Focalisation in Film Narrative

Focalisation in Film Narrative

CELESTINO DELEYTO

Deleyto's article is a pathbreaking attempt to adapt Genette's and Bal's concept of 'focalization' to the specific analysis of film narrative. Drawing on the distinction between narrator (who speaks) and focalizer (who sees), Deleyto rejects the traditional view that the 'camera narrates'. Following Bordwell ( 1985), he contests the position of 'invisible observer' theories according to which analysis of film narration should be exclusively based on the position of the camera. He accepts Bal's distinction between internal and external focalization, and concludes, disagreeing with Bordwell, that the external focalizer occupies the position of the camera. While Bal sharply separates focalization in the novel (which she ascribes to the story level) from narration (situated at text level), Deleyto argues that, in film, focalization has to be studied simultaneously with narration. Deleyto's most suggestive contention is that, whereas in the novel the two kinds of focalization alternate, in film several internal and external focalizers can appear simultaneously at different points inside or outside the frame, all contributing to the development of the narrative and to the creation of a permanent tension between subjectivity and objectivity. His analysis of classical film shows the general tendency of film narrative to combine internal with external focalization and to return to the objective presentation of external narration to make the internal gazes in the text understandable. Finally, Deleyto proposes an analysis of the relationship between internal focalizer and focalized through the use of four main techniques: editing, framing, movement of the camera and mise-en-scne.

Narratology is the study of narrative texts in general, not only novels. There are other ways of presenting a story, from the narrative poem

to the cartoon strip. Some of them do not use the written or spoken word as their only means of expression. Indeed, in some cases, spoken or written language is not used at all, as is the case in some paintings which clearly convey a narrative, or in certain silent films. It is through cinema, television and video, and not through novels that most stories are 'told' nowadays. For a narrative theory to be consistent and complete, it must work when applied to languages other than that of the novel. Most importantly in our culture, it must work when applied to the study of a film narrative. On the other hand, narratology has proved an efficient method of analysis when consistently applied to film texts. ¹ Mieke Bal ( 1985: 5) bases her analysis of a narrative text on a three-layer distinction. Two of these three layers, the fabula -- 'a series of logically and chronologically related events that are caused or experienced by actors' -- and the story -- 'a fabula that is presented in a certain manner' -- can be shared by several narrative texts expressed through different media. It is only at the last level, that of the text, that the analysis would vary depending on the language signs of which it is composed. The implication here is that the analysis of focalisation, which belongs to the story layer, could be applied, as it stands in her theory, to a film text, together with time manipulation, character and space. Both fabula and story are abstract instances, analytical constructs of the critic, which do not appear explicitly in the narrative text. Only the linguistic or visual signs, which form the text, are there to be seen. Focalisation is, therefore, like character or space, narrated in the text.

It is mainly in this distinction between story and text (which does not exist in Genette or in Chatman) that the application of Mieke Bal's theory to film narrative becomes most problematic. Is focalisation, as Bal defines it, not textual in a film? Is the focaliser not a narrating agent like the narrator? In more general terms, who is the narrator in a film text? In films like Rebecca ( 1940), Duel in the Sun ( 1946) or Double Indemnity ( 1944), there are clearly voice-over narrators, who may be character-bound ( Rebecca, Double Indemnity) or external ( Duel in the Sun). In the case of many silent films or some sound films, like Gone with the Wind ( 1939) or Arsenic and Old Lace ( 1944), the intertitles fulfil the function of the narrator. But the activity of the narrator in all of these examples does not appear throughout the text, but only intermittently, sometimes hardly at all. In all of the narrative theories discussed above, the existence of a narrative presupposes the existence of a narrator, who performs the activity of narration. No narrator, no narrative. Does this mean that films are only narratives in those cases in which the activity of narration can be clearly ascribed to a narrator's voice or printed words? What about those silent films in which there are no intertitles or sound films with no voice-over narrators or printed words? Are these not narrative films at all?

