Cinematic Free Indirect Style: Represented Memory in "Hiroshima mon amour"
Author(s): Leah Anderst
Source: Narrative , OCTOBER 2011, Vol. 19, No. 3 (OCTOBER 2011), pp. 358-382
Published by: Ohio State University Press
Stable URL: https://www.jstor.org/stable/41289309
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 support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Ohio State University Press is collaborating with JSTOR to digitize, preserve and extend access
to Narrative
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Leah Anderst
Cinematic Free Indirect Style:
Represented Memory in
Hiroshima mon amour
He: You saw nothing in Hiroshima. Nothing.
She: I saw everything . Everything .
- Duras 15 1
These opening lines of Alain Resnaiss Hiroshima mon amour inaugurate the films focus
on vision and highlight its distrust of vision as a form of knowledge. She, a French
actress, a tourist about to depart, wants to see Hiroshima. She wants to know the city,
the remains of its traumatic history; and the monuments it has erected to preserve
that history. He, a Japanese architect living in Hiroshima, knows that she cannot "see"
the city, cannot know its past through her vision. And so he disputes her vision of Hiroshima;
he disputes her knowledge of his city and his past. "From the first moments
of Hiroshima mon amour, therefore, the central epistemological question of the film is
laid out for us: how and what can we 'know,' through the epistemological instrument
of the gaze" (Craig 27).
Film narrates by "seeing" scenes and images, and it narrates with sound and with
editing. Hiroshima mon amour interrogates not just seeing but also cinematic agency
and narration, especially narrative representation of memory and history. This film
narrates memory, and it narrates history. In the fabric of its narration, however, in
its use of cinematic free indirect style, Hiroshima mon amour creates moments of
"plural narration," moments of dual or multi-visioned seeing where narrative agency
becomes uncertain, and the film thereby dramatizes the difficulty of capturing history
and memory through narrative.
Hiroshima mon amour tells the story of the brief love affair between an unnamed
Japanese architect played by Eiji Okada and Emmanuelle Rivas unnamed actress on
her final day shooting a film that takes place in Hiroshima. The film opens with a
Leah Anderst is a Visiting Instructor of Writing at Marymount Manhattan College. She completed her
Ph.D. in comparative literature at the Graduate Center of the City University of New York in June 2010,
and she was a 2010-201 1 fellow at the Writers' Institute at the Graduate Center.
NARRATIVE , Vol 19, No. 3 (October 2011)
Copyright 201 1 by the Ohio State University
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 359
fifteen-minute prefatory sequence, a series of scenes that juxtapose close-up images
of the lovers' bodies embracing with images of Hiroshima. The two characters speak
over the Hiroshima images, and their voices are "flat and calm, as if reciting" (Duras
15). With its combination of voice-over narration, documentary and fictional images,
and a score that at times complements the images and at other times conflicts
with them, this opening takes advantage of cinema's multiple tracks to move between
perspectives and distinct registers of discourse. This opening establishes the multiple
cinematic perspectives that will "see" and narrate this film.
The first portion of the opening sequence details what Riva saw while visiting
Hiroshima: a hospital, a museum visited four times, and a public square. The image
track matches her slow and deliberate voice-over narration. We see the hospital, the
museum, and the square while she describes them in voice-over. Among these images
are short intercut scenes of the lovers' bodies. Following a shot of the hospital
corridor, following an image of Peace Square outside the museum, following explanations
and ephemera related to the dropping of the bomb, and following reenactments
and archival newsreels documenting the bomb's immediate aftermath, the scene cuts
to extreme close-up shots of the intertwined arms of the two lovers. Complementing
this visual return is a corresponding return on the sound track. This opening
sequence begins with a slow, melancholy series of notes that becomes the musical
"theme" of Riva and Okada's relationship. Each time the image cuts away from Riva's
descriptions of Hiroshima and back to their lovemaking, the sound track marks the
same change, even if only for the duration of two or three notes.
In one of the newsreels of this opening, injured children and adults receive care
in a crowded hospital, and Riva's voice-over narration remains present over these
images. Nurses and doctors lightly dab the victims' deep wounds and burnt skin with
small pieces of cotton while Riva speaks. "Hiroshima was blanketed with flowers" she
says, "there were cornflowers and gladiolas everywhere, and morning glories and day
lilies that rose again from the ashes with an extraordinary vigor, quite unheard of for
flowers till then" (Duras 19). As we watch these victims receiving care, Riva describes
a city covered in flowers, a city in the process of rebirth rather than a city destroyed,
and her words no longer correspond to the images. The contrast between her words
and the images casts doubt over her testimony, over her ability to see and to know.
Following Riva's description of Hiroshima's flowers is a medium-shot of a woman's
head (Figure 1). Two pairs of hands insert medical instruments into her left eye
to open the lid, revealing an empty socket as they turn her head toward the camera.
The missing eye of this victim recalls the film's opening line: "You saw nothing in
Hiroshima. Nothing." Riva's words over this image belie her awareness of the limits of
her seeing and her memory: "Just as in love this illusion exists, this illusion of being
able never to forget, so I was under the illusion that I would never forget Hiroshima.
Just as in love" (Duras 19).
