Continuum: The Australian Journal of Media & Culture
Vol. 1, No. 1, 1987
Australian Film in the 1950s
Edited by Tom O'Regan


Horst Ruthrof

Review: Point of View in the Cinema, by Edward Branigan (Amsterdam, New York Mouton, 1984).


Edward Branigan's Point of View in the Cinema [1] joins a long list of works on narration and narrative, the more well-known of which are Lubbock's The Craft of Fiction, Stanzel's Narrative Situations in the Novel, Norman Friedman's 'Point of View in Fiction', Booth's The Rhetoric of Fiction, Chatman's Story and Discourse, and Uspensky's A Poetics of Composition. [2] Point of View in the Cinema impresses for a variety of reasons. It rests on a broad, sound assumption of narrative being a heterogeneous arrangement of signs which transcends any specific semiotic system, for the description of which the analogue of linguistic signification is useful, but which should be seen neither as the only critical tool nor even as the most privileged one. Narrative is understood as a dynamic structure recognisable independently from its material of expression and requiring for its analysis specific attention to the medium in which it manifests itself. In Branigan's book this medium is 'film' in the sense of 'cinema', not merely a text but also a cultural institution.

The book addresses itself to a large number of detailed problems of film theory and in so doing refers the reader to a rich body of film texts. At the same time the author manages to distinguish between levels of generality and specificity, no mean accomplishment given the complexity of the polysemiotic structures he has chosen to analyse. A particular strength, it seems to me, is the book's foregrounding of the active work of the film viewer. This has as a corollary the diminished responsibility of the author/auteur's construct, filmic images as such, the camera and other profilmic tools (though their importance as cinematic facts is by no means underrated).

In spite of the book's insistence on the generation of filmic meaning in the act of viewing, Branigan avoids a wholly relativist position. The reason for this balance lies in Branigan's view of the systemic nature of all cultural activity and its codes. These are argued to bind individual readers to a semiotic community and provide the relative freedom to generate an infinite number of filmic meanings from a limited set of basic rules. The book also draws our attention to a list of principles which distinguish our construction of narrative in classical film from that in modern film texts, along the lines of Roland Barthes' distinction between the 'readerly' and the 'writerly' composition. Throughout the book Branigan analyses film viewing from the perspective of the five units of classical representation: origin, vision, time, frame, and mind. The author uses these to theorise, among others, the concepts of point of view, subjectivity, narration and narrative, character and representation.

Compared with this general verdict, a recent review of Branigan's study by Seymour Chatman is not a friendly act, to say the least. [3] The book is blamed for its clumsy handling of theory, poor style, a confusion of representation and viewing, an overemphasis on the reader, poor editing, and its exorbitant price. Perhaps Chatman's attack is not so surprising when one recalls the relatively insignificant position which the reader/viewer is granted in his own schema in Story and Discourse. Chatman, it would appear, has misunderstood the fact that from the position of cinema, that is film viewing, all constructions of narrative meaning are primarily acts of reading. My own response to Branigan's work in this article is a detailed and, I hope, fair summary of his argument, followed by a brief, critical commentary.

Detail and Argument

There are eight chapters: "The Problem of Point of View", "Film as System", "Narration", "Subjectivity", "The Point-of-view Shot", "Character Reflection and Projection", "The Modern Text: Subjectivity under Siege from Fellini's 8 1/2 to Oshima's The Story of a Man Who Left His Will on Film", and "Metatheory". The book closes with a substantive, critical "Appendix: Orthodox Theories of Narration", a list of Suggested Further reading and an Index.

Branigan opens his argument by distinguishing between four subjects: author, narrator, character, and viewer. Together they constitute subjectivity or "the process of knowing a story - telling it and perceiving it", whereas 'subject matter' is "the result of a process of telling"(p. 1). Part of the recurrent terminology of Branigan's approach are such present participles as 'telling, presenting, portraying, representing, giving, framing, narrating' (p. 1f.) and a little later 'watching, listening', and 'displaying' (p. 2). This list which could have come straight from Husserl's Ideas where it is defined as a series of noetic acts. In fact, the central distinction between narration and narrative which the author sees as a specific form of the more general pair of enonciation and enonce needs to be related not only to the linguistic distinction drawn by Benveniste between discours and histoire (as Branigan does), but also both to the distinction in logic between modality and proposition and to Husserl's much broader distinction between the noetic and noematic, of which all those pairs constitute special cases. Though Branigan ensconces the connection between his own approach and phenomenology, the reader soon realises that his preference is for the performance of interpretative acts over the mere empirical registration of materials: any narrative is always the product of some narrational act rather than vice versa.