The answer to all these questions is that a film narrative does not need the existence of an explicit narrator, as this agent is defined by theories of the novel, for the activity of narration to take place. As Branigan ( 1984: 40) says: 'In film, the narrator is not necessarily a biological person, not even a somehow identifiable agent like in the novel, but a symbolic activity: the activity of narration'. This symbolic activity, which has been sometimes called narrative instance (instance narratrice'), does not solve the problem of how a film narrative works, since it seems to place the activity of narration outside the text, in a position which seems dangerously close, once again, to that of the implied author. The existence of some abstract or explicit instance, superior to the narrator, and which controls its activity, seems to be denied here, by equating the two figures. As if to corroborate this impression, Branigan redefines narration towards the end of his work: 'A set of frames within larger frames leading to a frame which cannot itself be framed within the boundary of the text -- an unavoidable and implicit omniscience which may now be called 'effaced'' ( 1984: 71). This is certainly far from Mieke Bal's concept of narrator and narration, and it does seem to associate the narrator (or that symbolic activity of narration which Branigan substitutes for it) with that persona of the real author which has been referred to with Wayne Booth's term 'implied author'. Notice how the adjectives 'implied' and 'effaced' have a very similar sense here. This 'narrator' does not utter any signs at all, visual or otherwise, at least not in the same sense as the narrator of words does, and cannot even be immediately identified with the explicit voice-over or onscreen narrator, when such an agent appears in the film.

There is another term which Branigan uses and which, in spite of the abstract, symbolic slant it is given in his theory, seems closer to the role of the narrator. That is the concept of camera, which he defines as a 'construct of the spectator' and a 'hypothesis about space' ( 1984: 54). The camera has often been used as the equivalent of the narrator in film, notably by such classical theorists as V. S. Pudovkin, Karel Reisz and Gavin Millar. ² The camera would define the position of that invisible observer that could be identified with the narrator, a more identifiable agent which would indeed seem to be the origin of narration in a film. We could even extend the concept of camera to account for editing devices. This invisible but identifiable narrator would be able not only to present, in various ways, the space contained in the frame, but would change from one shot to another when it is necessary for the development of the narrative. However, there is at least one textual code which could not be accounted for by any definition of camera. Apart from editing devices and the position of the camera, there is in a film text a group of elements which come under the general term of mise-en-scene, that is, the staging of the events in front of the camera. To say that these elements are not textual but correspond only to the level of the fabula would be the same as to say that a dramatic text does not exist. As far as the miseen-scene is concerned, a film narrative works in a similar way to a play. According to theorists of drama, there is no narration in a play but representation. ³ There is a story and, therefore, a fabula; but that story is not narrated but represented by means of actors and a dramatic space, with a certain relationship with the audience.A film is, therefore, at a textual level, a mixture of narration and representation. Narration, performed by a voice-over or onscreen explicit narrator or as a metaphoric activity whose origin is the camera, does not cover all the textual activities that appear in a film. The mise-en-scene of a film, a term which can only metaphorically be applied to the novel, falls outside narration. Because there is no way of including the mise-en-scene code within the concept of film narration, there is no point any more in identifying the camera with the narrator. Since we have to accept that some textual elements in a film are not produced by the narrator (however wide we make this concept to be), it is preferable to keep the term narrator for those cases in which there is an explicit narrator (voice-over, onscreen or intertitles), with a similar status to that of the narrator in prose fiction. For the rest, we cannot accept the phrase 'the camera narrates', since its activity is so clearly different from that of the narrator, even though it is obviously textual, in the sense that it produces visual Signs. ⁴ We are, therefore, still in need of a term which describes the textual activity that takes place in a film outside narration and representation. ⁵ The term camera is not satisfactory for three reasons:

I have rejected the phrase 'the camera narrates'. If we are to leave metaphors aside, what the camera does is look at something which it then defines in terms of what is inside the frame. [ . . . ] Words like the English 'gaze', the French 'regard' or the Spanish 'mirada' seem much closer to how a story is presented in a filmic text than 'narration'. Just as the activity of reading a novel implies a narrator at textual level, the spectator of a film, apart from reading and listening, looks and, therefore, his activity requires a textual agent that produces the signs he is looking at. This brings us back to focalisation.