Hiroshima mon amour was conceived as a documentary that Resnais would direct
along lines similar to his earlier documentary short, Night and Fog (1955). The
inclusion of Marguerite Duras as the film's screenwriter following Resnais's frustration
with the project lends the film "a fictionist's perspective" (Moses 161). For her part,
Duras refers to the film in the prefatory synopsis of her screenplay as a "false docuThis
content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
360 Leah Anderst
Fig. l.
mentary" (10). In addition to newsreels shot inside a Hiroshima hospital, this opening
sequence includes documentary images of decimated neighborhoods, of protests
and marches, of mass disposals of fish and other foods that the people now fear, of
tours and sites now visited, and of the city's streets and markets. In his monograph on
Resnais, James Monaco notes, "fully the first fifteen minutes of the film are given over
to the documentary-that-never-was" (38). Hiroshima mon amours opening "sees," in
part, with this documentary eye, with the agency of a third-person, documentary
narrator. The remainder of the film narrates the fictional story of the two lovers, and
the images alternate primarily between a non-documentary, third-person camera and
an internal, subjective camera. This documentary eye, however, will return; it will
return to combine with Rivas subjective perspective to form one variety of this films
cinematic free indirect style.
Though the city, its traumatic history, Rivas limited apprehension of that history,
and the city's current geography play a crucial role in this opening and in the film more
generally, the films primary subject is Rivas narration of the memory of her love affair
with a German soldier during the occupation of France some years earlier. Over the
course of one day, Riva tells Okada the story of this relationship that took place in Nevers,
her childhood home, when she was eighteen years old. She tells him of her desire
to flee France with her lover for Bavaria, she tells him of her lovers death upon Frances
liberation, and she tells him of the subsequent humiliation and madness she suffered,
branded a traitor by her town and forced into captivity in the basement of her parents
home. This memory is one that she has kept to herself, that she has, until now, resisted
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 361
telling. As the film reveals in a late scene, Riva equates narrating this story with forgetting
and with infidelity to her early lover, and she evinces a profound ambivalence
vis-a-vis this narration. "[T]he prototypical modern psyche split in two," Riva tells her
story, but as she tells, she holds back (Cohen 63). As she reveals, she conceals.
The film itself, famous for its use of flashbacks, evinces a similar hesitation. The
film narrates Rivas memories, but as it narrates, it wonders whether its task is possible.
As it presents many instances of represented memory, the film wonders whether
narration can "remember" memory and history. The viewer "sees" Rivas visit to a
museum in Hiroshima, we momentarily "see" her memory of her first lover dying
from a bullet wound, we "see" the Nevers of her youth and her memories of her love
affair there, and we "see" her as a young woman, a prisoner in her home. Watching
these scenes, the viewer imagines himself seeing from Rivas internal first-person
perspective; the viewer imagines seeing what Riva sees as she recalls her past. In portions
of these memory sequences, however, there is evidence of external perspectives,
perspectives that are not Rivas own. Representing these memories, then, Hiroshima
mon amour creates instances of cinematic discourse in which singular narrative agency
cannot be determined, instances of shared cinematic discourse that bear strong
resemblances to prose free indirect style (FIS - I shall use "FIS" rather than "FID,"
where the "D" is for "discourse" because "FIS" better captures the multiple tracks
of film communication). Unlike novels and stories, however, which most often operate
over only a single, verbal track, film is a multi-track medium. Agency within
the cinematic medium may change from scene to scene but also from one track to
the next. The image we see on the screen may originate with a third-person external
perspective while the soundtrack we hear embodies the first-person perspective of a
character. Like FIS in prose, with its careful blending of first-person and third-person
narration, the instances of represented memory in this film combine multiple agencies,
multiple perspectives, within single shots and single scenes but also over the medium's
multiple tracks. Cinematic FIS is just this: within single instances of cinematic
narration, narrative agency, narrative perspective is multiplied, and, because of this
proliferation of perspectives, uncertain.
PROSE FREE INDIRECT STYLE,
CINEMATIC FREE INDIRECT STYLE
In a brief discussion of Gustave Flaubert in S/Z, differentiating his writing from "
sic writing," Roland Barthes notes, "Flaubert, however, (as has already been sugg
ed), working with an irony impregnated with uncertainty, achieves a salutary dis
fort of writing: he does not stop the play of codes (or stops it only partially), so
(and this is indubitably the proof of writing) one never knows if he is responsibl
what he writes (if there is a subject behind his language); for the very being of wri
(the meaning of the labor that constitutes it) is to keep the question Who is speakin
From ever being answered" (140; emphasis original). Flaubert's well-known des
to achieve a level of impersonality, whereby the author is "invisible et tout-puiss
quon le sente partout, mais quon ne le voie pas" [invisible and all-powerful; eve
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
362 Leah Anderst
where felt, but never seen], undoubtedly contributes to this inability to definitively
answer the question Barthes raises (Flaubert 324). Flaubert's novels tend to resist
fixed agency, and his impersonal narrators provide his narratives with little moral
grounding. Flaubert also achieves what Barthes describes as a "salutary discomfort of
writing," a writing for which 11 one never knows if he is responsible for what he writes
with his frequent use of FIS. FIS in prose is a combination of first-person and thirdperson
narration that, in Dorrit Cohn's words, "[renders] a characters thought in
his own idiom while maintaining the third-person reference and the basic tense of
narration" (100).
