Narrative and the Film Image

Author(s): Alfred Guzzetti

Source: New Literary History, Vol. 6, No. 2, On Narrative and Narratives (Winter, 1975), pp.
379-392
Published by: The Johns Hopkins University Press
Stable URL: http://www.jstor.org/stable/468426
Accessed: 14/12/2009 16:13
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=jhup.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and extend access to
New Literary History.

http://www.jstor.org

Narrative and the Film Image

Alfred Guzzetti

in a fiction film-the framing, composition and

of
shots, the movement of the camera, the placesequence
ment of cuts, and all the rest-be explained by reference to
narrative alone? The most influential writers on film assume that it
can. Andre Bazin, for example, in analyzing the structure of the
image in a hypothetical film scene of the I93os, finds it sufficient to
cite only two principles:
AN THE IMAGE

(I) the verisimilitude of the space, within which the position of the character is always determined, even when a close-up eliminates the decor;
(2) [the fact that] the intention and effects of the decoupage are exclusively
dramatic or psychological.
In other words, played in a theater and seen from a seat in the orchestra,
the scene would have exactly the same meaning, the event would continue
to exist objectively. The changes in the camera's point of view add
nothing. They only present reality in a more effective manner. First by
permitting a better view, then by putting the accent where it belongs.1
Like most theorists and critics, Bazin shows no inclination to challenge
the idea that narrative, with its "dramatic or psychological" requirements, is the controlling factor in shooting and cutting and that the
image has only the very subordinate role of supplying "accent." In
fact, except in the limited class of instances where "the verisimilitude
of the space" is at stake, he proposes no terms at all with which to
explain the form of the image. Elsewhere, he makes this silence a
critical principle, siding with directors "who believe in reality" as opposed to those "who believe in the image."2 When he discusses a
specific shot, he does so only to the extent that the narrative informs
I Qu'est-ce que le Cinema? (Paris, 1958), I, I40. The translation is mine and
corresponds to that in What is Cinema? tr. Hugh Gray (Berkeley and Los Angeles,
I969), pp. 31-32. The term decoupage, which has no English equivalent, designates
the division of the action into shots; it is to be distiguished from montage, which
means editing.
2 Qu'est-ce que le Cinema? I, 132.

38o

NEW LITERARY HISTORY

it, never as it reflects the whole array of decisions made by the cinematographer, editor, and director. The requirements of spatial verisimilitude and narrative are, in his view, independent of these decisions to
such an extent that he willingly associates different sets of decisions
with "exactly the same meaning."
This conception of image and narrative, paradoxical and deficient
as it seems, merely reflects that implied in the style of classical sound
films.3 To understand the source and consequence of these deficiencies,
one must examine the conventions of that style in some detail, specifically those that link the "dramatic or psychological" to the spatial.
For this purpose, it is useful to refer to the practical manuals of cinematography and editing where these conventions are formulated as rules.
The fact that they appear in this guise rather than as theoretical or
critical propositions is significant in itself, since it reflects the consensus
of critics and viewers that "film" is so completely identified with
classical narrative that the workings of this style are not convention,
but "technique." Yet inasmuch as these rules are compatible only
with certain views of image and narrative, they undeniably constitute
a critical and theoretical position.
Two sets of rules, those governing continuity and screen direction,
are of particular relevance to the relation of narrative to image.4 The
first concerns the problem of showing an ongoing action in successive
shots. The basic rule of this set requires the editor to join the shots
so that the action is picked up at the beginning of the second at the
precise point it reaches by the end of the first. In shots that show
human beings, as in many that do not, the action normally includes
segments of continuous movement as well as resting points, called
"nodes." A second, and much looser, rule favors placing the cut on
the node. The consequence of both rules is to permit the logic of the
action to dominate that of the cutting. Cuts so made tend to be
"invisible": that is, they do not signify that two pieces of film have
been joined together, as they would if a segment of the action were
repeated or omitted or if the cut interrupted movement or stasis.
The continuity rules specify further that at the point of the cut
there must be a substantial change of scale or angle or both. Thus, if
3 I use the term classical more broadly than Bazin, who restricts it to the common
practices of decoupage among films of the 19305.
4 My account of these rules follows that in Karel Reisz and Gavin Millar, The
Technique of Film Editing, 2nd ed. (New York, i969), pp. 213-26, cited hereafter as "Reisz." The same rules may be found in Terrence St. John Marner,
Directing Motion Pictures (London and New York, I972), Arthur L. Gaskill and
David A. Englander, How to Shoot a Movie Story (New York, I967), and Kenneth
H. Roberts and Win Sharples, Jr., A Primer of Film-making (New York, I971).