Because of the prominent role which narrational activities play in Branigan's book (the author avoids the term 'acts'), he needs to return again and again to the problem of subjectivity. He does so, successfully on the whole, to make it easier for the reader to grasp the intricacies of filmic processes; but there is a certain loss of conceptual elegance when the author stretches the concept from covering 'character narration or perception' to including 'the perceptual context of every utterance within the text whether explicit or implicit' (p.2). To be sure, such an argument could be mounted on the basis of a theory of meaning which postulates the necessity of utterance acts for the existence of sign systems. Yet such a theory is not pursued.

One of the more important results of Branigan's axiomatic dynamics of narration/narrative is this observation. Since subject and object, the act of viewing and what is viewed, are in principle interchangeable, a film text can be seen as an 'hierarchical series of pairs of (nominal) subjects and objects in which a subject/object pair may at any time become an object for a higher-level subject' (p.2). This relation operates like a dynamic 'set of Chinese boxes, one inside the other, with each successive box or level introducing a new relation of subject and object' (p.2). Accordingly, Branigan quite rightly argues far subjectivity, in a functional rather than a humanistic sense, being part of all narration and present particularly in those shots which are commonly regarded as 'neutral'.

The relationship between narration and narrative is further classified by Branigan's analogy with the 'I am lying' paradox which, as we know from Goedel's explanation about logical systems, requires a metasystem for the dissolution of its contradiction. What is responsible for the paradox is the double subject concealed in the 1', one which operates at the level of what is said, the other at the level of its speech act. By splitting the subject into its two logical components we arrive at two levels of discourse, a metalevel for 'I am saying that' and a primary level for I am lying'. One could go on constructing a chain of such pairs ad infinitum. Likewise, the interaction between narration and narrative, a special case of the logical distinction, turns out never to be symmetrical, but asymmetrical in the sense that for each narrative we must always stipulate, a higher-degree, enabling narration. Hence the doubling of 'narration' in the title of this article.

While the multiplicity of subject positions or acts of narrating in language texts is indexed according to Branigan by linguistic shifters, it is the 'spatial point from which an image is viewed' which achieves the same for us in 'pictorial language' (p.4). But these positions are not absolute, fixed by the producer or the camera; rather, 'film is a discourse which itself creates a set of subject positions for the viewer, just as the viewer is able to frame and reframe the film and create subject positions for the presumed "author" of the film' (p.4).

At this point the author makes a neat comment about all theorisation. First, he shows that the sharp distinction he draws between narration and narrative exists only as the consequence of the application of a particular theoretical stance; in the actuality of viewing a film the distinction does not exist. And, second, he suggests that the very act of theorising is a kind of framing not unlike the process by which the viewer of a film chooses a particular perspective and subject position.

The remainder of Chapter One offers the reader a critical perspective on a number of theories of point of view: as perception and attitude, as identification, as language and as a logic of reading. It is with the last position that the author's sympathies lie. The approach to point of view as perception or 'a function of the position of the hypothetical observer who stands in for the viewer' (p.6) is shown to be limited in that it tends to favour one particular de vice, the point-of-view shot (of which Chapter Seven gives a detailed account), and because it tends to lead to a predominantly formal analysis. Its extension, the conception of point of view as attitude is criticised for its highly subjectivist explanations and ultimately fallacious reconstructions of authorial vision and thus its trivialisation of the transcendent socio-cultural phenomena of the text. Special attention is paid to the case of Sergei Eisenstein's Film Vision which seems to argue for point of view as attitude. But Branigan is at pains to point out that as far as Eisenstein proposes such an approach, 'that attitude is neither the personal attitude of the artist, nor an attitude determined by Pavlovian reflexology', nor is it 'a transcendental one (eg. "universal humanism")' but rather a 'socially relative attitude which is, in the widest sense, political and specific' (p. 9).