A brief passage from Flaubert's Sentimental Education illustrates FIS. Toward the
end of that novel, Frederic rejects Marie Arnoux, his perpetual yet inaccessible love
interest, just as she arrives to offer herself to him, "Frederic soup^onna Mme Arnoux
d'etre venue pour s'offrir; et il etait repris par une convoitise plus forte que jamais
Une autre crainte l'arreta, celle d'en avoir degout plus tard. Dailleu
serait! - et tout a la fois par prudence et pour ne pas degrader son
ses talons et se mit a faire une cigarette" (454; emphasis mine). [
that Madame Arnoux had come to offer herself to him; and onc
with desire
what a nuisance it would be! And partly out of prudence and partly t
his ideal, he turned on his heel and started rolling a cigarette" (455
The phrase uDailleursy quel embarras ce seraitV communicates wor
locates within Frederic's mind. Flaubert's largely impersonal narra
exclaim or emote in this manner, yet Flaubert embeds this senten
rator's description without a signal phrase such as "he thought." F
words of the character within those of the narrator so that the two
taneously present and so, to again quote Barthes, the reader canno
speaks," who, of these two agents, is "responsible" for the discours
Though film does not operate within a grammatical system tha
FIS immediately visible and identifiable as it often is in prose, film
use of FIS with its attendant uncertainty and irony. Hiroshima mon
combination of documentary and fictional narrative techniques e
opening sequence and with its mixing of distinct agencies deployed
tracks of the cinematic medium, creates an ideal viewing situation
of the distinction between perspectives. A number of shots and sc
memory sequences in particular leave the viewer unable to answer
to Barthes's regarding Flaubert: Who sees? Who or which agency i
the film's narration?
A handful of film critics and theorists have examined cinematic
bear resemblances to prose FIS. Jean Mitry's "semisubjective ima
and Psychology of Cinema, George M. Wilson's "indirect or reflecte
Narration in Light: Studies in Cinematic Point of View, and Pier Pa
indirect subjective" in "The Cinema of Poetry" all point in the d
multi-visioned cinematic narration.
Mitry is primarily concerned with external vision, with the combination of a
character's optical perspective and that of an external cinematic narrator. Mitry's examples
of "semisubjective images" comprise those moments in film when the camera
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 363
records what a character sees while that character is yet in the frame. "[I]n order
to 'experience' the feelings of a given character, all the audience had to do," Mitry
writes, "is be with the character, alongside him. Seen objectively, the character could
then assume the responsibility and motivations of a shared point of view. Thus instead
of the camera taking the place of the character, there were images framing the hero,
either from head to toe or from the waist up, following him as he moved, seeing with
him and at the same time as him. The image remained descriptive but shared in the
character's point of view" (215; emphasis original). These images combine the perspective
of a character, the "character's point of view," with that of a "descriptive" narrator.
In cases such as those Mitry describes, the camera often films over the shoulder
of an actor so that a portion of the actor's body, usually his head, remains on screen.
The viewer simultaneously sees what the character sees and sees the character in the
act of seeing. While this variety of cinematic image resembles the prose FIS in which
a character's spoken words are combined with the discourse of the narrator, I am
here concerned with the cinematic representation of internal rather than external
vision, the representation of thoughts, and more particularly, the representation of
memories.
Unlike Mitry's "semi-subjective image," George M. Wilson's "indirect or reflected
subjectivity" maintains a metaphorical relationship with the mind portrayed. In Wilson's
conception, "features of the projected image or the mise en seine are used to
depict or symbolize or reflect aspects of the way in which the character perceives
and responds to his or her immediate environment" (87). Such an image acts as a
metaphor for a character's perception. The doorman protagonist who is demoted to
bathroom attendant in F. W. Murnau's The Last Laugh , for example, is often recorded
from below at the film's beginning, an angle that emphasizes his height, his pride,
and the importance he attaches to his "front of the house" position. Following his
demotion, however, the camera frequently views him from above, lessening his visual
impact and rendering him small and insignificant within the frame. In these images,
these examples of "indirect or reflected subjectivity," "features of the projected image .
. . symbolize" the character's perceptions of himself at different stages of the film (Wilson
87). Although this definition of "indirect or reflected subjectivity" highlights the
"shared" aspect of the image, the image is third-person with certain of its "features"
simultaneously "symbolizing" the perspective of a character. As a result, Wilson's
concept is not particularly fit for representing precise thoughts or memories passing
through the mind of a character. Wilson's "indirect or reflected subjectivity" lacks the
sense of immediacy, the sense of direct access to a mind so typical of prose FIS and of
the cinematic FIS I will here consider.
In "The Cinema of Poetry," Pasolini describes FIS in this way: "[T]he author penetrates
entirely into the spirit of his character, of whom he thus adopts not only the
psychology but also the language" (549). This description shares similarities with my
conception of FIS. Substitute the word "narrator" for Pasolini's "author," and we see
that FIS in prose frequently appears when a narrator, such as Flaubert's in the above
example, momentarily subsumes the internal voice of a character into his own so that
the two seem to speak together. For Pasolini, however, there is not a stylistic distinction
between what he calls "interior monolog" and what he renames "free indirect
subjective." These two modes "look" the same on screen. Their distinction lies in the
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
364 Leah Anderst
difference or similarity in situation between the filmmaker and the character. "[T]
he interior monolog is a discourse relived by the author," Pasolini explains, "through
a character who is, at least ideally, of the same class and generation" (550). Pasolinis
"free indirect subjective" occurs, on other hand, when a filmmaker produces a film
in which the focalizing character lives a life totally distinct from the filmmaker s own
experiences. Whereas Pasolinis conception of "free indirect subjective" relies on the
social conditions of the filmmaker and his character, I am concerned here to describe
a cinematic FIS that is independent of the biography of the filmmaker.
Building from Pasolinis "free indirect subjective," Gilles Deleuze briefly explores
the formal characteristics of this mode in Cinema 1 The Movement Image. Deleuze
begins his remarks by distinguishing the subjective image from the objective image:
"It could be said that the subjective-image is the thing seen by someone qualified,' or
the set as it is seen by someone who forms part of that set
when the thing or the set are seen from the viewpoint of someone
nal to that set" (71). Deleuze correlates the subjective image with d
the objective image with indirect discourse (72). Deleuze rephras
tion of "free indirect subjective" in this way: "[T]he camera does
the vision of the character and of his world; it imposes another
first is transformed and reflected" (74). Following Deleuze s des
FIS is dual-visioned in the same way that prose FIS is "dual voic
on the formal qualities of Pasolini s discussion, and his definition
tant framework for identifying and characterizing cinematic FIS.