NARRATIVE AND THE FILM IMAGE

38I

the editor begins with a medium shot showing the actor from head to
mid-thigh, the rule allows a cut to a close-up of the face of the same
person seen from the same angle, but not, say, to a shot that shows
him, again from the same angle, from head to waist. The rationale
given is that the prohibited cut produces a change of scale slight enough
to seem jumpy and awkward.
But a more consequential explanation is possible. In the essay
"Dickens, Griffith, and Us," Eisenstein observes that the difference
in the American and Russian conceptions of the close-up is reflected
in the two respective languages.5 The Russian term literally means
a "large shot," as does the French "gros plan." Thus in Russian and
French, the shot is defined by its graphic aspect, the scale of the image
on the surface of the screen, whereas in English, it is defined by the
proximity of the camera to the subject. This observation suggests a
potential ambiguity in the instance I have given of the rule of scalechange: for any cut that takes the camera closer to the subject is
capable of signifying both "enlargement" and "approach." The small
change of scale prohibited by the rule would link two shots whose
overall graphic patterns within the rectangle are inevitably very similar,
thus calling attention to the surface of the screen and to the associated
signification "enlargement." By contrast, the close-up permitted by
the rule changes the graphic pattern sufficiently that attention to the
screen surface goes unrewarded, allowing the cut to be understood
unambiguously as a movement of viewpoint situated within the space
in which the action is played. Small changes of angle are excluded
on the same principle.
In addition, angle changes are subject to the rules governing screen
direction. A shot in which Actor A is shown screen-left and Actor B
screen-right must not be immediately followed by a shot in which their
screen positions are reversed. This condition will be met if both shots
are taken from the same side of an imaginary line, sometimes called
the director's line or center line, established between the two actors
on the floor of the shooting location. The simplest instance is a dialogue
scene in which A and B sit at opposite ends of a table: shots taken
from any angle within the I8o0 in front of the table may be cut
together; shots taken from behind the table can be cut in continuity
with these only if the camera or actors move, thus establishing a new
center line, or if a spatially "neutral" shot such as a close-up, insert,
or high-angle, is intercut. Similarly, a shot of an actor walking screen5 Sergei M. Eisenstein, Film Form, ed. and tr. Jay Leyda (London, 1963), pp.
237-38. The essay in question is included under the title "Dickens, Griffith, and
the Film Today."

