The following thoughts were prompted by an attempt to combine elements from Genette's narrative grammar with Metz's syntagmatic categories and Burch's scheme of articulations — or, perhaps more accurately, by some preliminary meditations on the extent to which such a combination would be feasible or fruitful. Some particularly interesting problems relate to identifying Metz's descriptive syntagma, or something equivalent to Genette's descriptive pause, in film — problems not merely related to applying a schema devized for prose narrative to a different medium but inherent too in Metz's own classification, with its identification of distinct descriptive syntagma within the wider category of chronological syntagma as well as separate types of achronological syntagma which can be seen as broadly descriptive in emphasis.
Underlying the following discussion are two factors inherent in the very medium of film. Firstly, virtually all film images have a descriptive dimension (probably the only real exception is the intertitle, where the film image becomes the equivalent of print). This has often been noted as an obstacle to applying the language model too literally to film : unlike prose, with its strictly linear, diachronic, syntagmatic form, film has two syntagmatic dimensions, both diachronic and synchronic, with each frame and each visual instant in the linear narrative, being saturated with descriptive detail. In prose narrative, however, the time flow of events has in effect to be disrupted — story time pauses while the discourse fills in the descriptive details. Secondly, there is the converse feature of moving film — its inescapably temporal nature. Metz places the syntagma he explicitly identifies as descriptive within the broader chronological category, since with normal projection speed real story time is obviously still passing during shots that seem predominantly descriptive, even though the fact of its passage may not be foregrounded in any particular way. Reel time, as McLuhan might have said, is also passing. As viewers we are bound to a temporal sequence, carried along inexorably by the running time of the film text itself. (One of the most intriguing aspects of the VCR is the way that readers' control of their reading ex-
115 Aust. J. Cultural Studies, 4:1 (1986)
perience with print texts has been extended to film texts; Philip Adam makes this point nicely in his plug for 'the print cartridge' (Bulletin, September 1, 1981.64).
Moving film, then, realizes literally what narrative prose can offer only notionally: those moments when discourse time and story time are truly identical. Because of this direct bondage to time, film has the special ability to slow down and speed up normal lived time literally too. Time lapse, fast motion, slow motion, freeze frame and the use of filmed still images (as in La Jettee) are all objective and quantifiable means of directly altering the relationship between story and discourse time. In actual practice of course fast motion is associated with comedy and tends to increase the narrative drive, but the conventional nature of this association can also be seen as tending to generalize and ritualize the action in a manner conducive to bringing out its descriptive aspect. All the other possibilities seem to be inherently descriptive, allowing us to focus our attention on features of the subject that normal living speed conceals and freeing the eye to explore the image more fully. (The slow motion action replay that interrupts the narration of a sports event is a straightforward example.) A similar effect can be generated in montage that simply repeats all or part of a shot (Genette's frequency scale applies readily here) or in the overlapping montage used by Eisenstein to stretch out actions without necessarily altering the speed of each shot (familiar examples are the breaking of the 'give us this day our daily bread' plate and the Odessa steps sequence in Potemkin). This descriptive emphasis is particularly strong in what one might call cubist montage, from Burch's account of it, in which one subject is shown from several different angles. Eisenstein's films, though, are a special case since their structure is usually deemed to be symphonic, rather than narrative in the everyday sense, with both spatial and linear composition strongly foregrounded. Indeed the prevalence of descriptive decelerations and pauses could be seen as particularly favourable to dialectic, giving the reader time to make a synthesis from the montage collisions without the distractions of a full narrative drive.
Both montage techniques like these and alterations in filming and projection speed foreground the medium of film itself. One might make the general point that any foregrounding of the medium serves to 'unforeground' the narrative and our sense of its movement through time, and hence increases the descriptiveness of what is shown. Camera's eye point-of-view shots, as distinct from unobtrusive or subjective points of view, can be added to the list of internally perceptible markers of description. Examples would be exaggerated cranes and zooms that draw attention to themselves, such as the long zoom out in Harold and Maud that begins by revealing that our
116 Aust. J. Cultural Studies, 4:1 (1986)
couple are picnicking in a cemetery and continues until they are reduced to a dot among dots and finally vanish altogether — a shot in which description becomes progressively more metaphoric.
