Narration and Focalisation in Wings of Desire.
David Bordwell initiated a cognitive analysis of filmic comprehension in his book Narration in the Fiction Film (1985). Bordwell argues that the logical form of a narrative film is initially incomplete, but is enriched, or completed, by the spectator's activity of inference generation. In order to complete a narrative film's logical form, the spectator must possess "narrative competence", a mental schema (that is, implicit, intuitive knowledge) that guides each spectator's comprehension of narrative films. The film's plot consists of a series of "cues" (such as gaps in the narrative events, monocular movement parallax, etc.) that trigger and constrain the spectator's activity of inference generation. This activity enables the spectator to gradually build up a mental representation of the film's story world.
Branigan's book significantly enriches and advances Bordwell's theory. Branigan has developed a sophisticated model of the film spectator's narrative competence, but especially of filmic narration, which determines the way narrative information is conveyed to the spectator. One of the key features of Branigan's theory is his account of the way spectators infer levels and agents of narration in order to process and comprehend narrative information. He persuasively argues that different agents operate on different levels of narration. In the following pages I shall focus on Branigan's theory of the narrator, character, and focaliser, three different agents who convey different types of narrative information to the spectator. (I shall also refer to two other agents: the implied author, who conveys non-fictional information to the spectator, such as the names of the actors, technicians, and the studio; and to the invisible observer.) My main aim is to apply Branigan's cognitive theory of filmic narration to a fil m that makes explicit and then playfully renders ambiguous the boundaries between narrative agents and levels of narration--namely, Wim Wenders's Wings of Desire (1987). I shall also consider other examples in passing, particularly Knick Knack (John Lasseter, 1989) and Psycho (Hitchcock, 1960).
Edward Branigan on Narrative Agents
To study narrative is to find out what is happening in a film. To study narration is to find out how spectators acquire knowledge about what is happening in a film's narrative. Filmic narration therefore mediates between the narrative and the spectator; it governs how spectators acquire knowledge about a film's narrative. Narration determines the flow of narrative information, and one of the aims of a theory of filmic narration is to describe how information is distributed by particular modes of narration. The narrative agent is an essential component of narration through which narrative information is filtered. For Branigan, a theory of agents requires a tripartite distinction between narrators, characters, and focalisers. A character is an agent whom spectators comprehend as existing on the level of the narrative--that is, an agent who experiences narrative events directly and who acts or is acted upon in the narrative world. A character whose experiences of the narrative world are then conveyed to the spec tator (by the narrator) becomes a focaliser. Narrators, on the contrary, do not exist in the narrative; they exist outside it on the level of narration. This means that they have the ability to influence the shape and direction of the narrative.
One of the most important contributions Branigan makes to the study of filmic narration is his rigorous definition of focalisation in film:
Focalization (reflection) involves a character neither speaking (narrating, reporting, communicating) nor acting (focusing, focused by), but rather actually experiencing something through seeing or hearing it. Focalization also extends to more complex experiencing of objects: thinking, remembering, interpreting, wondering, fearing, believing, desiring, understanding, feeling guilt. 
Branigan then distinguishes two types of focalisation, each of which represents a different level of characters' experiences. External focalisation represents characters' awareness, but from outside characters (it is semi-subjective, in traditional terminology). Internal focalisation represents characters' private and subjective experiences--ranging from simple perception (optical vantage point) to deeper thoughts (dreams, hallucinations, memories).