382

NEW LITERARY HISTORY

left to screen-right cannot be cut to a shot of him walking right to left

without signifying that his path has changed direction. Shots of A
walking left to right intercut with B walking right to left imply that A
and B will meet in a subsequent shot. Shots of A running right to left
intercut with B running in the same screen direction signify pursuit.
The meaning of these configurations depends on the hypothesis
that a direction on the screen signifies a direction in three-dimensional
space and that the shots are segments of a spatial whole. The perspective cues in the image are evidently sufficient to place the screen direction within the dimensions of the pictured space only if the fundamental
directional hypothesis remains uncontradicted. A permitted anglechange in the dialogue scene may differ from anything we perceive in
reality, yet it shows enough about the placement of the two figures for
the spectator to understand how the two viewpoints are situated in
space. By contrast, the prohibited angle-changes draw an equivalence
between one spatial direction and two opposed screen directions. The
result is disorienting because it cannot be immediately interpreted in
terms of the pictured space but momentarily only in the dimensional
dichotomy of the image plane.
Clearly, the aim of both sets of rules is to allow the information in
the flat image to be translated without ambiguity into the threedimensional space of the action: in Bazin's words, to produce "the
verisimilitude of the space, within which the position of the character
is always determined." The contemporary theorist Christian Metz
describes the resulting effect this way: "In a scene of everyday conversation, [the spectator] will more or less have the impression of having seen at every moment the whole of the room. Perception reconstructs a unitary space-time through a spontaneous (but coded) totalization of partial views and through the play of immediate memory."6
This process of spatial integration, consciously experienced by most
spectators only as the difficulty of recalling the decoupage of a scene, is
obstructed by styles that violate the continuity or directionality rules
or, like Eisenstein's, rely heavily on the close-up. Classical style, as
Bazin notes, can tolerate the spatial ambiguity of an occasional closeup, although to "believe in reality" fully means, as in Rossellini's films,
to avoid the close-up altogether.
But since the rules of continuity and direction have more to say about
what is not to be done in shooting and editing than what is, alone they
do not constitute a sufficient basis for explaining the image. A more
positive factor, which Bazin calls "dramatic or psychological," is
6 Christian Metz, "Ponctuations et demarcati. ns dans le film de diegese," Cahiers
du Cinema, Nos. 234-35 (Dec. 1971, Jan.-Feb. !972), p. 69.

NARRATIVE AND THE FILM IMAGE

383

described by his British contemporary Ernest Lindgren as follows:

"The fundamental psychological justification of editing as a method
of representing the physical world around us lies in the fact that it
reproduces this mental process in which one visual image follows
another as our attention is drawn to this point and to that in our
surroundings."7 Reisz points out that though this rationale may be
extended to cover changes in scale, which simulate changes in attention,
it does not adequately explain changes of angle, which correspond
to no simple fact of individual perception. Reisz's explicit answer is
to ascribe these changes to "the right of selection which we accept from
an artist," who gives us a dramatic, and therefore "ideal," view.8
But in passages like the following, he implies a more concrete
explanation:
A man is sitting at a table on which stands a glass of wine. He leans
forward, picks up the glass with his right hand, brings it to his lips and
drinks....
Just before the actor starts his forward movement, his facial
expression-a glance downward, possibly-will register his intention. If
the cut is made precisely at this point-i.e., just before the hand begins to
move-it will be smooth, because it will coincide with the mnomentof
change from rest to activity. (Reisz, pp. 217-18)
The "psychological" principle interwoven with the rule favoring cutting
on the node is more complex than it may first appear. The cut, like
the glance, registers intention; it externalizes something that happens
in the character's mind. Yet, oddly, the scale of the image changes not
for the character but for us, the spectators. As the actor looks downward, our tendency to identify with his gaze redefines the field of our
attention to conform with his. The image enacts this change by switching not to a downward view of the wineglass but to a still frontal but
tighter shot which locates the glass at the bottom of the frame.
What this example suggests is that, contrary to the explicit views of
Reisz, Lindgren, and Bazin, classical cutting does not mimic the attention of an observer to an action so much as it represents attention
mediated through the actor. Only infrequently does this mediation
take the form of an actual coincidence of a character's spatial viewpoint with the camera's. Even when the image shows a character,
as it must, from a spatial viewpoint which could not be his own, it
encourages us to identify with him, at least to the extent of imagining
ourselves in his position within the pictured space. This identification
7
8

Ernest Lindgren, The Art of the Film (London, 1949), p. 55.

Reisz, pp. 215-16.