Point of view as identification, Branigan suggest, needs to be understood in terms of its opposite, alienation, whereby both are aspects of a process of 'active participation with a text'. Not much can be gained from such an approach, Branigan says, as long as it is seen from a merely individualistic perspective. If, however, point of view as identification is put forward as an aspect of psychoanalytic criticism, especially in terms of a Lacanian rewriting of Freud's theory, a more productive analysis of film viewing can be achieved. A new understanding of the relation between viewer and author can for instance be accomplished on the basis of seeing the subject as split, whereby author and viewer become problematically linked alternatives of the same process.

As to point of view as language, Branigan notes a dearth of linguistic theory on the topic, with the exception of Benveniste's 'correlation of subjectivity', according to which a speaker (T) and an addressee ('You') confront one another in all language texts. The application by Mark Nash of Benveniste's criteria to an analysis of Vampyr is felt to be flawed mainly on the ground that the reader has been 'compartmentalised and only occasionally appears' and so is made into too passive a feature of film viewing (p. 16).4 In passing, Branigan also criticises approaches to point of view which take for granted linguistics as the master discipline the findings of which can be transferred without distortion to the analysis of the cinema.

Lastly, the approach favoured by Branigan himself is point of view as 'a property of a language system', which would permit him to work 'toward a general theory of representation in film' and an 'explanatory account of how film narration can be understood by a spectator'. Branigan's preceptor in this case is the Chomsky of a generative account of linguistic competence. Transferred to the cinema this means that 'a film spectator, through exposure to a small number of films, knows how to understand a potentially infinite number of new films' (p.17). At the level of theory this allows Branigan 'to give an account of the logic and procedures of our reading - how we come to understand not just a single text but a range of texts and potential texts' (p. 19).

At the heart of this logic of reading' lies the splitting of narration into the five elements 'origin, vision, time, frame, mind,' with narrative or 'object' as an unwieldy sixth (p.20). I contend that there is a theoretical flaw in this schema, an objection which will become clear at the end of this paper.

Chapter Two, 'Film as System', extends a number of structuralist linguistic tenets to provide an explanation of how we can understand film as a systematic formation. This is followed by a brief discussion of three major ways in which the film viewer's and critic's readings are constrained. The chapter ends a little abruptly with a summary of Roland Barthes' five codes: the hermeneutic, proairetic, semic, referential and symbolic, as set out in S/Z. 5

At the centre of Branigan's systemic view of film lies the Saussurean distinction between the axis of paradigmatic (metaphoric) relationships along which speakers of a community select language items from an absent but available reservoir and the axis of syntagmatic (metonymic) relationships which tells a speaker which items are required, permissible or disallowed by the current rules of langue. A point which Branigan is at pains to stress in his application to film of Saussure's (and Jakobson's) distinction, and its use in semiotics, is his rejection of the notion of 'inherent meanings'. All we can be sure of is 'a play of difference' within a system of 'texts and potential texts' (p. 29), so that no film can be regarded as constituting a system by itself.

A marginal point which arises from Branigan's stance is the problem of the ontological status of art, film, literature or theatre. The traditional question in aesthetics of 'what is film?' is dismissed as a futile way of putting the case. It is replaced by an investigation of 'how a text constructs a subject who will propagate the text' and a 'search for the conditions of meaning in a text' (p.30). Even though Branigan has argued against any ontological approach, he poses the question of the validity of reading. Incidentally, he does not employ the term ontology nor does he make it clear whether he opposes only empiricist approaches to ontology or also their phenomenological counterparts which describe the typical acts we perform when we engage in an activity which 'constitutes' the object. Branigan's answer, unlike E. D. Hirsch's in Validity in Interpretation and The Aims of Interpretation, does not seek succour in a theory of meaning but rather in Roland Barthes's theory of codes which act as a set of constraints on the viewing of any film. 6 There are three such restraints: first, multiple coding, such as repetition when 'identical signifiers appear two or more times and share a signified', doubling when 'different signifiers in the same material of expression share at least one signified', and multiple determinations when signifiers appear 'in different materials of expression, which share a signified' (p. 33); second, the limitation of readings by their being embedded of necessity in a cultural context always larger than the film text itself; and third, codes of theory which operate at a metalevel in that they provide the critical basis on which any particular reading occurs.

