, moving through the density the way Eisenstein crowded his deep planes in Ivan the Terrible. He wants Bazin's room and Eisenstein's cut in the same picture, and mostly gets them. What that produces is the Open — the sense that the frame is never a closed box but a slice of a whole that is still changing. The ballroom is continuous with the smoke behind it and the smoke ahead of it. Part II is where you feel this most, because there is no army on screen to do the feeling for you.

What it did to filmmaking: it pointed the reserved-seat epic apparatus inward. The 70mm roadshow existed for chariots and oceans; here it is spent on a teenager's year of waiting, and on a tree. It also refused Vidor's compression — Tolstoy's episode order kept intact, serial length instead of one feature — which is a narrative argument, not just a logistical one: the novel's sequence over the dramatic arc.

Its descendants are hard to name, and that is part of the story. For forty years the West saw a dubbed, shortened, faded thing that won an Oscar and misrepresented what Mosfilm had made. Only the 2019 restoration made the film legible abroad. So it arrives as an ancestor late, a chapter the discourse could not read when it was published.

Watch the ball twice. Once for Natasha, once for the camera, and ask which of them is doing the dancing. Then go back to the oak. Tolstoy tells you what it means. Bondarchuk declines to, and leaves the question sitting there: is the change in the tree, or in the man who rides past it, or only in the months the film refused to show you?