It will have been noticed how dose this concept comes to what I have been trying to describe here. 'Focalisation', says Bal in a more elaborate account, 'is the relationship between the 'vision', the agent that sees and that which is seen' ( 1985: 104). Focalisation, a purely narratological term, with no profilmic connotations like camera, and more precise and unambiguous than perspective or point of view, covers precisely the textual area that had been left empty by the restriction of the role of the narrator. To return to the questions on pp. 218 -19, the answer to the first one, 'is focalisation textual in a film?', has just been given and is affirmative. Whether the 'focaliser narrates' or not is a question of terminology, but there is no reason to accept the sentence if we had previously rejected 'camera narrates'. If we admit that focalisation is textual, then there is nothing to prevent us from saying that the focaliser focalises in a film text. In answer to the last question, 'who is the narrator in a film text?', we would have to say that the role which is performed by the narrator in a novel is, in a film text, carried out by both the narrator and the focaliser. Finally we must remember that narration and focalisation are not the only activities that take place in a film text. The code of mise-en-scene includes elements which fall outside the realm of either of them. ⁷ We would say, therefore, that where there is narration in a novel, there is 'narration, focalisation and representation in film (and probably also the activity performed by nondiegetic music).

Focalisation and narration, therefore, exist at the same level, and simultaneously in film. It is not my aim to prove the inadequacy of Mieke Bal's theory for the analysis of film, but it remains to be seen whether the differentiation between story and text would still hold after my discussion. This basic difference between novel and film exposes, in any case, Bal's theory as a theory of the novel. The very example used by her to illustrate the importance of focalisation is from a visual text, the basrelief Arjuna's penance, in Southern India. The gaze of the mice and the supposed gaze of the cat in this text are not narrated but 'narrative'. They are elements of the text ( 1985: 102). It is no coincidence, however, that the example was chosen from the visual arts, because it is in them, with no mediation between the vision and its representation, where the crucial importance of focalisation can best be evaluated. This modification of narratological theory in its adaptation to film narratives is not without problems. Jost ( 1983) is aware of it when he distinguishes between focalisation and ocularisation. It seems to me, however that its use in film is unnecessary since both terms designate much the same thing. What Jost does realise is the fact that not all focalisation is textual in a film. When the narrator or the characters speak, at least, the contents of what they say is focalised, but much in the same way as in prose fiction. The perceptual selection operated in such cases is then narrated textually. This, however, does not mean that we are faced with two different activities but only that, in film, they can be textualised in, at least, two different ways, through narration or through focalisation. In a way, therefore, the story/text differentiation still holds, but, as far as our analysis is concerned, focalisation (and probably other aspects too, like character and space) must be studied simultaneously with narration.

Summing up what has been said so far, we are now in a position to point at two general differences between novel and film from a narrative point of view. In both genres, focalisation and narration are key concepts in the analysis of the narrative. In the novel focalisation is not explicit in the text, but must be elicited by the critic from the information given by the narrator. We read what the narrator says but only metaphorically do we perceive what the focaliser perceives. In film, focalisation may be explicit in the text, in general through external or internal 'gazes' and works simultaneously and independently from narration. Both activities, focalisation and narration, are textual. More specifically, the almost permanent existence of an external focaliser in a film narrative accounts for the general tendency of the medium towards narrative objectivity. Regardless of the various subjectivities that may appear in the text, the almost permanent external presence of the camera ensures a vantage point for the spectator, which continually tends to dissociate itself from and supersede that of the various characters involved in the action; in the novel, on the other hand, both narration and focalisation can be exclusively subjective and are so on many occasions. Even in the inappropriately called 'third-person narratives', internal focalisation, through the perception of one character, subjectivises the narrative in a way which has hardly been achieved in film.

In studies of the novel before Bal, subjectivity has usually been associated with narration. ⁸ In film, narration as I understand it here has generally been overlooked in studies of subjectivity, and the emphasis has been on visual aspects. ⁹ Mitry ( 1965), Kawin ( 1978) and Branigan ( 1984), among others, suggest different taxonomies of subjective images. [. . .] There are, however, problems in all of these categoxisations. Mitry only thinks in terms of the shot, although film theory has for a long time rejected the idea that the shot was the equivalent unit of the word in human language. He has, therefore, no use for eyeline matches, shots/reverse shots, etc. Kawin seems to mix different levels of generalisation: subjective camera seems much more specific, for example, than self-consciousness, which could include the other three. Branigan firmly bases his system on the position of the camera and largely ignores such important aspects as mise-en-scene, movements of the camera, etc. The very special relationships between objectivity and subjectivity in film -their simultaneity, the unmarked transitions from one to the other, and the distance between apparent and real subjectivity in the classical film -- fall outside the scope of any of these studies. The study of visual subjectivity in film can, on the other hand, be approached from a more abstract, less taxonomical standpoint; that is, once again, through the analysis of focalisation.