Chapter Three, "Narration", discusses narration as a symbolic activity - in terms of a series of levels, in terms of two theories of reading, and in terms of representation. The first section opens with a definition of narration as 'a dialectical process between narrator and reader through which is realised a narrative' (p. 39). Traditional approaches to narrative which try to discover its origin in the 'essential subject' of an author, 'the giver of narrative' (p.39), are ditched by Branigan on a number of grounds. First, the notion of the transmission of messages from a source to receivers is based on an inadequate, one-way communications model, especially dissatisfactory when we dealing with works of art. Second, such a theory rests on 'expressionist assumptions' and the mine field of unclarified authorial intentionality. Third, such approaches establish dubious causal links between author and work, connections which ultimately lead to a naive referential fusion of art and 'reality'. Fourth, traditional views of authorship tend to be elitist in that they imply a hierarchy of privilege with an all-knowing author as master interpreter and a mass of readers relegated to the lower rungs of understanding. And fifth, even where the author/auteur appears inside the text (cf. Hitchcock or Renoir) those images do not guarantee any direct access to origins, for as soon as any figure appears in a film it is 'trapped as an object of a film process' (p. 40).

Instead of searching for the author as a biological entity, Branigan suggests we must look for a 'symbolic activity - the activity of narration', a phrase which is used throughout the book as interchangeable with 'narrator' (p. 40). The author, on the other hand, does not speak directly to us but has to be deciphered as a 'subcode of the code of narration', he is not a person but 'an hypothesis constructed by the reader' (p. 41), not a reliable origin we can turn to but merely a necessary ingredient of reading/viewing. Consequently, the auteur who is interviewed about his or her film should not even be regarded as a privileged critic but 'only as another critic'; the author exists only 'as the product of a specialised attention by the reader' (p. 42).

And yet, in Branigan's view, all texts are ultimately controlled by a single, overarching, third-person, omniscient narrational activity. This is the case even where narrational authority has been delegated to subordinate narrators. Looked at from the perspective of origin narration could be said to emanate from a 'single source' but generating 'multiple readers', viewed from the perspective from which 'the reader's position of reception is constant' the narration is realised as spliced into 'different levels each with a nominal narrator' (p. 42). These two perspectives, origin and destination, in conjunction with variations in diegesis, allow Branigan to classify narration into different kinds of reading.

The Aristotelian notion of diegesis is redefined for the purposes of film analysis to stand for 'those elements which give rise to the fictional world of characters, landscapes and events' (p. 43). In this, diegetic features are sharply separated from non-diegetic aspects, such spatio-temporal elements outside the 'world' of the characters, voice-over narrators and voice-over narratees, such as for instance 'the laugh track in television comedies' (p. 47). Using the concepts of 'origin, destination and diegesis' Branigan proposes the following schema for narrative production: omniscient, third-person narration as a level of production without any origin in the text, and subjective, first-person narration as diegetic production assigned 'to a particular origin' or non-diegetic, as in the case of 'identifiable voice-over narrator' (pp. 43f). On the side of reception, Branigan's schema distinguishes between a text which denies its viewer and so produces a 'voyeuristic reader' and text passages in which second-person shifters are dominant and so foreground an implied viewer. This, again, may occur as diegetic narration where a character is the addressee or a non-diegetic relation when a character speaks at or looks at the camera. The resultant relationships are sketched by Branigan thus:

Narrator Reader
1. No origin (omniscience) 1. No destination (voyeurism)
2. Origin located in non-diegetic time and space 2. Destination located in non-diegetic time and space
3. Origin in diegesis 3. Destination in diegesis

