The film lacks rhythm! It is often by virtue of its absence that rhythm gleams for an instant in the eyes of film critics. It is true that we tend to mistake it for that briskness which Cicero once made the essential quality of a narrative. But conciseness – the cinematic definition of which is no better known – can no more be confused with rhythm than a film simply amounts to the narrative it carries.
The difficulty of the problem – but also what is interesting about it – lies in its complexity. Some narratologists have proposed that we compare the duration of the various parts of a story with the space they occupy in the narrative text. The variations in this relationship would constitute a rhythm. This solution already seems inadequate in literature: a story is made up of events, not hours or months; and since a sentence has its own rhythm, a text cannot be measured in lines. In cinema, moreover, there is hardly anything but fragments of real, invariable time. For his part, David Bordwell has taken pains to count the number of shots and compare their average length.  But this statistic is misleading: it reveals, for example, that the art of silent cinema cut much more often than the talkies, which doubled the average shot length. Yet the impression is generally shared that silent works are slower than sound films. On the one hand, it is absurd to count the intertitles as shots, despite the cuts that surround them; and since it is impossible to do otherwise, we can see that the comparison is inevitably lame. On the other hand, speech drastically changes the content of a shot and accelerates its unfolding. Finally, cinematic time, for all its realism, has nothing to do with chronometers – and even if one uses them, the results remain resolutely subjective.
On top of all that, one cannot make a shot a rhythmic unity in the musical sense: a measure that would allow us to organise the configuration of the rhythm. In fact, conversation, action and music are not limited to suggesting broader cadences; they can render imperceptible, or at least inoperative from the rhythmic point of view, the dividing line that constitutes the cut. Moreover, the shot possesses its own rhythm.
Finally, rhythm is probably the morphological component in which the problem of the relationship between small-scale and large-scale forms is posed with the greatest pertinence. The rhythm of a shot and the rhythm of the film, the rhythm of a gesture and the rhythm of the dramatic representation. We know, or we think we know the rules of a good script: the alternation of high points and lulls or temps morts; the final acceleration; the intensification that must mark the end of each part. But these rules (which are, anyhow, contestable) quite plainly ignore the fact that no segment of a film is inert as far as rhythm is concerned and that, consequently, lulls and high points are notions completely relative to a context. For light and shadow, body movements, the actors’ phrasing, the respective length of dialogue lines and of silences, the varying intensity that the dramatic representation imparts to the implementation of the narrative – all make up rhythm in cinema.
The start of Angel Face (Otto Preminger, 1953), exterior night: the ambulance driven by Frank (Robert Mitchum) disappears at the right of the frame; coming out of the house, Diane (Jean Simmons), followed by a movement of the camera, slips toward the garage, at the left; she exits it in her top-down convertible, whose movement toward the right is shown for only an instant. The shot presents a rhyme: the double exit of headlights. It is composed of a measure with three beats, which are not simple durations, but determining of temporalities. The first beat is brisk and simple, a matter of pure execution; but it includes an acceleration, shown less by the vehicle than by the abrupt exit from the frame: the speed of a breakaway. The second beat surprises: as in a syncopation, the beginning of an unexpected action takes the place of the ending of an expected one. The ambulance driver leaves: is that all? No! The mysterious stepdaughter leaves her home furtively. Even and discreet, the pace of her steps along the façade contradicts the narrative accentuation. This beat is thus contrasting and tense. The third beat has the same clarity as the first: we grasp that the girl is following the hero, and the straightforwardness of mechanical movement is substituted for her furtive gait. But this beat retains the tension of the second: why is she tailing him? And the sports car accelerates faster than the ambulance!
Without pausing to admire the fine animation that rhythm introduces into this segment, we can observe that it expresses the art of “folding” with which Gérard Legrand characterised Preminger’s genius.  The contrast between Mitchum’s resoluteness and the uncertainty that surrounds Simmons is resolved in the apparent affinity that, to finish the sequence, brings their respective movements closer together – but not without giving cause, via the deferred character of this affinity, to suspect that there is something threatening in the acceleration which the heroine generates. Through the way its rhythm is composed, the shot thus fits together with the film as a whole; but not as a rhythmic part of a total rhythmic schema, nor as a predictable element of a regular, general rhythm. The rhythm of the shot is, rather, part of the definition of the characters as elements in the representation. The contrast between what is accented and the shiftiness of gesture with which Simmons executes her exit – the central form of this segment – allows an organisation of the variety of tempos, and has decisive value for our knowledge of the character. This demonstrates that dramatic aspects control musical form. For this reason, Juri Tynianov’s definition concerning the “interaction of stylistic and metrical moments in the unfolding of a film, in its dynamic” shows the weakness of an overly narrow conception of forms.  Rhythm governs a representation; no metre can be postulated in a general way.