The narrator is the third major agent in film. A character can become a narrator--or, at least, a homo-diegetic narrator, that is, a narrator portrayed in the narrative by a character. This type of character-narrator is prevalent in films such as Double Indemnity (1944) and Sunset Boulevard (1950), in which a character narrates the narrative events in the form of flashbacks. But whether a film contains a homo-diegetic narrator or not, it is still narrated by a higher level narrator, one who does not exist in the world of the narrative. This implicit narrator is an agent who exists on the level of narration--s/he is an omniscient "master of ceremonies" who does not see anything from a perspective within the narrative, but can only report what happens from a position outside. (In the following pages, the term "narrator" shall be used to refer exclusively to this implicit narrator.) Although the narrator is absent from the narrative, his/her presence is nonetheless felt on the level of narration. For example, fi lm techniques that cannot be directly attributed to (or motivated by) a character in a film attest to the existence of the narrator. These include: unmotivated camera movements (camera movements not motivated by the movements of a character or narrative objects), intertitles, foreshadowing effects, and so on. (Shot changes are usually motivated by character movement, character glances off-frame, or by off-screen--but still diegetic--sounds and voices.) Such forms of discourse are not produced by characters in the film, so they can only be attributed to a narrator. The work of the narrator is always present in the film--as is evident in the very discursive activity of the film; in other words, the very acts of seeing shots on screen and hearing speech and sound attest to the existence of a narrator. Yet the narrator's presence is concealed from the spectator most of the time, for the film's discursive activity is usually motivated by the characters. But when such activity cannot be motivated by the characters, then the presence of the narrator becomes apparent.
Finally, we need to briefly consider one other type of agent--the implied author. This term refers to an element of narrative competence the spectator applies to a film to comprehend (more specifically, identify a source for) non-fictional information, such as the film's title, credit sequence, non-diegetic music, explanatory titles, and so on. More generally, Branigan defines the "voice" of the implied author as "an implicit extra-fictional narration" that "defines the limits of what can be seen and heard by us in the film but without defining the conditions of its own existence; one, moreover, which is able to predict events and anticipate the moral of the story ... ."  Seymour Chatman  argues that the implied author is the silent source of all information in a film; he/she is the inventor who remains silent but who delegates the narrator to speak for him/her. This distinction between silent inventor and transmitter is crucial for Chatman because it enables the film analyst to account for ironic and f acetious texts--in other words, it enables the analyst to distinguish between the literal meaning intended by a narrator and the ironic meaning intended by the implied author. It should be evident that the competence required by the spectator to comprehend the implied author is "top-down"--in this instance, extremely contingent, and can only be applied with a degree of uncertainty (as we shall see ahead). This also applies to a lesser degree to the narrator, whose presence can be more strongly comprehended. Finally, the characters and focalisers are the agents most clearly comprehended by the spectator.
Branigan's distinction of agents (narrators/characters/focalisers) and types of focalisation (external/internal), enables us to identify the following four basic types of shot:
(1) non-focalised or "objective" shots, or shots directly motivated by a narrator;
(2) externally focalised shots, which represent a character's awareness of narrative events. Such shots are comprehended by the employment of the eyeline match: we see what the character sees, when s/he sees it, but not from his/her exact spatial position;
(3) internally focalised shots (surface), which represent a character's experience of narrative events, as in optical point-of-view shots: we see what the character sees, when s/he sees it and from his/her exact spatial position; and
(4) internally focalised shots (depth), which directly represent a character's complex experiences--dreams, memories, and so on.
As I pointed out above, Branigan argues that these various levels and agents of narration constitute spectators' narrative competence, which aids their comprehension of narrative films. In other words, the levels of narration and the agents of narrative information are hypotheses and inferences spectators impose on a film to comprehend it (motivate its shots, understand the relevance of narrative information, and so on): "Such concepts as 'narrator', 'character', and 'implied author' (and perhaps even 'camera') are then merely convenient labels used by the spectator in marking epistemological boundaries, or disparities, within an ensemble of knowledge; or rather, the labels become convenient in responding to narrative."  The cognitive approach proposes that comprehension of a narrative film is not fixed and determined by the film itself. The spectator is engaged in an ongoing activity of applying hypotheses and inferences to films.
We can go so far as to say that what exist on the movie screen are simply changing patterns of light and shade, which are perceived together with a series of sounds that somehow relate to these patterns on screen. In comprehending a film, the spectator transforms this raw data into a coherent mental representation (the film's fabula or diegesis). Some of the spectator's activity is automatic (bottom-up processing), but other aspects are more contingent; that is, dependent upon his/her specialised knowledge or competence of cinema (top-down processing). This emphasis on top-down processing suggests that the spectator generates non-demonstrative inferences--i.e., inferences which are defeasible (that is, flexible and open to revision). The spectator's construction of a mental representation of the film's diegesis/fabula is not, therefore, a simple, linear, automatic process, but is an ongoing activity that is continually modified, corrected, and improved.