Though character may appear foregrounded in this analysis, we should not, Branigan cautions, confuse origin with character. Rather, character must be realised as one possible and specific manifestation of origin, which is a 'structuring principle' at a higher-level of generalisation. Further, since we do not have direct access to any level of narration, what is important in theorising film is the identification of 'markers' which permit us to see the text as split into a number of diegetic and non-diegetic domains. In this, the camera requires special attention. 7 The camera, or better a series of cameras, is understood as at the same time 'a materialisation of a visual symbol system and the labels we apply to that system' (p. 44). For the sake of analysis, Branigan divides camera placement into motivated and unmotivated positions. A camera should be regarded as unmotivated he says, only if it does not fulfil one of the following functions: establishing 'scenographic space', relating to the 'movement by a character or object', intimating a 'glance', selecting a 'significant detail', or producing 'character subjectivity'. Put differently, unmotivated camera is nondiegetic without the production of 'invisible omniscience' (p. 45).

The section on 'Levels of Narration' concludes with the significant point that no matter how many or how few levels can be distinguished, they are all subject to higher planes of narrational organisation. It is for this reason that we find it difficult in the last analysis to draw sharp boundaries between the film text itself and the cultural context in which it is embedded and which acts as a master narration, a text without which film viewing would be unthinkable.

In the section 'Two Ways of Reading' the author looks at two kinds of explanation as to how we construct the continuity of narration. He names one the 'error' theory, and the other the 'hypothesis' theory. According to the former our reading proceeds by attending to the detail of the text as it unfolds, the viewer a faithful recipient of the signs which permit him or her to construct a network of codes. But the viewer is, of course, as much unguided and even misguided by the narration as s/he is guided. As a result, s/he is bound to commit a series of errors which have to be corrected before the conclusion of the viewing. Branigan finds this explanation faulty for two reasons. First, he asks what precisely is an 'error' in the production of film meaning. After all, without such misreadings the film text would not be what it is. Hence, to speak of errors and their correction in the cinema is pointless. Second, the error theory is a one solution explanation, an 'all or nothing interpretation', according to which the viewer is right or wrong. Instead, given the complex interaction of so many levels and codes in film it is far more productive to stipulate a polysemic way of reading.

The hypothesis theory, by contrast, includes 'misreadings' and 'making mistakes' and 'even forgetting'. To shore up this explanation Branigan, as so often in his study, resorts to the work of Roland Barthes according to which the text is 'a nomination in the course of becoming a tireless approximation, a metonymic labour' (p.52). However, the hypotheses of our readings are anything but random; owing to our intersubjective relations with a community, they are structured, systematic (one might prefer systemic), the result of multiple codings, and they are shared. The hypothesis thesis is able to map the 'process of reading' because of its double perspective of the fluidity of reading and its systemic background. It favours the possibility of multiple functions, it gives room to the viewer's input and it replaces the sharp distinction between presence and absence by the more sophisticated concepts of foregrounding and effacement. 8

Branigan's view of representation also relies on the Barthesian formulation of decoupage. This is the cutting out of an object for the vision of a subject from a spatio-temporal frame 'under the pressure of logic or purpose' (p. 56). For the visual arts Branigan stipulates six 'elements' which constitute 'classical representation': origin, vision, time, frame, object, and mind. Origin refers to the spatial locus 'from which the representation derives'; vision is the 'gaze' as 'force', or one might say directional activity, which produces representation from a source point; time is the process by which all aspects of representation are linked as a sequence, continuous or discontinuous; frame is a 'perceptual limit or boundary' which separates the represented objects from those that are not, and hence 'the frame is the measure and logic of the simultaneity of parts'; object is the represented item; and mind is 'that condition of consciousness' which is given as the 'principle of coherence of the representation', such that an object can be said to be represented as a memory or a vision or a perception. Branigan illustrates the role which the six units play in classical representation by referring to the point-of-view shot. In the POV shot origin marks the spatial position of a diegetic character; vision is the character's glance; time is a continuum; frame is recognised as a spatial detail previously identified for the viewer; the object is an item of the diegesis related to the character by way of his or her glance; and mind is given as a major condition of realist art and its concomitant assumptions about the homogeneous subject.