Focalisation in film can be external or internal. In a novel, external and internal focalisation can appear simultaneously in what Mieke Bal calls double focalisation ( 1985: 113-14); or it may not be clear whether a character or an external agent are focalisers (ambiguous focalisation). In either case, the origin of focalisation is the same (the external focaliser associating itself perceptually with the character, or not) and the focalised object identical. In film, on the other hand, there can appear, simultaneously, several focalisers, external and internal, on different points of the frame (or outside). It is through the study of the relationship between all these different agents, their possible positions on the frame, and the relationship between them, that the study of focalisation can contribute to the analysis of subjectivity in film.

At a textual level, the focaliser always occupies the position of the camera. ¹⁰ A movement of the camera or an editing device (cut, dissolve, fade, wipe, etc.) imply a change in the position of the focaliser. One striking exception would be the split screen, in which the external focaliser would occupy several positions at the same time. Such is the case in films like Napolon ( 1927), or, more recently, Twilight's Last Gleaming ( 1977). A superimposition would also imply, in theory, the same multiplicity of vantage points for the external focaliser, although, generally, superimpositions are used to express the mind of the character, and would therefore denote an internal focaliser. Superimposition and the balloon technique are used, for example, in The Crowd ( 1928) with this particular function. Similarly, gradual transitions from shot to shot, like wipes or dissolves, would momentarily imply this characteristic omnipresent power of the focaliser. There is a spectacular use of wipes and split screen in Dr Jekyll and Mr Hyde ( 1932). A characteristic description of New York through several dissolves happens again in The Crowd. In North by Northwest ( 1959), the omniscience provided by the dissolve is put to a more original narrative use in a quick double dissolve from the United Nations building to the Capitol in Washington and the CIA office, as is explained by James Monaco ( 1981: 190).

As a general rule, we could say that the external focaliser yields focalisation to a character less readily in film than in the novel. Subjectivity is often expressed in a film without the complete disappearance of the external focaliser as a distinct agent from the character whose vision or mind we are made to share. There are, however, several cases in which the external focaliser totally disappears. In the usually called point of view (POV) shot or subjective shot, we have a shot of a character looking offscreen and a cut to what s/he is looking at, from the exact position s/he is looking. Although there is external focalisation in the first shot -- the character looking is being focalised from the position of the camera -- after the cut, this same position becomes internal as it expresses the exact origin of the character's gaze.

The eyeline match is not absolutely necessary in order to express total identification between camera position and character gaze. Other clues to subjectivity may be given in the dialogue, voice-over narration, movement of the camera or, most characteristically, when another character looks straight at the camera (and straight at the spectator) as he addresses the character whose vision we are supposed to share. An interesting example of consistent use of subjective camera is La mort en direct ( 1979), in which a character has a device installed in his eyes in order to record visually everything that he sees. In this case, not only does the camera place itself in the position of his gaze, but his gaze imitates, within the world of the fabula, the profilmic activity of the camera. The best known examples of a sustained use of subjective camera are The Lady in the Lake ( 1946) and the first section of Dark Passage ( 1947). Unlike La mort en direct, in these two cases, we never get to see the character whose vision we are made to share. He is always focaliser and never focalised. The lack of effectiveness of these two films seems to imply that film, unlike the novel, needs the alternation or simultaneity of external and internal textual focalisation in order to express subjectivity efficiently.