Despite the dominance of top-down processes in comprehending narrative films, it is important to emphasise, as Branigan does, that both processes operate simultaneously in comprehending characters, objects, levels of narration, narrators, various types of shot, camera movements, and so on from the patterns of light and shade on the screen. This raises the questions: How do we infer a level of narration or comprehend a camera movement? And which process dominates the comprehension of these particular "objects"? The film presents spectators with stimuli, or a series of traces, which require comprehension. A series of traces we need to consider, especially in relation to Wings of Desire, are those spectators label "camera movement".
To demonstrate that camera movements are not signified entirely by the text itself (are not wholly bottom-up stimuli), we shall consider an image from one of John Lasseter's short computer-generated 3-D animations. In Knick Knack (USA 1989) spectators comprehend one of the opening shots as a camera movement--the "camera" "tracks" from a collection of souvenirs from sunny countries to a snowman encased in a snowy landscape in a glass bubble. But this shot was not produced by a tracking camera; indeed, it was not even produced by a camera, for it was entirely generated on a computer. (Conventional cartoons, of course, are filmed using a camera, but the camera is usually fixed.) The computer has been programmed to manipulate the background and produce a series of traces on screen that the spectator comprehends as "camera tracks to the right", even though no camera was involved in the production of these traces. David Bordwell makes this point (in relation to conventional cartoons) in a 1977 essay "Camera Movemen t and Cinematic Space."  He argues that we cannot adequately develop a theory of camera movement by relying on the notion of a profilmic event filmed by an actual camera. He suggests we must divorce the cues, or traces on film from the camera and the profilmic events. In other words, it is not sufficient to theorise the production of these traces--a camera physically moving through space; rather the film analyst must instead concentrate on the reception of these traces--how the spectator comprehends them as a moving camera.
For Bordwell, a number of perceptual cues in a shot can be comprehended as subjective movement. One such cue is the monocular movement parallax, which applies to both a perceiving subject moving through space, and a stationary subject perceiving a moving object. The concept of monocular movement parallax states that a moving subject will perceive objects gliding past his/her field of vision; the closer the objects, the faster they glide by, due to a differential angular velocity between objects close to the perceiving subject and objects far away from the perceiving subject. In other words, a moving subject perceives a continual displacement of his/her visual field. By contrast, a stationary subject perceiving a moving object will perceive a partial displacement of his/her visual field.
Bordwell then applies the concept of monocular movement parallax to the cinema: "For the camera movement effect to occur, monocular movement parallax must be read from the entire visual field. If only a part or item in the field yields that differential angular velocity across time, then camera movement will not be specified--only the movement of the object will be specified."  Cues comprehended as camera movement "install the viewer as a subject moving through a fictive scenographic layout." 
Comprehension of camera movement is predominately determined by cues in the film--in other words, is predominately comprehended in bottom-up fashion, whereas other aspects of a shot (if it is internally or externally focalised, for example) rely much more strongly upon top-down competence. All top-down comprehension is defeasible and thereby open to revision, for it is not as deterministic as bottom-up comprehension. To comprehend a particular shot as focalised is to impose upon it an anthropomorphic reading in which the "camera" imitates a character's experience of the narrative world. Bordwell points out that American and Canadian avant-garde films challenge an anthropomorphic reading of shots, and cites Michael Snow's La Region Centrale. To what extent do the camera movements in Wings of Desire challenge an anthropomorphic reading of shots? And how easily does the spectator comprehend one of its dominant oppositions--the distinction between black & white/colour as a convention for coding focalisation (resp ectively, focalisation around angels and children/focalisation around adults)?
Branigan's theory of levels and agents of narration (and Bordwell's theory of camera movement) can be applied to any narrative film, whether it strongly adheres to these levels and agents, or creatively manipulates and transforms them. In Wings of Desire, the boundary between character and narrator becomes ambiguous, whereas in "classical" films such as Hitchcock's Psycho, for example, the boundaries are clear and distinct (comparatively speaking, for, as I pointed out above, spectators infer narrative agents, and such inferences are defeasible). First I shall analyse the opening of Psycho, emphasising the gradual transition from the dominance of the implied author and narrator to the dominance of characters, before moving on to analyse Wings of Desire.
