“If an image, looked at separately,

expresses something clearly,

if it includes an interpretation,

it will not transform itself in contact with other images.

The other images will have no power over it,

and it will not have power over other images.

No action, no reaction.

—Jean-Luc Godard, King Lear (1987) (and elsewhere)

We are on the way to montage. But before trying to say what montage is, we should step back and ask: what makes it possible? What must an image be, if it can be combined with other images? Will any image work, or are there certain kinds of images that lend themselves to montage?

“Montage” simply means “construction” or “composition,” and it is not restricted to cinema. For our purposes we can say that it happens whenever we make something out of what we see, piecing images together while breaking them up. By giving the interplay between each image and all the others a particular shape and trajectory, the process of montage constructs a special visibility and makes it last, if only for a little while.

The lines above, spoken in English by Godard, are drawn from Robert Bresson’s Notes sur le cinématographe. Bresson’s book is a compilation of aphorisms, slogans and verbal sketches, encompassing philosophical observations, personal exhortations, and story ideas, all taken the “log books” he kept between 1950 and 1974. In this clipped and gnomic form, Bresson’s Notes resembles nothing so much as a book of spiritual exercises for aspiring filmmakers. This particular aphorism dates from 1950-58.

On this account, images acquire their power not because of their inherent qualities, but because they prove themselves to be transformable, that is, because they can enter into relations of composition with other images. This claim must be surprising to anyone who has admired a particular shot in a movie for its pictorial qualities; it is hard to resist the idea that a great film would be one where every shot could stand on its own. (The photographic ideal dies hard.) Yet any image that presents itself as self-sufficient—one that offers up its own meaning, that tells you everything you need to know, that can be taken at face value—will be useless for the purposes of montage, that is to say, for cinema and for thinking.

In an earlier aphorism, Bresson explains: “An image must be transformed by contact with other images as is a color by contact with other colors. A blue is not the same blue beside a green, a yellow, a red. No art without transformation.”  Just before the passage quoted by Godard, Bresson proposes a  “Cinematographer’s film where the image, like the words in a dictionary, have no power and value except through their position and relation.”

Even after we have grasped the principle that images are irreducibly multiple, dynamic, and changeable, there is still something enigmatic in these suggestions. We are being warned that the impotence of images arises exactly where they seem strongest: when they bring us a clear message or when they dazzle us with their beauty. On the contrary, Bresson asserts, images must not be too “definitive.” When expression leads directly to interpretation, the image claims too much attention for itself and thus is immediately exhausted. How, then, can images evade this kind of fatal clarity or affective overkill, without succumbing to the opposite problem, vagueness and incoherence?

How can images be made, so that they will make themselves available to other images? How can we tell which images will transform themselves “in contact with other images”?

It might seem obvious that film directors compose their images in order to assemble them into “strong” montage sequences. That is what the vast majority of films offer us: there is not one image out of place. To the degree that every image aligns with the next one along a well-marked path—guided by the overarching unities of narrative, genre, design, etc.—such films will present themselves with built-in interpretations, which the spectators will be more or less able to recognize every step of the way.

In opposition to that way of working, we have to proceed as if it is possible to practice montage without the advance guarantees of formulaic composition. In that sense, montage can be a radical experimentation, in which we discover which images are strong enough to overcome aesthetic and semiotic inertia by seeking only the strongest combinations. Each image, no matter how definite, will acquire its force only when it is seen in the midst of other, equally provisional, images.

It is a question, as Dirk Baecker said to Alexander Kluge, of what the images know: “What does the image know about the next image, what does every shot of a film know about the next shot, which is only possible when this or that has been shown before?” Baecker insists that Godard makes every image according to the knowledge that he can find a way, however unexpected, to put them together. Kluge replies: “Godard puts incompatible images next to each other in the hope that meaning will arise out of the gaps between them.”

Perhaps neither of these positions is quite correct. No image really “knows” what is coming, even if Hollywood tries to make images that never fall off the rails. But that does not mean that images “know nothing” before they are assembled by montage. Every image anticipates the next, although the next does not always arrive, just as every image remembers the last, even when it was never there.


Quotations from Robert Bresson, Notes on the Cinematographer, trans. Jonathan Griffin (Copenhagen: Green Integer Press, 1997) pp. 20-1.

Quotation from Dirk Baecker and Alexander Kluge, Vom Nutzen ungelöster Probleme (Berlin: Merve Verlag, 2003), p. 141-2.

Godard uses Bresson’s note again in Histoire(s) du cinéma, episode #1B, where it is read on the soundtrack by Julie Delpy. Unlike King Lear, this citation includes the final line: “It’s definitive and unusable in the cinematographer’s system.”