384

NEW LITERARY HISTORY

is rewarded not only in reverse-angle cuts, but, as Reisz's example

demonstrates, also in subsequent shots that continue to show him, as
it were, "objectively." The common shot-reverse-shot pattern of dialogue scenes, which Reisz finds difficult to justify theoretically, may be
referred to this principle. The shot of Actor A taken from over Actor
B's shoulder, usually explained as a rendition of B's "point of view"although the camera never precisely occupies B's position (A never
looks directly into the lens) and although frequently a portion of B's
head or body is visible-also causes the spectator to identify himself
with A. If it did not, the cut to the "reverse" angle, showing B from
over A's shoulder, would be unintelligible.
A relation may be drawn between this psychological principle and
the aim of spatial verisimilitude. The continuity and directionality
rules support what Joel Haycock calls "the analogous traits" of the
image, the contention that the pictured space is analogous to that in
which the spectator is situated.9 In this sense, they are preconditions
to the further analogy between the spectator and the figure or figures
shown. The resulting identification, supported by the decoupage, allows
the spectator to understand the space as the characters do. Thus a shot
of an actor looking offscreen-right is made to imply that the following
shot will represent a segment of the space located in that direction.
As Haycock argues, this structure promises that we, like the characters,
may look into, and in this way place ourselves within, the continuous
space extending beyond the boundaries of the frame.
The narrative illusion so constituted, the notion that the spectator
is witness to a complete world informed by the alternating subjectivities
of the characters and is able to identify himself with them, thus rests
on the spatial organization of the image. It is clear, given this dependency, that the continuity and directionality rules function as much to
discredit the image as a two-dimensional surface as to affirm its threedimensional reference: that is, they imply that a full acknowledgement of the complexity of the image is incompatible with the narrative.
The relation of the two, therefore, may be more accurately described
as opposition than subordination-a fact which at once defines "classicism" in narrative film and demonstrates the futility of attempting
to explain the image adequately within the terms of classical styles.
However, there exists a second category of narrative conventions,
which may be called "modernist," to which these conclusions do not
apply. The term "modernist" in this definition requires no demonstrable
9 Joel Haycock, "Radical Ideology in the Cinema: An Essay on Jean-Luc
Godard," Diss. Harvard 197I, pp. 52ff. My argument also owes a more general
debt to the position taken in this essay.

NARRATIVE AND THE FILM IMAGE

385

link to historical movements of the same name in painting, music, or

fiction, but simply designates film styles which encourage reflection on
their constituent means of expression, specifically their sounds and
images. It is thus opposed to classicism, which organizes the image so
as to obstruct reflection of this kind.
By definition, then, modernism resists the sort of practical or theoretical codification possible (perhaps even necessary) in the case of
its opposite. There has been no major spokesman for modernism in
film since Eisenstein and, of course, no rule book of modernist style.
Further discussion of image and narrative must, therefore, be conducted within the terms provided by an instance of modernist style.
Since modernism embraces a spectrum of styles, the film I propose to
cite-Godard's Deux ou Trois Choses que je sais d'elle, made in 1966has no more valid claim to be regarded as typical than, say, Zero de
Conduite or Muriel. Yet while in those films the conception of image
and narrative remains implicit, in Godard's it is conveniently made the
subject of speculation delivered on the sound track in the voice and
persona of the director.
The little narrative that there is in Deux ou Trois Choses que je
sais d'elle is organized around the movements of Juliette Janson, a
middle-class married woman and part-time call girl, during a twentyfour hour period. The "elle" of the title refers to her and, as the main
credits explain, to "the Paris region" where she lives and works. The
scene in which the voice-over commentary touches on issues closest to
the present discussion occurs almost midway through the action;
Godard, speaking in the whisper he uses throughout, introduces it with
a pair of questions on the problem of rendering narrative as image:
"How to give an account of events? How to show or say that on this
afternoon, at about 4: Io, Juliette and Marianne came to a garage at
the Porte des Ternes where Juliette's husband works?"10 Though
it is tempting to take these questions solely as theoretical speculation
on narrative, it is important first to specify the ways in which they are
bound to their particular dramatic context. For as Godard asks how
he may account for events, the accompanying sounds and images, as
much "his" as the voice, set to work making just such an account. He
speaks as if the "events" he mentions, like the Paris region, existed
independently of his film-as if Juliette and her friend Marianne, as
o My translation of Jean-Luc Godard, Deux ou Trois Choses que je sais d'elle
(Paris, I971), p. 64. Godard's word for events, evenements, is the same that Bazin
uses in the text cited in n. I: "jouee sur un theatre . . l'evnement continuerait
d'exister objectivement." The phrase "show or say" suggests the equal weight
given to image and sound in modernist films, and to language in Godard's.