Focalisation in Psycho
As the credit sequence of Psycho unwinds, the implied author's presence is dominant (which is apparent in all films, of course). The non-diegetic music is also a mark of the implied author's presence. The first three shots of the film are long shots, panning right, across a city sky line. The camera's movement and spatial location are not motivated by a character--these shots do not represent the point-of-view of a character looking across the city sky line. Furthermore, what city is this? The name of the city is superimposed over two shots (the name appears just before a dissolve from one shot to another, masking the transition). We are informed that this is Phoenix, Arizona. But who informs us? Here, a measure of uncertainty enters the analysis. Or, more positively, we can say that the titles can be described in more than one way. It is arguably the implied author, informing the spectator of the non-fictional name of the city. But at the same time, we can argue that the name of the city has been appropriated and absorbed by the narrator into a fictional world. Then the day, date, and the time are imposed over the shots (again, just before shot transitions)--it is Friday December the eleventh at two forty-three p.m. This is information relevant to the fictional world, and is conveyed to the spectator by the narrator.
Furthermore, both pieces of narrative information could have been conveyed to us by the characters in the narrative. For example, Sam could have told Marion the time, or one of them could have been reading a local newspaper, indicating the day and the fact that they are in Phoenix. This mode of conveying narrative information is conventional and unexceptional, and has the effect of making the narrator less noticeable. But in Psycho, the implied author and/or narrator directly supply us with the name of the city, the date and the time. Their presence at this moment in the film is quite dominant. Their presence is also indicated in the following two shots. Being omniscient, they know exactly where two of the film's main characters, Sam and Marion, are located; in shot 4 they direct the camera to their hotel room window and in shot 5 allow it to transcend space by entering the room. In the hotel room, the camera movement gradually attaches itself to the movement of the characters, making the presence of the impl ied author and narrator less overt. After Sam and Marion have dressed, we are presented with two shots of Sam (shots 13 and 15), looking out of the window, from Marion's approximate point-of-view. In other words, the spectator comprehends these shots of Sam as being externally focalised around Marion's vision. Sam is seen returning the look; he does not look directly into the camera, but off-screen left, indicating that the shot of him is not internally focalised around Marion. These shots of Sam are alternated with two shots of Marion (14 and 16). But in shot 16, Marion turns around (turns her back on Sam). Shot 17 returns the spectator to Sam standing by the window, framed in the same way as shots 13 and 15. Whereas shots 13 and 15 are externally focalised around Marion, the spectator implicitly comprehends shot 17 as non-focalised (as an "objective shot"). In other words, the shot of Sam is repeated for a third time, but is comprehended differently. As shot 17 progresses, Sam walks towards the middle of th e hotel room where Marion is now located--she has turned around again, although the spectator did not see her perform this action. Even though shot 17 is initially composed in the same way as shots 13 and 15, it is comprehended as non-focalised because Marion has turned her back on Sam in shot 16 (her look does not motivate the cut back to Sam). This comprehension of shot 17 as non-focalised is then strengthened as Sam walks into the middle of the room and faces Marion in a two-shot.
Branigan's definition of focalisation depends on a rigorous separation between characters and narrators. For Branigan, only characters can act as focalisers because they are able to experience directly the narrative events, whereas narrators cannot, although it is the narrator who conveys the experiences of the focalisers to the spectators. Focalisers only experience; they do not convey their experiences to spectators. It is the task of agents higher up to convey these experiences to spectators.
Focalisation in Wings of Desire
Branigan has accurately defined the two main levels of a narrative film (narrative/narration) and the main agents who inhabit these levels (characters and focalisers/narrators respectively). These levels are rigidly adhered to in "classical" narrative films, but in Wings of Desire they are rendered ambiguous. I would argue that the uniqueness of Wings of Desire is that the main protagonists--the angels Damiel and Cassiel--possess qualities of characters and narrators at the same time, and inhabit both the level of narrative and of narration, or, more accurately, exist between the two levels.