Such immediately discernible effects alter the relationship between story and discourse time in an objectively quantifiable way that fits neatly into Genette's narrative grammar. Outside of these effects, however, categorizing a shot or sequence as predominantly descriptive becomes more problematic. In some cases there may be distinctive internal evidence that time is somehow to be deemed to be slowed down or irrelevant. Last Year at Marienbad and India Song, both provide clear examples of this in scenes in which the human actors keep unnaturally immobile and silent while the camera explores the scene at what seems a normal shooting speed. In establishing and scene-setting sequences the flow of time may seem irrelevant simply because of the absence from the image of any human or animal presence — or conversely through an overabundance of people or animals, provided no individual is singled out or dwelt on more than another. Normal time is still passing, but it is not being foregrounded internally through the narration of a particular character's activities. Identifying these sequences as being establishing or scene-setting, though, is probably more dependent on convention than on internal evidence; we expect the opening shots of a film to establish the scene and will treat them as descriptive until the moment when the camera homes in on and begins to follow one character, giving us the sense of time being lived through individually and so shifting into narration. A roughly similar sequence later in the film may well be highly narrative in its effect, as we imagine the protagonist and/or antagonist moving through time in off-scene space towards us; indeed, once the plot is under way and we are caught up in the action, we are likely to give narrative sense to whatever we see, unless a different possibility is explicitly signalled by means such as those discussed above.
Confidently classifying a shot as narrative or descriptive, then, frequently depends on external evidence and not on anything inherently present in that shot. Consider a medium close-up of moving coach wheels and how it might be read depending on diachronic syntagmatic relationships within a particular text: in a documentary on early forms of transport it might well be descriptive of 'coaches in motion;' in a Western it might have a primarily descriptive function as part of the familiar iconography of the genre or a strongly narrative one as the vehicle rushes towards or away from some important event. These examples show too the vital role played by the expectations that are aroused by our placing a text within a genre. The contribution to decoding of choices from media and genre paradigms is spelt
117 Aust. J. Cultural Studies, 4:1 (1986)
out by Fiske and Hartley (1978): the image of two children leaving school shifts in general significance as the medium changes from family photograph to advertising poster, or the genre from documentary to fiction play. If an actual genre label or other metalinguistic framing suggests a predominantly descriptive genre, such as 'wildlife documentary,' we will tend to foreground this aspect throughout our reading of the text, including during any narrative sequences it may contain (images of, say, one cheetah running down, killing and eating its prey will seem primarily descriptive of a process). Incidentally, one might argue that slow motion and repeat shots have become a recognizable part of the technical codes of this genre precisely because of their inherently descriptive tendency, which in this case underpins the significance the particular genre conventions have given them. But even with shots like these, it is impossible to overestimate the extent to which actually reading them as descriptive depends primarily on the genre conventions.
Convention too has established certain types of music, especially but by no means exclusively extra-diagetic music, with mood rather than action; indeed music usually performs a distinct metalinguistic function in shaping our reading of the shots it accompanies. Even in chase scenes, where the musical accompaniment is used to heighten the pace and excitement of the action being narrated, the very presence of that music seems to generalize and ritualize the action and so bring its lyrical and descriptive aspects to our attention, much as fast motion does with action peaks in comedy. This role of music is seen clearly in Westerns. 3.10 to Yuma is a particularly interesting case, since the persistent musical motif combines with some unusual technical choices to generate an elegiac mood that dominates the action (although admittedly this is made easier by the fact that the plot itself stresses situation and potential for action rather more than action in the 'action-packed movie' sense); moreover the sound track also draws attention to itself by the way it seems to play with the distinction between intra- and extra-diegetic sound — for example, at one point the theme is heard sung by a female voice and seems unmo-tivated, given the film so far, until a character makes a comment that seems prompted by his hearing what we are hearing.
Metaphoric shots, which do not add to our knowledge of physical details of characters or locations involved in the story, are usually classed separately from 'descriptive' shots, yet they do effectively focus attention in a descriptive rather than a narrative manner — we backtrack mentally, against the flow of the narrative as it were, matching details and drawing analogies. Shots that are purely metaphoric — non-diegetic inserts like Eisenstein's peacock that mocks the pretentious official — clearly do suspend story time. If
118 Aust. J. Cultural Studies, 4:1 (1986)
they are filmed and shown normally such inserts retain a natural temporal existence, but one so distinctly unrelatable to the time (and, in this example he space) of the events we are following that
they both interrupt the main time sequence and hold it momentarily frozen. Metz's bracket syntagma do the same thing, providing as they do images that function primarily as illustrations, being drawn from the same paradigms as items present in images that are part of the story but having no direct involvement in either the space or the time occupied by the narrative. However, even shots that are firmly
rooted in the narrative action can have a similar pausing effect if they possess some kind of metaphoric potential. This is demonstrated in particular by intra-diegetic match-cuts. Consider, for example, the well known bone weapon/space station match-cut in 2001; although this represent a daring acceleration of narrative time, the interaction of possibilities of seeing each of these objects as somehow like the other is what intrigues our imagination. Similarly, in Potemkin, even as the narrative suspense builds towards the moment when the marines will either obey or disobey the order to execute the rebellious sailors, a shot of an officer tapping his sword hilt is matched with one of a priest making a rhythmically similar movement with his crucifix and attention shifts to the political implications that this explicit visual parallel suggests. In such cuts as these the emphasis is upon likeness of form at least as much as on temporal sequence; the relationship between the two shots is one that foregrounds atemporal descriptive possibilities.