In addition to Deleuze s reconsideration of Pasolini, my concep
FIS borrows from Bruce F. Kawins exploration of first-person cin
Mindscreen: Bergmany Godard and First-Person Film. A method f
jectivity," "mindscreen" is narration that portrays what a character
"the field of the minds eye" (Kawin 10). During "mindscreen," th
agency to the character whose visual thoughts are represented on
rator of mindscreen is offscreen; the indicators of his presence
Kawin cites Hiroshima mon amours well-known "hand twitch" sce
"mindscreen." Early in the film, Riva watches her Japanese lover sle
ter they meet, and the sight of his twitching hand as he awakens p
the films image track- a cinematic Proustian involuntary memor
the similarly twitching hand of her German lover, slowly dying f
on the banks of the Loire. The viewer "reads" the image on the s
Riva as a younger woman embracing her dying lover, as a visuali
sees in her mind s eye while she watches her lover in Hiroshima. Ri
"the agency" of this image, and as the narrator she is "offscreen"
bered self is onscreen (61). Kawin also cites instances from the
of Last Year at Marienbad and Hiroshima mon amour as "minds
fluid tracking shots through the corridors at Marienbad or the h
seum at Hiroshima are movements of the mind more than of the
"Mindscreen" is a prominent narrative mode in Hiroshima mon
however, complicates agency and vision by infusing key instance
narration, a first-person, subjective mode, with markers of third-p
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 365
perspectives external to Rivas. In these moments of the film, "the camera does not
simply give" Rivas vision of her "world; it imposes another vision in which the first is
transformed and reflected" (Deleuze 74).
REPRESENTED MEMORY IN HIROSHIMA MON AMOUR
Hiroshima mon amour includes many instances of filmed memory sequences,
instances of "mindscreen." The "hand twitch" scene that Kawin references is on
early instance. This moment is short, lasting less than five seconds. Later, as Ri
rates her memory to Okada, the film presents the viewer with images that corre
to her tale. She tells him of her relationship with the German soldier, of her disg
and of her madness, and the film "envisions" the memories she speaks. The
that accompany Rivas tale of her love affair, in particular, feature highly amb
combinations of distinct cinematic agencies, and it is here that I will examin
film's use of FIS.
As Riva and Okada lie together in his bed late into the afternoon on the
after their first meeting, Okada asks only one question that prompts Rivas na
of her first love affair: "Was he French, the man you loved during the war?"
47). Rivas response consists of only a few sentences, the basic structure of a na
without details or embellishment: "No ... he wasn't French
... At first we met in barns. Then among the ru
else
Riva narrates the affair to Okada, and in
lipses in the above citation, the scene cu
home to images of Nevers, images of Riv
with the German soldier.
The camera films Riva and Okada fro
highlights the easy intimacy of their br
ing throughout the sequence. In the alte
shown of Nevers, there is a strong distinc
era remains very close to Riva and Oka
shots of Nevers from a greater distance f
style highlights the intimacy between R
of intimacy in the manner that Nevers
man you loved during the war?" and Riva responds, "No
(Figure 2). From Riva and Okada embracing in bed, her hair
lips, the scene cuts to an image of a soldier walking toward
square outside of a cafe (Figure 3).
The camera films from inside the cafe, watching the soldi
a window. Over a few brief moments, the soldier walks in th
without revealing any particular intention, never looking at t
having as if he recognizes or notices anything. Reminded of
mentary photographs of the film's opening sequence, Siobha
particular shot of the solider in Nevers as a "quotation" of th
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
366 Leah Anderst
Fig. 2. "No
Fig. 3. Nevers Public Square
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 367
his brief shot," Craig writes, "with its quotation of the authoritative genre of the documentary
film embedded diegetically in the 'memory' of a woman we know to be - or,
at least, to have been- mentally unstable . . . appears deliberately ironic" (31). Craig is
correct to note the strong stylistic similarity between this shot of the soldier in Nevers
and the film's prior use of documentary footage, and, indeed, irony frequently results
from the combination of narrator and character discourse in prose FIS. The presence
of FIS in Flaubert's Sentimental Education and, to a greater extent, Jane Austen's
Emma , very often results in an ironic reading of those novels' protagonists. In the
case of this film, however, more than irony and more than a reminder of Riva's mental
state, this shot and those of this memory sequence convey discursive uncertainty.
Who "sees" this image of a soldier as he crosses the square?
For the simple fact that the image follows Riva's words about her former lover,
the viewer links the image with Riva. The viewer "reads" this man as the lover she
speaks of, and we place this image in her mind. The viewer "reads" the image as
"mindscreen": it appears to be a flashback that represents what she sees in her mind's
eye as she tells Okada that her lover during the war was not French. As Craig notes,
however, the camera records the soldier from an unidentified, apparently objective,
third-person position inside a cafe. Even as the film prompts a first-person reading,
nothing of the shot itself suggests Riva's first-person perspective, and the question of
agency surfaces. Who or what is this shot's originator, its agency? Does it envision
Riva's mind's eye, or is it the perspective of an external cinematic narrator?
The soundtrack poses an additional complication. When this scene begins, the
music that opened the film, the lovers' "theme" introduced in the film's prefatory sequence,
returns here. In this memory sequence, this music remains constant over
the shots of Hiroshima and over those of Nevers. The soft, slow music we hear as we
watch Riva and Okada together in bed continues over this image of the German soldier.