If the angels are not characters or narrators, then perhaps they conform to the conditions of the invisible observer, a familiar agent in traditional film theory, but which Branigan re-defines as part of the spectator's narrative competence; that is, a necessary agent the film spectator employs to motivate and comprehend a shot or sequence of shots.
What is being analyzed is not simply a technique, or a device or a style, but the elements that define a mode of comprehension. As a convention of reading, invisible observation asks the spectator to accept (or impose) a restriction on the amount and type of knowledge available from the text at a given moment. A convenient way to describe this particular restriction on knowledge is to imagine an "observer" with specific qualities. 
We shall consider in detail Branigan's discussion of the invisible observer to see if it constitutes part of the knowledge necessary for a spectator to motivate and comprehend the shots in Wings of Desire focalised around the angels.
Branigan identifies six characteristics of the invisible observer: 
(1) Invisibility. The observer has no causal interaction with the events which are witnessed; s/he is an eavesdropper who is unheard as well as unseen. Although, in Wings of Desire, the angels do not interact with narrative events (a source of displeasure for the angel Damiel), they are seen by children and, of course, by the spectator. They are invisible only to adults (although their presence can be felt by ex-angels such as Peter Falk).
(2) Ubiquity. The observer has the ability to move instantaneously through scenic space and to remember what has been seen and heard, but cannot move forward or backward in time. The angels move instantaneously through scenic space, which becomes noticeable in Wings of Desire because the invisible observer can actually be seen by the spectator (although not by characters). For example, early on in the film, Damiel effortlessly walks through a number of apartment buildings, observing their occupants. He then looks out of the window and sees an ambulance travelling along the auto-bahn. Outside the apartment buildings, and high up, the camera begins to follow the ambulance. Conventionally, spectators would comprehend this shot as non-focalised, but here it is focalised around Damiel. We then cut to inside the ambulance, and see a pregnant woman. The camera moves towards her stomach, and Damiel's hand is suddenly shown entering the filmic space from off-screen space shown to be empty moments before. Furthermore, the angels remember what has been seen and heard because they make notes. In two other sequences, Cassiel transcends time--a sequence in which he travels along the streets at right is filmed in fast action, and, travelling in the back seat of a period car, he looks outside the window (eyeline match) and sees the bombed Berlin of 1945 (documentary footage). The car then arrives on a film set which is full of Nazi soldiers, momentarily confusing the spectator as to the exact date of the narrative event. There is no evidence that the angels can move forward in time.
(3) Alertness. The observer is attentive to events so as to be able to avoid obstructions and assume the best or perfect angle on the actions at all times. This characteristic applies to the angels, since they are not bound by the physical constraints of the diegesis (hence giving them the characteristic of being narrators).
(4) Neutrality. The observer assumes a "standard", often straight-on angle to the action. The angels have the ability to observe narrative events both from human eye-level (hence imitating the invisible observer found in other films) and also from high up. Some camera movements can be comprehended as the movement of an angel moving from a high position to eye level position. In one scene Damiel sits at the top of a circus tent and watches Marion on the trapeze. Later in the scene Damiel moves from the top of the tent to ground level. The shot of him descending is initially internally focalised (we experience his optical point-of-view as he descends from the top of the tent to ground level). When the camera reaches ground level, we then see Damiel enter the shot. Most spectators initially comprehend the shot as Damiel's optical point-of-view (although not automatically; some spectators, finding this shot unusual, cannot comprehend it as the point-of-view of a descending angel). That the same shot is then compr ehended as a non-focalised shot of Damiel is quite common in Wings of Desire, and indeed in film in general, as both Bordwell and Branigan have emphasised. Bordwell writes that: "It is a permissible play with convention to have a character enter a shot which has been initially established as her or his point of view." For Branigan, "the presence of the POV shot as a narrative structure cannot be determined mechanically by measuring such material divisions as shots. Even in a wider context, the relationship between narration and the editing of shots is not fixed but must be discovered."