Metz's other achronological relationship, that between parallel stories, which has its classic instance in Intolerance, extends the potential for metaphoric as opposed to metonymic connections over a whole film. This effect is intensified in Resnais's La Vie est un Roman, by the juxtaposition and mixing of mediums (live action, cartoons) and of genres (fairy story, expressive realism, expressionism and musical).
To sum up then, in applying narrative grammars to the analysis of specific texts, our identification of certain segments as being descriptive in function rather than narrative seems largely reliant on evidence other than that actually contained within a particular segment. Some reasonably reliable evidence internal to the individual shot does seem to exist however. Whenever the medium of film itself is strongly foregrounded within a shot the image tends to take on a descriptive bias. This is most obvious in alterations to normal filming and/or screening speeds — time lapse, slow motion, freeze frame and even fast motion, despite the increase in narrative pace — and in camera eye shots, such as exaggerated cranes and zooms. Such technical choices might reasonably be said to be describing the act of narration itself, rather than any subject matter within the frame, but
119 A ust. J. Cultural Studies, 4:1 (1986)
they do help the viewer escape from the linear impetus of narrative and redirect attention to more descriptive aspects of what is shown. Internal markers of description can also be provided by the mis-en-scene — reasonably unproblematically in the static groupings of actors found in films by Resnais and Duras, but much less reliably in the lack of any character to whom narrative importance can be readily given, since in the latter case the actual position in the overall syn-tagmatic chain of the text or the genre to which the text is judged to belong are likely to be determining features. Indeed, with any examples of strictly internal evidence, it is extremely hard to distinguish some sort of intrinsic descriptive potential from the identification of certain devices as descriptive based on reading practices triggered by some external metalinguistic frame, to decide whether the potential has produced and underpins the device or has merely been produced by it. Certainly, whatever degree of inherent descriptivity one might claim for a particular type of shot, one can scarcely discount the extent to which genre conventions shape and influence interpretation, making available certain meanings and closing off others.
In most shots the markers of description are not objectively present; the 'evidence' we rely on is not something that is foregrounded internally but is the product of syntagmatic relationships within the text. For example, position in the text alone may be what distinguishes a scene-setting from a suspense-building function. Recognising certain shots as extra-diegetic and hence descriptive obviously depends on their relationship with the rest of the text, as does the dividing of such shots into the sub-categories of purely metaphoric, illustrative etc. In a range of cases the relationship is one that draws attention to the film medium in some way, with the consequent increase in descriptiveness noted above with internal foregrounding — as in aggressive montage (except when used in the stereotyped action climax), overlap, repetition, match cutting, and indeed overtly metaphoric shots.
And finally, all such intratextual evidence is conventions of a culture, the genre expectations, and a range of metalinguistic cues, from the style of a poster advertising the film to the known programming policy of a given cinema, be it art house or porno pit — and even more generally on the subject positions that the type of text and type of society makes available.
To return to the narrative grammars with which this discussion started — the problems that emerge from trying to apply the categories proposed by Metz or Genette to film texts clearly do not mean that they have little practical use; at the very least they provide an invaluable ready made means of separating out the elements of a text
120 Aust. J. Cultural Studies, 4:1 (1986)
for the purposes of closer scrutiny, as their use continues to show. Moreover, the very nature of the problems, rather than working against the status of such grammars, merely bears out that crucial perception of which the structuralist and semiotic enterprise has made us acutely aware — that 'meaning' does not reside in individual units, be they words or images, but precisely in those relationships that structures of choice and of assemblage set up.
Jane Crisp teaches at Queensland University.
Notes
1. Poetry is the exception here — obviously in concrete poetry but also more generally, both in its layout which provides a constant metalinguistic reminder and in the related practices of form and reading method which mean we will investigate the possibility of spatial relationships of units both within and across the lines.
References
Burch, N.., (1973), Theory of Film Practice, London: Secker and Warburg.
Fiske, J., and Hartley, J., (1978), Reading Television, London: Methuen.
Mast, G., (1971), 'Soviet Montage' in A Short History of the Movies, New York: Pegasus.
Monaco, J., (1981), How to Read a Film, New York: Oxford University Press.
Rimmon-Kenan, S., (1983), Narative Fiction, London: Methuen.
New: 20 October, 2019 | Now: 6 November, 2019