The association of a high point with a less marked cadence seems, moreover, rather frequent. In Au revoir, les enfants (Louis Malle, 1987), a shot in which Julien (Gaspard Manesse) plays ragtime with Bonnet (Raphaël Fetjö) is abruptly followed by a broad, silent, location shot: the wind, blowing the branches, sweeps flakes of snow along in light, glittering eddies. Only this visible, almost imperceptible rhythm recalls the liveliness of the music while attenuating it; whereas the understanding between the children was previously conveyed by the exchange of glances and smiles, their alliance from now on is made manifest, in a more montonal fashion, by their appearance side by side at the far end of the courtyard. It is the evident de-dramatisation that gives this cut its value as a dramatic accent. The presto of a nascent friendship is succeeded by an andantino that is more melancholic, but above all less defined. This rhythmic confluence also harmonises with a major theme in the film, in which the lively tone of the chronicle comes up against the gravity of a historical tragedy that is only present on the horizon. Witness the long scene of the treasure hunt that the encounter with German soldiers abruptly ends.
To speak of rhythm is to uncover the ordering of temporalities whose forms govern heterogeneous materials of various durations. The subtlety of some of these orderings is hard to get a grip on. A little later on in Angel Face, a tracking shot follows Simmons and Mitchum in a convertible. Their conversation has its own rhythms, determined by the speed of delivery, the length of the dialogue lines, and the silences; but in the background set they are driving through, probably by means of some technical artifice and therefore excluding any necessary concession to pure chance, the bright or sombre tones of the parked cars and, above all, the black holes that correspond to the cross streets express another rhythm that is specifically filmic: can it be combined with the theatrical rhythm of the dialogue? Should one, on the contrary, consider it a counter-rhythm that undermines the intelligibility of the other? Since it is impossible to perceive both rhythmic forms at the same time, the second solution seems more appropriate. But it contradicts the idea of a single rhythm that envelops all those phenomena that might justify the use of metrics.
In an ‘avenue for reflection’ such as this Positif column, it would be unwise to reach a firm conclusion.  Rhythm remains an enigmatic form. What matters most of all is to call attention to it. Sketching a method to deal with it is of somewhat lesser importance: it is on the level of the smallest forms that it seems least troublesome to describe a rhythm. Gestures, sounds, the movements of people and things, dialogue exchanges, camera movements and cuts, but also music – can they be perceived or at least grasped, in whatever slice of a film, as a unique configuration? Or rather, should we distinguish several figures organising, at the same time, several types of temporal material? That is the first question. It recalls problems of classical versification, for example the coincidence or non-coincidence between linguistic accents and metric ictus in a verse constituted on complex roots, such as Latin verse. But our case is more obscure, since cinema does not specify any kind of generalisable measure.
Assuming that a rhythmic schema can be determined, how is it to be integrated into the film’s totality? The examples cited above postulate an analogy between the rhythm of the shot and the rhythmic dispositif of the representation. Two other possibilities deserve to be explored. The analogy could define a strict and general correspondence between the rhythm of the whole and the rhythm of each part. Or the integration could be activated step by step, in a linear fashion, without any proportion governing the totality. Now, the grasping of global rhythm is not at all foreign to filmmakers’ conscious preoccupations. Arlene Croce has eproduced a diagram, taken from director Mark Sandrich, which clearly shows the length and positioning of the musical numbers in Follow the Fleet (1936).  The heterogeneity of scenes in the musical genre allows us to discern three rules: the first fixes a ternary rhythm; the second makes the in-between time into a weak time; and the third opposes an initial accentuation (at the start) to a final accentuation (toward the end). This implies a progressive transformation rather than an all-over organisation. Such an analysis might merit some new research.
 See the collectively authored The Classical Hollywood Cinema: Film Style and Mode of Production (London: Routledge, 1985).
 See Gérard Legrand, Cinémanie (Paris: Stock, 1979), pp. 292-298.
 Anthologised in François Albera (ed.), Les Formalistes russes et le cinema. Poétique du film (The Russian Formalists and Cinema: Poetics of Film) (Paris: Nathan Université, 1996).
 This article appeared in the magazine’s regular opinion/reflection section, “Chantier de réflexion”.
 In Croce, The Fred Astaire and Ginger Rogers Book (New York: E.P. Dutton, 1987).
Original French text “Notes sur le rythme” appeared in Positif, no. 439 (September 1997), pp. 63-65. Translation by Jacqueline Kaminski (2006), completed and revised by Adrian Martin (2015). Thanks to Bo Palinic for conveying this document to Screening the Past, which is reprinted with kind permission of the author and of Positif.