Whereas the cut between Hiroshima and Nevers in the image track underscores
the strong visual contrast between the shots, the continuity in the soundtrack builds
an aural link. As this scene and the film continues, this aural link will build a connection
between these two "impossible" love affairs, as Riva will later characterize them
both, but in this individual shot, the consistency of the soundtrack serves as a strong
marker of Riva's perspective. With its "quotation" of the documentary filming techniques
introduced earlier in the film, with its placement following Riva's description
of her lover, and with the soundtrack linking the two shots together, this image of the
soldier in Nevers presents a strong example of cinematic FIS. In this one instance of
cinematic narration, the viewer finds evidence of multiple perspectives: Riva's subjective
perspective, her memory, and the film's "documentary" narrator. Both are here
present; neither predominates, and agency is uncertain.
This shot of the soldier fades out very slowly while an image of Riva and Okada
fades in. Sitting up now, he is positioned just above her as she remains lying on her
stomach. As he and the viewer watch her, Riva continues her narration with another
very brief line, "Yes, it was at Nevers," and again, the scene cuts to Nevers. Riva is now
present in these images. She is younger. Her hair is long, and, over a series of seven
shots, she rides her bike in the streets of the village and in the surrounding countryside
(Figures 4 and 5).
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
368 Leah Anderst
Fig. 4. Biking in Nevers.
Fig. 5. Biking in Nevers.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 369
The melancholy soundtrack, the lovers' "theme," remains, again tying these
"present" and "past" images together. In these shots, Riva rides her bike toward or
away from a fixed camera recording her. "[T]he flashbacks become lengthier," Craig
writes of these and subsequent memory images in this scene, "but no less disorienting:
scenes of countryside rush past as if we were travelling at speed" (31). The camera
films from a distance that prevents the viewer from fixing on Rivas face. She is a small
figure, and the viewer can neither read emotion from her expressions nor perceive
what or as she perceives. Though these shots correspond to Rivas narration, "Yes, it
was at Nevers," the camera films from a position that appears not to engage her mind's
eye. As she speaks to Okada, does Riva chiefly recall her bicycle rides in and around
her village? And if so, do Figures 4 and 5 represent how she "sees" those memories
in her mind's eye? Rather than conveying the subjective perspective that the viewer
expects from the alternation between the present in Hiroshima and the flashback to
Nevers, the camera does not appear to penetrate her consciousness of this memory.
It does not reveal the emotions and the desires she doubtless remembers having
experienced on these bike rides traveling to meet her lover. As in the earlier shot
of the soldier walking in an outdoor square, these images resist Rivas perspective.
Though third-person, however, this sequence of bicycling images no longer "quotes"
the documentary eye of the films opening as the shot of the soldier does. Carefully
and artfully composed as they follow Rivas ride, these images appear to be "seen" by a
third-person fictional eye. These are "objective" images according to Deleuze's definition;
they appear to be "seen from the viewpoint of someone who remains external"
(71). Though these images follow and illustrate her words to Okada, and though the
music on the soundtrack continues to link Nevers to Hiroshima, these images do not
"look like" Rivas visual memories of her love affair. Over these images, Resnais has
"impose [d] another vision in which" Rivas "is transformed and reflected" (Deleuze
74). Resnais films these moments of "mindscreen" with a third-person camera.
This series of bicycling shots in Nevers continues with Riva riding by the banks
of the Loire, and a curious shot retrospectively places these images within the vision
of the German soldier (Figures 6 and 7).
As Riva rides, the camera pans with her along the river (Figure 6), but rather
than continuing to follow her or cutting away to another shot, the camera follows its
own path, panning quickly over a blurry landscape until it finally arrives at a medium
shot of the soldier (Figure 7). The camera stops abruptly to capture him in profile,
and the viewer imagines him watching her continue out of view. When these bicycling
shots begin, the viewer wonders who sees Riva riding around Nevers. Do these
highly distanced shot envision her memory of this story? As she speaks to Okada, is
this what she imagines? These images appear, on the other hand, similar to the films
third-person vision. The shot of the soldier looking further complicates the viewer's
sense of agency. Now, the viewer retrospectively assigns agency to him. He sees her
riding her bike in Nevers. He watches her as she traces a circuitous path around the
countryside. The viewer sees what he sees. Though this ascription of agency appears
to "work" for these shots, the perspective of the German soldier strikes a strange note
mixed as it is within Rivas tale to Okada so many years later. Just as quickly as Resnais
picks up the soldier s point of view in this shot, however, he abandons it for another.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
370 Leah Anderst
Fig. 6. Riva bikes by the Loire.
Fig. 7. The soldier watches Riva.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 371
Fig. 8. Wide shot of a Nevers field
The camera position of the final shot of this bicycling sequence (Figure 8) suggests
the perspective of an external vision, a third-person, objective cinematic narrator.
Positioned far above a large open field with a few rows of small trees and bushes
and flanked by a dense forest, in this shot a fixed camera records Riva on her bike
as she rides down into the field where the soldier waits for her at an opening in the
row of trees and vegetation. The camera remains still as she nears him. Riva and the
soldier are the visual points of interest in this shot- she is in motion and he is her
destination- but they are small against a deep and imposing background. This image
fades very slowly back to Riva and Okada in bed in Hiroshima moments before she
reaches the German soldier. If the viewer can tentatively and retrospectively ascribe
agency to the soldier in the early shots of this biking sequence, here that ascription
fails. Like Riva, the soldier is a tiny figure within the frame, and he cannot now be the
viewer's "eyes." Rather, this shot, embedded within Rivas memories, emerges from the
perspective of an external, fictional narrator.