?~,~1i
i~~
i :::
::
f:

::

''?e
'-~:'::'
....

,,

0-=...

O
t.

. A

NARRATIVE AND THE FILM IMAGE

387

a matter of historical fact, which it is his task merely to record, "came

to the garage." We of course understand this pretense as a convention
of fictional narrative. A pair of shots in the opening sequence of the
film assures us that this understanding is relevant here: in the first,
Godard shows the heroine, identifying her as Marina Vlady, the actress,
and describing her hair, clothing, and Russian origin, while in the next,
an almost identical view of the same woman, he tells us we are seeing
Juliette Janson, whose hair, clothing, and national origin he describes
in precisely the same terms.
These contradictions plainly establish that the whispered commentary is not to be understood as that of the historical Jean-Luc Godard,
whose statements can have the status of theoretical assertions, but of
the persona of director, whose speculations exist within a drama that
includes him and his efforts at composition: "Yes .... How to say
exactly what happened? Of course there is Juliette . . . there is her
husband . . . there is the garage. But are these really the words and
images one must use? These alone? Aren't there others? Am I speaking too loud? Am I watching from too far or too close?" (Deux ou
Trois Choses, p. 65). The shot accompanying this text (shown in
Illustration I) suggests no narrative criterion by which these questions
may be definitively answered. Too far for what? The conversation
between Juliette and her husband, which we see but do not hear,
could be shown with equal clarity in a tighter framing, though this
would have the effect of isolating it further from the background. The
relation of the background action-the attendant filling the tank, the
waiting cars-to the protagonists suggests other such relations possibly
visible in a wider shot. But since none of these functions in the narrative, the narrative gives no means of settling the question. It is true
that Godard prejudices matters by not allowing us to hear what is said.
Nonetheless, we recognize that his text validly points to the problematical and uncertain nature, admissible only in a modernist style, of the
relation of narrative to at least one aspect of the image-namely, scale.
Yet despite the uncertainty indicated by Godard's questions, the
accompanying shot is not, and probably cannot be, in any sense "uncertain" or "undetermined." This fact is explained in some degree by
the text. In asking whether he is too far or too close, Godard indirectly calls attention to the particular distance he has chosen and
thereby invites us to understand this choice as his responsibility. The
"modernism" of this stylistic gesture in fact may be said to consist in
the attribution of the choice to the director rather than to the internal
logic of the narrative. Though other modernist films may not follow
the strategy of explicitly dramatizing the agency of the director, their

388

NEW LITERARY HISTORY

narratives often in some way concern the act of making images: Blowup, 8 2, Muriel, The Decameron, and Le Petit Soldat are obvious
examples.
The shot I have been discussing begins with some action not evident
in the illustration. An attendant opens and closes the car door on
Juliette's side, apparently to ask her what she wants, then crosses
behind the vehicle to fill the tank; Juliette's husband enters and squats
in the position shown. The chosen framing is thus in part justified by
the action, since it allows us to see that the transaction between
Juliette and the attendant continues while she speaks with her husband. However, this logic serves only to raise a further question, since
the attendant is an anonymous figure in the drama, an extra, and this
piece of action has no narrative consequences.
This issue, unresolved here, is taken up four shots later by means of
the nearly static image (shown in Illustration 2) in combination with
the following text: "There is also another young woman, of whom
we know nothing. We wouldn't even know how to say that honestly"
(Deux ou Trois Choses, p. 65). Though the young woman is placed
in the frame in a way ordinarily reserved for the main characters of
the drama, she indeed does not appear elsewhere in the film. The text
stresses that it is not the director, the first person singular, who knows
nothing about her, but "we," the spectators. His pretended inability
to say this "honestly" jokes about the fact, unreported in the film, that
the young woman is the second assistant director, Isabel Pons. But the
text is more than just a joke. We do indeed know the sort of thing
about her that Godard in the opening sequence tells us about Juliette:
namely, the color of her hair and sweater, the expression on her face,
the movements of her head. The sort of thing we do not know is what
the narrative might tell us. If we were in a position to get Godard's
joke, we would understand that he knows something about her that he
can easily express neither in the image nor in the narrative. His text
and image stress instead that we are not in this position. Not only are
we ignorant of the specifics narrative might report, but narrative does
not even give us terms for stating this ignorance "honestly": that is,
for recognizing the sort of "knowledge" it promotes and suppresses.
Including himself among the spectators, the "we" of the text, Godard
in this rather opaque fashion suggests that narrative itself, quite apart
from the particularities of its execution, has a "problematic," a structure which encourages us to think only of certain things as in need, or
capable, of explanation.
The following shot (Illustration 3) and its commentary make a
parallel argument about the image: "There is also a cloudy sky, on the