In the last section of Chapter Three, 'The Elements of Representation in Film', Branigan links these six elements with his theory of reading by hypothesis. In this, he illustrates how the film viewer in trying to come to grips with film narration assigns origins. The importance of gaze and vision is underlined by the observation that they allow us to 'define narration in the visual arts as a positioning of the viewer with respect to a production of space, and subjectivity as a production of space attributed to a character' (p. 64). Our construction of time in film is regulated to a large extent by such devices as a 'match on action', continuous sound which bridges a cut, eyeline matches, and camera movement' or, 'slow motion, the freeze frame, or superimposition' (p. 64). Frame in film is always decoupage in the sense of making present and setting something off against what is absent. At the same time Branigan insists that the traditional pair in film analysis of 'selection and organisation' cannot escape the rule of his six elements of representation. The fifth element, object, in its multiplicity in the process of the shots and cuts which make up an actual film, ultimately constitutes narrative, namely that which is represented. As to mind, the last of the units, Branigan, surprisingly perhaps, goes beyond the mere technical significance of such devices as memory or flashback. He suggests that this unit guides the viewer to invoke of necessity the human quality of intelligibility: 'The apprehension of a system of intelligibility in a specific text is the apprehension of a human quality in the text' (p. 66).

Chapter Four addresses the problem of subjectivity in film from the perspective of the analytical apparatus established so far. Subjectivity is defined as a special 'level of narration where the telling is attributed to a character in the narrative and received by us as if we were in the situation of the character' (p. 73). Branigan here restricts the discussion of subjectivity to the realm of the diegetic, leaving aside such cases as identifiable voice-over and other extradiegetic narrative situations. Instead, the focus is on the logic which associates framed space with character as its origin.

When we are dealing with POV shots, Branigan reminds us, we are naturalised into making a series of assumptions, not all of which square with ordinary space-time reality. He refers especially to the convention of the camera assuming the specific spatial coordinates of a character. But such contradictions are resolved ideologically, in the sense of a viewing convention which makes us take the camera 'to be invisible and the character to be real' (p.74). One of the defining features of subjectivity in film is that all of the six stipulated units of representation refer to character. Subjective flashbacks have their origin in character, vision becomes a character's introspection, time the subjective time of a character, frame is the decoupage determined by a character's mental activity, object the result of that activity, and mind the state of a character's consciousness. When these six elements are associated with character, the author says, we face the 'character as subject' and the concomitant 'telling or representing is called subjective' (p.75).

Contrasted to character narration is non-character or reference narration. Of this there are a large number of types such as: voice-over narration separated from story time by a temporal gap, voice-over narration which deceives the viewer, the sarcastic, the mischievous, the ironic narrator, the voice of the dead, of angels, or God, unidentified voices, titles, or epitaphs, and many more. The defining characteristic of such reference narration can be any of the following: 'non-character origin, vision, time, frame, and object' (p.76). Reference narration usually serves to highlight subjective narration.

Branigan's types of subjective narration are functions of the three variables time, frame and mind, with the other three, origin, vision, and object remaining non-defining units. Time can be constructed in four different ways: as present, past, future or undefined time. The last category is of special interest since it includes the phenomena of 'non-continuous and non-simultaneous time' with such variants as 'temporal repetitions, reversals, expansions, contractions, ambiguities, contradictions, etc'. Frame in subjective narration is either 'from the point', ie. when the viewer has available conventions which allow him to link the framed representation to its origin in a character; or it is 'not from the point', in which case the absence of a 'cultural reading practice' prevents such an association. We are then dealing with reference narration. Lastly, the mental state of a character as a defining characteristic of subjectivity should be understood not so much as a psychological entity as an internal organising system which allows us to characterise mind as 'memory, dream, fear, foreboding, hope, fetish, and so forth', all of which can be 'cited by the text' (p. 78).

Branigan suggests that what is important about subjective vision is not so much what is seen, but that a character 'experiences difficulty in seeing' (p. 80). The author also draws a sharp analytical distinction between the POV shot and the perception shot, according to which only the latter signifies a mental condition. However, in the actuality of film viewing this distinction is not guaranteed. On the contrary, there are a number of situations which introduce a degree of indeterminacy into the cinema leading to different and sometimes contradictory reading hypotheses (pp. 83f).