External focalisation may also disappear in the case of dreams, fantasies and flashbacks. In the famous dream sequence from Spellbound ( 1945), it is the mind of the character that is visualising the contents of the frame. Focalisation is here only internal (and supported intermittently by internal narration). Flashbacks, a very common narrative aspect in classical film, are frequently supposed to express the memory of a character who would then become focaliser. In theory, external focalisation would disappear throughout flashback. This is actually not so in most films. In Rebecca, we have two flashbacks. The film starts with a dream which the character played by Joan Fontaine has the night previous to the film's present. Internal focalisation is adhered to during this first flashback, as we see Manderley, the house where most of the subsequent action takes place, through her 'mind's eye'. There are several cues to the strictly internal nature of textual focalisation in this scene: for example, the dream-like tracking shot and the connection with the narration (also internal) as when she crosses the gate or when she sees the house. The rest of the film is formed by the second flashback, in which we are shown the events that led to her arrival at Manderley and its eventual destruction by the fire. This section is cued as a memory of the character by voice-over narration at the beginning, but from then on, all textual signs in this direction disappear and they do not even reappear at the end of the film. Although internal focalisation falls more heavily on the main character, this is not the same internal focaliser as the one who is supposed to be telling us the story from the present, but the character from the past. Although there has been no explicit sign that signifies that what we are looking at is not her memory any more, the spectator forgets that he is supposed to be witnessing a memory of a character of the fabula, and interprets the focalisation that does not correspond to any of the characters of this second level of the fabula as external.

[. . . ] Even in films in which flashbacks have a much greater load of subjectivity, like Out of the Past ( 1947) or Double Indemnity, the tendency towards interpreting focalisation as external, when there are no recurrent explicit cues, is still strong. Again the mixture of external and internal focalisation seems to be a crucial fact in film narrative. A Letter to Three Wives ( 1948) is formed in its story structure by a present time and three flashbacks which occupy the central part of the film and consist of memories of the three female protagonists. The beginning of each flashback is marked by more or less complex transitions which cue the contents of what follows as subjective (although narration corresponds to a fourth agent, Addie Ross, whom we never see but whose presence is constantly felt). The first shots of the second and third flashbacks, so conspicuously marked as subjective, show two rooms (two kitchens): in both cases, the person to whom the memory corresponds enters the room a few seconds after. They are not in the room when the flashback starts and focalisation must therefore be external, although this contradicts, at a superficial level, the subjective character of the flashback. It is as if film narratives required a constant return to objective presentation for a better understanding of the internal gazes that occur in the text.

One reason for this is the difficulty that film has to present the mind of a character outside narration. In a novel, in a passage in internal focalisation, the mind of the character can be shown without a change in focalisation. The character can be focaliser and focalised at the same time, while keeping perceptual control over other characters or objects. In film, although, as we have seen, dreams, hallucinations, memories, etc., can be shown, it is problematic to express the characteristics of the vision while showing its object, hence the resource to shots in external focalisation, in which the focaliser becomes focalised and in which we can analyse better how what he perceives affects him. The completely subjective shot, on the other hand, because of its relatively rare occurrence, and because of the total disappearance in it of external focalisation, may become an effective tool to express some particular mental state in the character whose origin it is, whether it is curiosity, surprise, puzzlement, or a mind at work. This special state is sometimes intensified through what Branigan calls perception shot, which he defines as a shot that shows the heightened attention of a character ( 1984: 81). The typical example is the blurred shot that presents the vision of a drunkard. In The Stranger ( 1946), a character sees the image of her brother blurred before she faints. [. . . ] In Vertigo ( 1958), towards the end of the film, we get two subjective shots that express the vertigo that the main character suffers from as he ascends the staircase to a church tower, through a mixture of tracking and zoom shots. Branigan finds another perception shot in The Birds ( 1963). In this case, a woman sees a dead man, and the eyeline match shows two quick cuts into a dead man's face ( 1984: 81). This, according to Branigan, expresses the horror of the internal focaliser at the sight of her lifeless neighbour. I, however, would interpret the shot after the second cut as an intervention of an external focaliser that places itself closer to the focalised so that the spectator can 'enjoy' a better view of it. Here, once more, we come across the apparent smoothness with which internal focalisation is replaced or accompanied by the external agent in the expression of subjectivity. It is a clear case of Bal ambiguous focalisation, 'in which it is hard to decide who focalises' ( 1985: 114).