The shots of this memory sequence thus far, these flashbacks that illustrate the
words Riva speaks to Okada while the camera films them in Hiroshima, visually
narrate her story. Intercut within her words to him, their placement, their context,
prompts the viewer to assign Riva as the "narrator" of these images. They are instances
of "mindscreen" that indicate her first-person perspective; "the narrator of mindscreen
is offscreen; the indicators of his presence are contextual" (Kawin 61). However, in
these instances of represented memory, "the camera does not simply give us the vision
of the character and of his world; it imposes another vision in which the first is transThis
content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
372 Leah Anderst
formed and reflected" (Deleuze 74). "Imposed" over these images is "another vision,"
the vision of an agency external to Riva, an agency that does not form part of the set.
This external vision at times recalls the documentary images and newsreels of the films
opening sequence and at other times appears similar to the films fictional third-person
vision. The film complicates the viewer s "seeing" of these memories; it complicates the
visual and aural narration of Rivas memories and of her past.
Riva continues with her tale- "At first we met in barns. Then among the ruins.
And then in rooms like anywhere else" (Duras 48) - and the scene cuts again to images
of Nevers. In these shots Riva and the soldier are together, and they are filmed
entering various structures and spaces. As in the previous sequence, in these shots
the camera stands at a far distance from the couple as they climb into a door- sized
opening in the stone wall surrounding the city, as they enter an abandoned shack
within a large field, and as they embrace standing among ruins in the countryside.
Again, these shots purport to belong to Rivas memory. They match her description of
the places where she and her lover meet, and the cut from the films "present" to these
shots signals Rivas first-person memory, "mindscreen." Again, however, the filming
of these scenes suggests an external agency, Deleuze s "objective image," disconnected
from Rivas vision.
The opening shot of this sequence (Figure 9) in particular suggests a third-person
perspective. After Riva speaks, the scene cuts to a shot of large brick houses. The
camera is positioned well below the houses and on the opposite side of a stone wall.
The image remains on the houses for only a few moments before the camera pans
down along the wall until Riva and the soldier, on the ground below, come into view
(Figure 10). They climb into an entrance near the bottom of the wall; he helps her step
onto her bike and up into the entrance. The initial shot of the houses functions as an
"establishing shot," a shot traditionally impersonal and third-person. The shot tells
the viewer that the two lovers sneak into abandoned spaces beyond the walls of the
city - their relationship is forbidden and secret- but nonetheless very close to Nevers.
Rather than envision her minds eye, the style of this shot resists Rivas internal,
first-person perspective. Riva and her lover are very small within the frame while
the city's wall and the houses tower over them. The remaining shots of the sequence
illustrating the various locations where Riva and her lover meet all behave similarly.
Holding hands, the young lovers enter a small wooden shack while an unmoving
camera records them from across a large, wide field. The camera films them from a
far distance standing and embracing in the ruins of a stone structure, her hair and
her skirt blowing in the wind. While these shots convey the intimacy and the danger
of this secret relationship, the images themselves, with Riva and her lover positioned
at such a great distance from the camera that records them, do not communicate a
strong sense of Rivas memory to the viewer. The viewer does not "see" these images
as "mindscreen." These shots, like the earlier images of Riva riding her bicycle, appear
to be seen by an "objective" eye external to the scene.
As this sequence illustrating Rivas narration to Okada approaches its end, an
image that follows Rivas penultimate words appears to correspond less directly to
the words she speaks. Following her description of the places where she and her lover
met, she says, "And then he was dead," and the image below (Figure 11) appears on
the screen (Duras 48).
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 373
Fig. 9. Village Wall, Nevers.
Fig. 10. An entrance in the village wall.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
374 Leah Anderst
Fig. 11.
For the first-time viewer, there is no apparent relationship between this image
and the soldiers death. Not only does the viewer here wonder who sees, the viewer
also wonders what is seen. In a later scene, the film reveals the identity of this structure.
Rivas lover was shot as he awaited her by someone who stood at this outdoor
terrace overlooking the Loire. Riva and her lover were to meet there to flee together
for Bavaria where they would marry. She rushes down to the river bank to join him
where he lies prostrate and injured. She reaches him, and, looking up and around in a
panic, this is the image she would have seen. She would have seen this outdoor terrace
viewed from below just as it is shown here. Within this memory sequence, this image
is perhaps the only instance of "mindscreen," but without the requisite information,
the viewer cannot see it as such. The viewer "reads" the first-person perspective of
this image only retrospectively. Agency here may not be dual or multiple, then, but
ambiguity renders it uncertain.
Riva reaches the end of her narration as she tells Okada, "I was eighteen and
he was twenty- three" (Figure 12) (Duras 48). The scene cuts from an image of Riva
reclining on her side to the Nevers flashback (Figure 13). As the film reaches this
portion of the memory sequence, the soundtrack, the "lovers theme," which has remained
constant throughout this back and forth alternation between Hiroshima and
Nevers, ends with an image of the younger Riva closing her piano (Figure 13).
This image illustrates Rivas inclusion of her age as the final point in her narration,
but it serves an additional purpose: it ties the soundtrack to Rivas first-person
perspective. This image retrospectively renders the soundtrack diegetic. As she recalls
these memories during her narration to Okada, this image (Figure 13) suggests that
the soundtrack heard by the films viewer is also what Riva hears in her "mind s ear"
as she speaks. While the images of this memory sequence resist Rivas first-person
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 375
Fig. 12. aI was eighteen . .
Fig. 13. Lovers' theme ends.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
376 Leah Anderst
perspective in favor of a largely third-person perspective, the soundtrack retains her
subjective perspective.