NARRATIVE AND THE FILM IMAGE

389

condition that I turn my head instead of looking steadily in front of

me without moving. And inscriptions on the walls" (Deux ou Trois
Choses, p. 66). In order to understand the phrase "on the condition
that I turn my head," we must envisage the speaker present within
the real space of the filming. Our impulse to identify with him, to suppose that we too will see a cloudy portion of the sky if we turn our
heads, is obstructed by our situation as spectators. The sky that we do
see is not cloudy, but hazy blue. Godard does not qualify this point,
as he might have, by inserting a shot showing the cloudy sky. Instead,
he stresses the strangeness of having to look fixedly ahead and of having
one's view limited by the edges of the frame. In this way, he characterizes the image not as a segment integrable within a continuous
space, but an excision, differing in this respect from the view of a live
witness. The combination of text and image suggests a critique not only
of the rhetorical figure of Godard's phrasing, which identifies the
image with the director's gaze ("Do I watch from too close?" rather
than "Have I placed the camera too close?") but of the functional
identity, posited by classical styles, of image and glance.
A few slight movements in the shot allow us to relate this argument
to the psychological principle of narrative. After the phrase "I turn
my head," the station attendant, having finished lighting his cigarette,
turns his head to look toward the camera; after the phrase "without
moving," he turns his head screen-right. The relation of the first of
these gestures to what is said on the sound track invites a parallel between the attendant and the director, for what the attendant does the
director proposes to us he may do also. Our covert knowledge of
film production, the understanding that the "voice-over," as the phrase
implies, is an element added after the filming, enters into the interpretation, ensuring that we do not take the actor's movements as a response
to the director's words. Because the conditions of our vision as spectators differ as they do from the director's, the parallel between director
and character enforces the recognition of how our situation also differs from the character's. Here, too, Godard withholds any structural
gesture, such as a shot representing the attendant's point of view, that
might challenge this recognition. Instead, he insists that to be before
an image is to be deprived of the freedom to move one's gaze in the way
one does in space and to recognize one's dissimilarity to the figures
shown.
The spatial premises of this style thus drastically limit the possibility
of identifying with the characters of the drama. But at the same time,
they open the possibility of incorporating into the meaning of the discourse significant organizations of the image plane. For example, the