Subjectivity in film is, however, not restricted to instances in which a character focaliser seems to completely take over the external focaliser. Most times, the perception of one or several characters of the fabula is emphasised by the text while the external focaliser still keeps its enveloping position. In these cases in which two or several focalisers coexist, all the gazes present contribute to the development of the narrative. The degree of subjectivity of the scene win depend on our awareness of its internal gazes. Typical cases of external focalisation prevailing would be the long and extreme long shot. The influence on our reception of the impressions provided by the text of the gazes of each one of the characters included in such shots is usually minimal. However, there are no fixed rules and an internal gaze can be of crucial importance in a long shot. In The Man Who Knew Too Much ( 1956), an overhead long shot shows the character played by Doris Day watching a church from the street, whereas at the other side of the wall that separates the back entrance to the church from the street, the villains lead her son into a car and escape. The fact that she cannot see this action because of the wall in spite of her watchfulness, is the main narrative function of this shot. Film language is so flexible that any set of rules or classifications of textual elements is always risky and becomes invariably incomplete. The most we can do is observe and point out recurrent elements in classical texts, which are more highly codified. In this restricted sense, and without any attempt at covering the whole spectrum, there seem to be four textual codes that are frequently used to establish relevant internal focalisation, without making the external focaliser disappear. These are editing, movements of the camera, framing and mise-en-scene.

To start with editing, two of the most important techniques in continuity editing, the eyeline match and the shot/reverse shot, reinforce internal focalisation. ¹¹ Like the subjective shot, the eyeline match, one of the basic rules of classical film narration, relates two shots by means of the gaze of one or several characters. As in the former case, shot A focalises externally on a character looking offscreen. Shot B shows what the character is looking at, but, unlike the subjective shot, from a position which is not the one occupied by the character (the order could be reversed). The external focaliser remains present throughout and its position typically ensures for the spectator a better vantage point than those of any of the characters concerned. Sometimes it is difficult to decide whether an eyeline match is a proper subjective shot or not. However, for the difference to be relevant, some more marks of subjectivity are necessary, apart from the complete coincidence between character and camera in the origin of the vision. A very usual mark is camera movement, which, in the truly subjective shot, imitates the movement of the character. Such instances are common in Hitchcock's films, usually at moments in which the character discovers some relevant visual information. [. . . ]

The second editing element that emphasises subjectivity is the shot/ reverse shot, used frequently, although not exclusively, in dialogues. In Rear Window ( 1954) there is a long dialogue between Jeff and Lisa (played by James Stewart and Grace Kelly) filmed in shot/reverse shot, with the two characters sitting at opposite ends of a sofa in his apartment. The external focaliser shifts alternately from one end to the other of the imaginary line that joins both characters together, depending on where the higher point of interest in conversation lies. As one of the characters is focalised, the position of the other one is very close to that of the external focaliser. Our position with respect to the frame is, therefore, very close to that of the character, and internal focalisation is thus activated. Again, character perception becomes crucial in the narrative. In this example, the external focaliser stays close to Jeff for much longer periods than at the other end of the line and therefore, the narrative depends for its development more on his internal focalisation than on hers. This is only right in a film which plays with our position as viewers by having as its main protagonist a person who performs diegetically a very similar activity to ours. As the scene comes to an end, the shot/reverse shot strategy becomes a set of reciprocal eyeline matches between the two characters. In this brief set of shots only one character appears on the frame. The difference between the two sequences of shots in the scene emphasises the increasing separation that exists between the two characters, which is emphasised here by focalisation.

The shot/reverse shot directs our attention to one or another part of the film space, which is organised along the axis line, and makes, consequently, for some of the most striking changes in textual focalisation in film narrative. In Grapes of Wrath ( 1940), a character narrates in several flashbacks the takeover by the banks of the house and land where he and his family lived. In one shot, the external focaliser is placed next to the car where the representative of the bank sits, and the emphasis is here on the precarious conditions of the laborers who are focalised. Later on, as the tractors arrive to demolish their house, the tractor is seen from a position close to that of the labourers in what is a virtual shot/reverse shot structure, although some film action has intervened between the two. Here, the opposite vantage point selected by the focaliser underlines the growing awareness of the peasants with respect to the previous shot. Also, the effect of the tractor destroying the arable land is much intenser if we are placed in the approximate position of those who are being most affected by its action within the diegesis.