Marguerite Duras notes in the preface to her screenplay for Hiroshima mon amour
that Resnais asked her to "annotate" the films sequences that take place in Nevers. She
wrote these annotations after the completion of the screenplay but prior to the films
shooting. Durass annotations are included as part of an appendix to the published
screenplay, and she titles them "Nocturnal Notations (Notes on Nevers)." The following
passage is excerpted from Durass annotations on the memory sequence I have
been examining: "We kissed behind the ramparts. Deathly afraid, but utterly happy,
I kissed my enemy. ... I discovered his hands when they touched the gates to open
them before me. I soon wanted to punish his hands. I bite them after making love.
... I no longer remember the gate at the end of the garden. He waited for me there,
sometimes for hours. Especially at night. Any time I could find a free moment. He
was afraid. I was afraid" (88). Durass annotations present the love story the viewer
might expect from Rivas character. The annotations convey the character s conflicting
emotions, her fear as well as her joy, "deathly afraid, but utterly happy," and they
include intimate details that a lover might recall, "I bite [his hands] after making love"
(88). When Riva tells her story to Okada, and when this memory sequence is filmed
and edited, however, these personal, emotional details are largely absent from the
films narration. Rivas words to Okada are clipped and limited. She speaks only the
most basic details of her love affair, and the film envisions her memories similarly. The
film imposes perspectives external to Rivas mind over these images so that together
with her vision, her "mindscreen," the images carry markers of perspectives external
to Riva. The film narrates this memory, then, but even as it does so, it resists showing
her story; it resists her solitary perspective, the expected perspective. Resnais films
and edits the scene, instead, in a manner that suggests cinematic FIS, a manner that
suggests a plurality of perspectives, perspectives combined so that the viewer cannot
know who sees just as the reader of prose FIS cannot know who speaks.
After the lovers "theme" ends with the image of Riva closing her piano, a new,
playful piece of music begins, and the image cuts to a series of shots of Riva and the
German soldier, their faces hopeful as they run to meet one another beyond the walls
of Nevers. In this second and final portion of this memory sequence, Riva no longer
narrates to Okada. She no longer speaks, and the image track no longer cuts between
Hiroshima and Nevers. Rather, the shots alternate quickly between the German soldier
waiting, watching, and signaling excitedly to his lover and Riva running through
fields and forests and climbing over hills and fences before the two finally meet and
embrace. Riva no longer tells her story, but the film continues its narration. In these
shots the viewer sees Rivas face and her hopeful expressions as she approaches her
lover, the viewer sees the happiness on both of their faces as they rush into each other
s arms and embrace, and the viewer sees the lovers in close-up as they lie together
on the ground in a barn. Unlike the earlier memory shots, these the viewer "reads" as
"mindscreen." The subjects in these images, their increasing proximity to the camera,
the upbeat, exciting soundtrack, and the rapid speed of the editing together communicate
Rivas perspective, Rivas vision of these memories.
Rather than analyzing these images individually, I have cited them in a block
below (Figures 14-19) in order to better convey their swift progression.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 377
Fig. 14.
Fig. 15.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
378 Leah Anderst
Fig. 16.
Fig. 17.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 379
Fig. 18. Lover's theme returns.
Fig. 19.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
380 Leah Anderst
These images, these "mindscreens," convey Rivas memory of one meeting between
her and the soldier. She recalls and imagines the anticipation she felt as she
rushed to meet him, she recalls the physical barriers she had to traverse to arrive at
their meeting point, she recalls seeing him signal excitedly to her as she neared, and
she recalls their running embrace and the time they spent together, sheltered from
outside eyes in a cold barn. Much more so than the images of the first portion of this
memory sequence, these carry markers of Rivas vision, of her mind s eye, and the
viewer has the sense of seeing with her, of seeing what she imagines as she remembers.
Few novels or stories retain FIS over lengthy blocks of narration. More often,
brief instances of FIS are found within indirect or direct discourse. In the example
cited from Flauberts Sentimental Education , "Dailleursy quel embarras ce serait!" [Besides
what a nuisance it would be!], is surrounded by the narrators indirect report of
the characters thoughts and actions. FIS in prose is a narrative mode that is turned
to quickly but just as quickly turned away from, and this tendency is no different in
film. Resnais's film does not maintain cinematic FIS throughout this entire scene.
Rather, the films narrative moves from an objective camera to cinematic FIS, back
to objective, and finally on to subjective narration, to "mindscreen." Reader recognition
of prose FIS depends on contextual clues as well as on the modes distinction
from direct and indirect discourse. Cohn, who proposes the term "narrated monologue"
rather than FIS, writes that w[t]he narrated monologue is thus essentially an
evanescent form, dependent on the narrative voice that mediates and surrounds it,
and is therefore peculiarly dependent on tone and context" (116). Cinematic FIS is
also "evanescent" and "peculiarly dependent on tone and context" (116). The viewer
recognizes the moments of shared cinematic discourse, the moments of FIS, in this
memory sequence because the film earlier established distinct discourses that are here
combined. The viewer also recognizes FIS because the scene encourages its viewer
to "read" the memory images of the sequence as if from a first-person perspective.
Cinematic FIS, perhaps more so than FIS in prose, is, then, "peculiarly dependent
on" viewer expectations. Resnaiss viewers, having already "seen" Rivas memories
portrayed on the screen, expect the images accompanying her telling of this memory
to be first-person, subjective images rather than the largely third-person images the
film presents.
CONCLUSION
"In Hiroshima mon amour" Michael S. Roth writes, "[Resnais] has the . . . diffi
task of projecting forgetting onto the screen- first, the forgetting of historical me
ory, the withdrawal of the destroyed Hiroshima from our consciousness; second,
forgetting of personal memory, the evaporation of the traumatic memory of love
the woman in the film; and third, the connection between forgetting and narrat
(94). Later in his essay, Roth calls Hiroshima mon amour "a film that remembers f
getting" (101). As Riva tells the tale of an old lover to her new lover, she begins
forget. Narrating memory, she leaves out details and she alters the experience by fa
ioning it into a narrative. Her ambivalence vis-a-vis narrating the memory of her fi
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
Cinematic Free Indirect Style in Hiroshima mon amour 381
lover bookends the film. In the opening, prefatory sequence she admits her forgetting
to Okada, "Like you, I too have tried with all my might not to forget. Like you, I forgot.