(5) Impersonality. The unique personality traits as well as the gender, age, race, class etc. of the observer are muted--made indefinite--so that what is displayed is strictly a perceptual experience and, in principle, inter-subjective, i.e., regulated by general norms of seeing. This point certainly does not apply to the angels, precisely because they are symbols of Christianity and because they are depicted in the diegesis. But the perceptual experience displayed is inter-subjective in that it represents the experience of all angels (and children). Spectators automatically share this perception for the simple reason that we can see the angels and because we see the narrative events in black & white (the shift from black & white to colour will be discussed below).
(6) Passivity. Observation is governed by the limits of immediate space and time. This last point assumes that the invisible observer exists within the film's diegesis, is constrained by its physical space-time boundaries, a condition which does not apply to the angels, even though they are visible in the diegesis. The angels can also read the minds of characters (conveyed to the spectator as voice-over) and see their memories of Berlin in World War II conveyed to the spectator in the form of documentary footage).
In Wings of Desire, the angels, together with a number of unusual camera movements they motivate, challenge the spectator's mode of narrative film comprehension, because the amount and type of narrative information they impart to the spectator is radically different from the information imparted to the spectator from other agents depicted in the diegesis (that is, characters). When we first see the angels, we expect them to behave like characters (especially when children notice them in the opening sequence), but we soon realise they are invisible to adults and have characteristics spectators usually attribute to invisible narrators. The restrictions spectators normally impose on knowledge conveyed by characters must be revised, for the spatio-temporal restrictions we attribute to characters are not applicable to Damiel and Cassiel. The angels do not conform to the boundaries each spectator imposes upon the film to comprehend the status of the information. The film establishes a number of unique intrinsic nor ms the spectator learns only after seeing the film a number of times, for the spectator versed in classical narrative film does not possess the necessary knowledge to make sense of these intrinsic norms on first viewing.
That children can see the angels, whereas adult characters cannot, again raises the issue of the ambiguous status of the angels, for they must, to some extent, be diegetic. But the angels, like narrators but unlike characters (potential focalisers) do not participate in the narrative world. They merely observe, look, and narrate. They cannot experience the colour, weight, texture, smell, and spatio-temporal dimensions of the here and now. This produces conflict for Damiel, who has fallen in love with Marion, the trapeze artist, and wants to experience that love with her. On a narratological level, Wings of Desire is about Damiel wanting to lose his non-diegetic narrator status and become a character.
Focalisation is explicitly codified in Wings of Desire by the opposition black & white and colour sequences (another of the film's intrinsic norms). Black & white signifies focalisation around an angel (or occasionally around a child), for it represents his/her detached awareness/experience of the narrative world (an experience devoid of most sensory data, as we have just seen). Colour signifies focalisation around potential/actual human characters (implied diegetic focalisers, or actual focalisers--that is, characters on screen). The shift from black & white to colour marks the distinction between detachment from the narrative world (by the observing angels, who act as potential narrators) and full experience of the narrative world (by characters who act as potential focalisers).
The first use of colour occurs in the scene in the tent discussed above. The scene consists of Marion on the trapeze swinging back and forth, Damiel at the top of the tent, and the ring master on the ground. We first see Damiel in long shot looking off-screen, moving his head from left to right. Cut to Marion on the trapeze, swinging from left to right, with the camera following her movement. The camera's placement and movement are therefore motivated by and are internally focalised around Damiel. But within the context of Branigan's theory, this shot should be comprehended as non-focalised, because it does not represent the experience of a character. However, the shots of Marion are comprehended as focalised because spectators actually see the non-character agent (Dansiel). We then cut back to Damiel, this time in medium close-up, as he follows Marion's movement. In the following shot we cut back to Marion (hence, maintaining a shot/reverse shot pattern). However, the following shot does not cut back to Dami el, but instead shows Marion, in a colour shot, from an eye level point-of-view. The camera looks up at Marion, and the space occupied by Damiel is shown in the background. Yet Damiel is absent from the space, strengthening the assumption that the black & white shots of Marion are not focalised around a character and indicating that the colour shot represents the visual experience of an adult (rather than an angel or a child). Damiel still occupies this space, but the spectator he can no longer see him. We then have an extraordinary analytic cut-in, as the camera shows Damiel, in a black & white shot, in medium close-up again. The shot depicts a detail of the previous space, but in black & white rather than colour, and with Damiel now shown occupying that space, who was shown to be absent in the previous (colour) shot.