Although in the shot/reverse shot strategy a character might be directing his/her gaze to an object or character onscreen (this is the case in the example used from Rear Window), we have been dealing so far with units of two shots -- shot A that establishes the origin of internal focalisation and shot B that shows (internally or externally) the focalised. But in a film both internal focaliser and focalised may appear simultaneously onscreen, while the external focaliser occupies a specific position, different from theirs, in the film space. In these cases, the relationship between the camera and the filmic space, or, to be more precise, the composition of the frame is usually crucial for the establishment or intensification of subjectivity. As a general rule, the most active internal focalisation would correspond to the character that is framed closest to the camera. This is the case of the individual shots in the shot/reverse shot sequences studied above. It is also the case of shots in which foreground and background are clearly separated. The most usual device of this kind has the camera placed somewhere behind a character who is looking while the object of the gaze is also included in the frame. This is what Mitry calls a 'semi-subjective shot' or a way of 'subjectivising the objective' ( 1965: 75). ¹² Two elements are at least necessary for this kind of internal focalisation to be active: the framing and the gaze. In The Burmese Harp ( 1956) and in most of Orson Welles's films the constant appearance of characters in an extreme foreground position and the use of wide-angle lens, which keeps both foreground and background in focus, do not usually activate internal focalisation because their perception of the rest of the frame is not underlined. In these films, the object of this near identification of character and external focaliser is mainly used in order to create onscreen a much larger film space than in other films. We must remember that focalisation is not only a textual element. It originates at the level of the story, an analytical layer in which the very existence of a camera and a frame are immaterial. The camera and the frame are some of the vehicles through which perception may be textualised but not the only ones. The textual element that comes into play here is again mise-en-scene. In a film, perception can be intensified independently of the camera. A character's gaze may be crucial for our understanding of the narrative regardless of the position of the external focaliser. What matters sometimes is simply the relationship between the figures in the film space and the importance of the gaze of one or several of them. In The Thief of Bagdad ( 1924), the main character goes into a mosque, in which some people are at prayer, through a window in a wall opposite the position of the camera. His amazed perception of what is going on inside is one of the basic narrative elements of the scene and yet he is nowhere near the camera. In Chimes at Midnight ( 1966), Hal and Poins eavesdrop on Falstaff and Doll Tearsheet from the top of the bed in which the other two are lying. The important factor in these two examples is the fact that the three internal focalisers are looking (or listening) without being seen or heard. Focalisation, consequently, can originate in the fabula, become active in the story and the textualised by the mise-en-scene.

Summing up what has been said so far, the relationship between internal focaliser and focalised can be established through editing (eyeline match, shot/reverse shot, subjective shot), framing or miseen-scene. Another textual device which is often used is camera movement. We have seen how a camera movement, as a part of a subjective shot, can contribute to express internal focalisation. The classic example here would be the shot at the end of Vertigo, already referred to, when Scottie climbs the staircase to the church tower. There a mixture of an ascending tracking shot and a descending zoom shot transmits the impression of vertigo that the character is feeling and that has been so central to the development of the story. In this case, however, the tracking shot expresses the subjectivity of the character but keeps him offscreen. Sometimes, however, the movement of the camera can relate both object and subject onscreen, like the other devices that I have revised here. We would say, in general terms, that internal focalisation, as a relationship between a subject A and an object B in the diegetic space of the film, can be presented textually by cutting from A to B, by including both in the frame and, if necessary, emphasising A's gaze, or by tracking, panning and tilting from A to B. As a general rule, the eyeline match is preferred to camera movement in classical film because the latter calls the viewer's attention to the heterodiegetic position of the camera. The eyeline match is a much more unobtrusive choice (like all the other strategies of the system of continuity editing) and, therefore, much preferred by a system that finds the best way of manipulation through character subjectivity in giving the impression that the film narrates itself. Some modern films, however, are less concerned with transparence. [. . . ]

Camera movement can be combined with framing to present as subjective a shot which was apparently objective. In Jubal ( 1955), a man on horseback is following a woman through a wood without being seen. His gaze is intensified by framing as the camera is placed somewhere behind him and the focalised is left in the background. After two cuts, the focaliser becomes focalised when a quick tracking shot discloses the presence of another character, who had been following the first one, and had been left offscreen occupying a hypothetical position behind the external focaliser up to that moment. The framing after the tracking shot underlines the new subjectivity in a similar way to the former one.