390

NEW LITERARY HISTORY

shot shown in Illustration 2 can, and must, be taken to mean that the
young woman is to be understood in relation to the objects-in particular, the advertisements-grouped around her in the frame. Since
we have no means of knowing exactly the size of the signs on the wall,
and since we do not see the plane of the ground linking figure and
wall, we have no certain sense of the space between them. Godard
thus strengthens the flattening tendency of the image by composing
the elements chiefly in reference to a single plane, an effect that is
even more striking in the color original, where the yellow and blue
signs are set against a white wall.
It is clear even without taking the analysis further that in this style
the classical chain of logic linking image to narrative is at its every point
opposed and questioned. If the dichotomy "classicism versus modernism" is understood (to use the language of structuralism) "diachronically," then it is sufficient merely to record these oppositions as differences. Some distinguished recent theorists, despite the synchronic
nature of their method, take this view. The philosopher F. E. Sparshott argues that critical differences about film cannot be resolved
by discussion of the properties of the image, which is sufficiently complex to support differing, even opposing, views. Christian Metz portrays film language as a heterogenous ensemble of elements, none of
which occupies a controlling place or authorizes a critical principle.11
In reply, I should point out that since modernist styles take account
of more aspects of the image, they are in a position to relate it to narrative in a more complex and more complete way. This is true not only
in cases like that of Deux ou Trois Choses, where the interrogation of
classical conventions is carried to exhaustion, but even of such films
as The Gospel According to Matthew, whose modernism consists
chiefly in certain discordantly documentary habits of cinematography,
which function, as their author observes, only "to make the presence
of the camera felt." 12
Completeness of this sort is not, however, a criterion that may be
simply applied in the judgment of films. It does not, for example, imply
that because Pasolini's style habitually calls attention to the presence
of the camera, his films are to be preferred to those that do not. Rather,
Metz develops this view most fully in Language et Cinema (Paris, 197 ).
I
Sparshott's essay, "Basic Film Aesthetics," is reprinted in Film Theory and
Criticism, ed. Gerald Mast and Marshall Cohen (New York, London, Toronto,
1974), pp. 209-32.
Pier Paolo Pasolini, "Le cinema de poksie," Cahiers du Cinema, No. 171
I2
(Oct. 1965), p. 64. Pasolini uses the phrase in a general discussion without
specific reference to The Gospel According to Matthew, though this film clearly
belongs to "the cinema of poetry."

NARRATIVE AND THE FILM IMAGE

39I

it suggests that the modernist model of image and narrative has a more
valid claim to set the terms for interpretation and criticism. This means,
at a minimum, accepting the premise that narrative and image in
film, like text and music in opera, are distinct strands, one of which is
not fully reducible to the other, and that critical positions, which, like
Bazin's, reject this premise are compromised at the root.
To apply this principle in the interpretation and criticism of classical
films is, obviously, to resist not only the manifest intentions of their authors but the assent given those intentions by spectators. The very
intelligibility of classical styles demonstrates that these intentions are
viable, that one may "succeed" in using narrative to suppress certain
attributes of the image. What, then, justifies this resistance? As the
sequence from Deux ou Trois Choses demonstrates, a given characterization of the image, two or three dimensional, bounded or "analogous,"
is neither a question of personal predilection, as Sparshott supposes,
nor an isolated choice, but rather entails a whole ensemble of propositions, not the least of which involve the relation of the spectator to the
image. Thus one cannot argue that the image can be determined
exclusively by style, if style is defined in isolation from the conditions
that circumscribe the relation of spectator to screen. Modernist styles
show that what is truly at issue in style extends beyond such exclusively
formal questions as whether the image is taken to be surface or spatial
reference to whether it is characterized as a material component in the
work of author and spectator. The rules of continuity and directionality suggest the extent to which classical styles are predicated on
denying this. The critic who acquiesces in this denial foregoes the
recognition not only that we, spectators and directors alike, live in a
material world, to which the cinema belongs, but that the conditions
of that world-in particular, the physical situation of the spectator
and the material attributes of the image as they originate in optical
and chemical processes and in human choice and labor-are pertinent
to viewing or making films.
Such a recognition requires neither hostility to the classical film
nor prejudice to critical questions that might be asked of it. Rather, it
simply proposes that as a first principle criticism treat the organization
of the image plane as significant, even when the style insists that it is
not, and understand that such significance originates not in style alone
but in the relation of the film image both to the space to which it
refers and to the conditions that bound its presentation to the spectator.
In taking account of these relations, interpretation, along with the
criticism based on it, necessarily outruns the terms of narrative. Even
when such criticism is in accord with the intentions of the director as

392

NEW LITERARY HISTORY

they are recorded in the style, it is justified not by them but by the
aspiration to understand and respond to narrative film in its complete
manifestation as a significant object.
HARVARDUNIVERSITY