This phenomenon of overlapping space--first filmed in black & white, then in colour--is another intrinsic norm of Wings of Desire, and is employed a number of times in the film, particularly towards the end, when Damiel is transformed into a character and talks to Cassiel. The shift from black & white to colour is not determined by the angels, but a more conventional (and higher level) narrator (or possibly implied author) existing outside the fictional world, who decides whether to portray that world in black & white or in colour. Furthermore, this shift signifies a radical change in point-of-view, from narrator-angels to characters (the colour shot is most plausibly focalised around the ring master). I suggest that the reason for the shift occuring at this particular time is to introduce this unique intrinsic norm to the spectator, so that s/he will be better prepared to comprehend the second time the shift from black & white to colour occurs--in Marion's caravan.
In the caravan Damiel observes Marion's belongings and reads her thoughts. He turns his back on Marion and looks at a number of photographs and "picks up" a stone. When he turns around, he sees Marion sitting on the edge of her bed, beginning to undress (filmed as a two-shot). We cut to a shot of Damiel watching, then to a close-up of Marion's back and left shoulder as Damiel's hand enters the frame and runs along her shoulder, as Marion's thoughts continue to be heard on the sound track ("I only need to be ready, and every man in the world will look at me"). We then cut to a shot of Damiel; the camera remains still as he walks out of the frame (the camera therefore disengages itself from him). We then cut to a long shot of Marion sitting on the bed. The shot slowly changes from black & white to colour as Marion puts on her dressing gown and walks to the other end of the caravan. This transition is extremely rare in narrative film, and so its comprehension involves contingent top-down processes that most spec tators may not possess. For me, this colour transition signifies the disappearance of Damiel from Marion's caravan (he's a polite angel!), leaving the shot to be comprehended as objective. 
This discussion of Wings of Desire does not pretend to be exhaustive, but merely begins to highlight those moments in the film that challenge the spectator's activity of comprehension. For example, it raises questions such as how the spectator comprehends agents, camera movements, and changes from black & white to colour in narrative films. Wings of Desire therefore draws attention to spectators' narrative competence (their implicit, intuitive knowledge) in comprehending narrative films, and tests the boundaries of that competence.
Warren Buckland is Associate Professor in Film Studies at Chapman University, Orange County, California. He is author of The Cognitive Semiotics of Film (2000), Teach Yourself Film Studies (1998), and editor of The Film Spectator (1995). He has recently co-authored (with Thomas Elsaesser) Studying Contemporary American Films: A Guide to Movie Analysis (2001). He is currently writing a book on style and narration in Steven Spielberg's blockbusters.
(1.) Edward Branigan, Narrative Comprehension and Film (New York and London: Routledge: 1992), 101.
(2.) Branigan, 90-91.
(3.) Seymour Chatman, Coming to Terms: The Rhetoric of Narrative in Fiction and Film (Ithaca and London: Cornell University Press, 1990), especially Chapters 5 and 6.
(4.) Branigan, 85.
(5.) David Bordwell, "Camera Movement and Cinematic Space", Cine-Tracts, 1, 2 (1977), 19-26.
(6.) Bordwell, 22.
(7.) Bordwell, 23.
(8.) Branigan, 172.
(9.) Branigan, 171-72. The following six points are shortened summaries of Branigan's text, although I have added the information concerning their relation to Wings of Desire.
(10.) Bordwell, 25.
(11.) Branigan, 142.
(12.) After reading this passage, Edward Branigan offered me a far more accurate description of this transition. He writes that: "This transition from b & w to color seems to be the transition from Damiel's POV (i.e., internal focalization of an explicit fictional, nondiegetic narrator) to implicit fictional, diegetic narration (i.e., an invisible, objective diegetic observer" (personal communication).
|Printer friendly Cite/link Email Feedback|
|Date:||Jun 22, 2001|
|Previous Article:||DVD the Shift to Film's New Modernity.|
|Next Article:||Narrative and Spectacle in Gladiator.|