Another possibility is illustrated by two similar shots from North by Northwest. At the beginning of the film, several characters are talking in the bar of a hotel. Roger Thornhill calls the bellboy to ask him for a telephone, and, simultaneously, the boy calls out the name George Kaplan. At this precise moment the camera tracks left and forward to disclose the presence of two strangers who had been standing just outside the lobby, overlooking the scene. The simultaneity of the two events leads them to believe that Roger Thornhill is George Kaplan, a misunderstanding which changes Thornhill's life and sets the narrative mechanism of the film in motion. The movement of the camera, with the unexpected presence of the two internal focalisers that it reveals, underlines the coincidental nature of the beginning of the protagonist's misfortunes. Later on, the auction scene starts with a close-up of Philip Vandamm's hand lying on the back of Eve Kendall's neck. Then the camera tracks back and laterally to what is apparently an establishing shot of the new film space (the auction room), but finally discloses Thornhill standing at the back of the room, looking towards the position occupied by Eve and Vandamm. In the first scene, the most relevant information was the mistaken perception of the two men. In the second one, the discovery that Thornhill makes of the relationship between the other two. In both cases, an apparently objective presentation is abruptly shown to be internally focalised by a movement of the camera from object to subject. In neither of them does framing contribute to the establishment of internal focailsation at the end of the tracking shot, as the two subjects in the first scene and the one in the second one are ostensibly focalised externally. In both, the need for the viewer's concentration on the relevant gaze asks for this dcalage between external and internal focalisation in the same shot. Mise-en-scene, therefore, replaces framing once again in the expression of subjectivity.

I have discussed here only a few elements that contribute to the textuahsation of internal focalisation in a film text. Many others could have been referred to: lighting, colour, camera distance, internal sound, etc. [. . . ]. In most cases, the different codes work simultaneously and inseparably, in such a way that the function of each particular one cannot be fully understood without the contribution of the others. I prefer, therefore, to keep them all in sight and observe how they work in each individual scene. Focalisation is, at any rate, an essential code in film narratives. As opposed to the novel, it is also textual and, therefore, explicit, working at the same level as narration and the rest of the codes. More specifically, it is the permanent tension between internal and external focalisation that has been revealed as most relevant for our study. This tension could, in general terms, be described as the tension between the cinema's natural tendency towards objectivity and the centrality of the gaze in film narration. The apparent inconsistency inherent in the eyeline match or the classical flashback is in fact an element of richness and complexity, which provides film texts with the unique possibility of combining simultaneous internal and external gazes in such a way that, most times, the coexistence of both is taken for granted by the viewer, thus constituting a permanent source for subtle fabula manipulation and irony.

Notes

References

ASTRUC ALEXANDRE ( 1948), ''Naissance d'une nouvelle avant-garde: la camra-stylo'', cran franais No. 144 ( 30 Mar. 1948).

BAL MIEKE ( 1985, 1980), Narratology. Introduction to the Theory of Narrative, trans. Christine van Boheemen ( Toronto: University of Toronto Press).

BORDWELL DAVID ( 1985), Narration in the Fiction Film ( Madison: University of Wisconsin Press).

BORDWELL DAVID and THOMSON KRISTIN ( 1990, 1979), Film Art. An Introduction ( New York: Alfred A. Knopf).

BRANIGAN EWARD R. ( 1984), Point of View in the Cinema. A Theory of Narration and Subjectivity in Classical Film ( Berlin: Mouton).

COHN DORRIT ( 1978), Transparent Minds ( Princeton: Princeton University Press).

GENETTE GRARD ( 1972), Figures III ( Paris: Editions du Seuil).

--( 1982), Nouveau discours du rcit ( Paris: Editions du Seuil).

JOST FRANOIS ( 1983), ''Narration(s): en dea et au del'', Communications 38.

KAWIN BRUCE F. ( 1978), Mindscreen: Bergman, Godard and First-Person Film ( Princeton: Princeton University Press).

KOZLOFF SARAH ( 1988), Invisible Storytellers. Voice-over Narration in American Fiction Film ( Berkeley: University of California Press).

MITRY JEAN ( 1965), Esthtique et psychologie du cinma, vol. II ( Paris: ditions Universitaires).

MONACO JAMES ( 1981), How to Read a Film ( New York: Oxford University Press).