Like you, I wanted to have an inconsolable memory, a memory of shadows and
stone" (Duras 23). Toward the film's end, Riva chastises herself for having spoken her
memory aloud. Alone in her hotel room after telling Okada her story, she plunges her
face into a sink full of water and stands in front of the bathroom mirror. She looks into
the mirror, her face and hair dripping, and she says aloud, "In Nevers she had a German
love when she was young
She never went to Bavaria. I dare those who have never gone t
her of love" (73). The first line, enunciated like a child reciting
Nevers un amour de jeunesse allemand" (110)- is delivered in a
a tone close to boredom. Hers is an unexceptional story without
nificance. The remaining lines, however, the final line in partic
greater intensity. While she continues to look at herself in the hot
hears Riva in voice-over: "You were not quite dead. I told our sto
to you tonight with this stranger. I told our story. It was, you s
be told. For fourteen years I hadn't found ... the taste of an im
Since Nevers. Look how I'm forgetting you. . . . Look how I've
Riva tells the story that she has until now guarded. It is a story
"racontable" (110). Her hesitation compels her to provide only
outline of her love affair, and the narration of this scene, with its
reflects her hesitation and envisions her ambivalence. The film itse
a parallel ambivalence. The film narrates her memory, but as it
that narration. The film wonders whether its task is possible; it
ema can narrate, can represent or "remember" memory. During m
rated memory, the film imposes an external eye over Riva's me
film asks: which agency here "sees" these memories? It also asks:
memory, "sees" history, more clearly, more accurately?
The instances of cinematic FIS examined here, instances of fi
tion that carry strong markers of a third-person vision, draw s
discursive uncertainty of prose FIS. FIS is a common narrative m
Its practitioners are diverse and its usages and effects wide rang
likely a common narrative mode in film. Hiroshima mon amour
of a film that makes use of FIS to represent memory or consc
thematic focus on an epistemology of vision makes its usage of
ral form of narration, particularly fruitful. The cinematic mediu
tracks, its use of words, images, sounds, and editing, presents th
other, many additional varieties of "shared" cinematic narration,
ENDNOTES
A version of this essay was presented at the 2010 International Society for the Study of Narrati
ference in Cleveland. I would like to thank Jerry Carlson, James Phelan, and the anonymous
for their very useful suggestions for revising this article.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms
382 Leah Anderst
1. Unless otherwise specified, all English quotations from Duras are from the Richard Seavers
translation, and all French quotations are from the Paris, Gallimard edition that transcribes the
original French of the film. The same is the case for Flaubert's Sentimental Education, which I cite
in both the original French and in the English translation by Robert Baldick.
WORKS CITED
Barthes, Roland. S/Z. Translated by Richard Miller. New York: Hill and Wang, 1974.
Cohen, Keith. "Pleasures of Voicing: Oral Intermittences in Two Films by Alain Resnais." L'Espr
Creature 30.2 (Summer 1990): 59-70.
Cohn, Dorrit. Transparent Minds: Narrative Modes for Presenting Consciousness in Fiction. Prince
Princeton Univ. Press, 1978.
Craig, Siobhan S. "Tu rias rien vu a Hiroshima: Desire, Spectatorship and the Vaporized Subject
Hiroshima mon amour!' Quarterly Review of Film and Video 22 (2005): 25-35.
Deleuze, Gilles. Cinema 1: the Movement-Image. Translated by Hugh Tomlinson and Barbara Ha
berjam. Minneapolis: Univ. of Minnesota Press, 2009.
Duras, Marguerite. Hiroshima mon amour. Translated by Richard Seaver. New York: Grove Pres
1961.
Flaubert, Gustave. Correspondence. Edited by Bernard Masson. Paris: Ga
Higgins, Lynn A. New Novel , New Wave, New Politics: Fiction and the Repr
Postwar France. Lincoln: Univ. of Nebraska Press, 1996.
Hiroshima mon amour. Directed by Alain Resnais. Screenplay by Margue
Criterion Collection, 2003. DVD.
Kawain, Bruce F. Mindscreen: Bergman, Godard and First-Person Film. Pr
Press, 1978.
The Last Laugh. Directed by F.W. Murnau. 1925. New York: Kino Video,
Mitry, Jean. The Aesthetics and Psychology of the Cinema. Edited by Benoit
Lewis. Bloomington: Indiana Univ. Press, 1997.
Moses, John W. "Vision Denied in Night and Fog and Hiroshima Mon Am
terly 15.3 (1987): 159-163.
Pascal, Roy. The Dual Voice-. Free Indirect Speech and its Functions in the Ni
Manchester: Manchester Univ. Press, 1977.
Pasolini, Pier Paolo. "The Cinema of Poetry." In Movies and Methods Volu
542-58. Berkeley: Univ. of California Press, 1976.
Roth, Michael S. " Hiroshima Mon Amour: You Must Remember This." In Re
and the Construction of a New Past, edited by Robert A. Rosenstone,
eton Univ. Press, 1994.
Wilson, George M. Narration in Light: Studies in Cinematic Point of View.
Univ. Press, 1986.
This content downloaded from
86.49.244.69 on Mon, 29 Mar 2021 16:14:15 UTC
All use subject to https://about.jstor.org/terms