Special attention is given to 'mental process narration', often referred to as 'dream sequence'. It results from the presence of 'undefined temporal markers and the existence of a character's mental condition' which guarantees the logic of what is represented. Traditional ways of introducing mental process narration are dream balloons, rear screen vision, superimposition, matte shots, or split screen images. Inside the mental narration the realist temporal markers of continuity and simultaneity are abandoned. The new undefined time structure can be marked in a variety of ways: 'on the sound track, with superimpositions, titles, slow motion, spinning images, intercuts, lighting, or colour', and sound, as Branigan outlines at the end of this chapter, as well as 'unusual camera locations, angles, and movements' or 'contradictions in the mise-en-scene' (p. 90). But no matter how distorted the mental process narration may be, it is always premised on reference narration against which the viewer must measure the degree of deviation from what is conventionally given as the norm.

The central concern of Chapter Five is the analysis of the point-of-view shot 'in which the camera assumes the position of a subject in order to show us what the subject sees' (p. 103). It normally consists of two shots and six elements, specific instances of the six units discussed earlier. Shot A is made up (1) the Point in space and (2) the Glance from the point. Between shots A and B there is (3) a Transition which suggests continuity or simultaneity. Shot B consists of (4) From the Point which locates the camera at or close to the point and (5) the Object of the Glance. Lastly, shots A and B are linked by the viewer through the construction of (6) Character, the presence of a subject (p. 103). This schema is fleshed out by examples and enriched by a number of useful qualifications. Branigan suggests, for example, that no point may be given - or a number of points can compete with one another, that there may be uncertainty as to whether a glance has actually occurred; that the POV may be metaphorical rather than literal; that though no more is needed for us to realise a transition between two shots other than 'that the last fragment of shot A' should be 'temporally joined to the first fragment of shot B' (p. 105) the transition may never become quite clear; that in fact all elements of the above schema may be modified to produce deviant structures, such as the subjective travelling shot, the absence of an expected object (p. 107), or the ambiguity of character presence and absence (p. 108f.)

Of the POV shot, the author shows, there are two principle kinds with eight variants. Shot A (point/glance) may be followed by shot B (point/object); but the sequence can also be reversed. POV shots may be termed closed when shot A is repeated and so produces a 'high degree of narrative stability' (p. 112); delayed when shots A and B are kept apart, often used to effect suspense; open if we never receive the object promised in shot A; continuing when the glance is directed at the same object more than once; cheated or forged in cases where the object is closer than it would appear to the glance; multiple or interlocking when more than one character glance at the same object; embedded when 'a POV structure of one character is nested or contained within a larger POV structure of another character' (p. 117); and, finally, reciprocal when the object revealed by the glance is another person who returns the glance, a device often employed in conversation sequences.

Chapter Six discusses the representation of subjective states in the form of character reflection which gives us 'only the presence or normal awareness of the character', as in eyeline matches and mirror shots, and character projection which offers 'a specific mental state of the character' (p.123). Thus we should speak of character reflection 'when a character projects his or her body into space, and only the body, not a mental state' (p.127) and of character projection when 'a character's mental state is made explicit' (p.132), usually accomplished by metaphoric framing. This occurs 'where the framing is not strictly from the character's point in space', so that a 'metaphoric transfer' is effected from 'what is looked at to the character who looks' (p. 125). Significantly, such a transfer entails a shift from the metonymic process of narration to a paradigmatic break. Also, Branigan concludes, these devices produce a form of displacement of surface as well as psychical phenomena and a technique of framing which subjectivises the objective (Mitry). But it cannot actually be regarded as first person, it is only metaphorically so.

The confrontation between the expectations accompanying the viewing of the classical film with those of the modern text has been well prepared by Branigan's text so far. It is given a full description in Chapter Seven, 'The Modern Text: Subjectivity under Siege from Fellini's 8 1/2 to Oshima's The Story of a Man Who Left His Will on Film'. Both films are critically metafictional in that they present the making of a film and, more importantly, address 'the conflict between the maker of this inner film and a system of production' (p. 143). Despite their many differences both films fundamentally challenge each of the six units of classical representation. Not that these elements are irrelevant to the reading of a modern text. Quite the contrary; it is only against the system of classical viewing conventions that we can ascribe meaning to such films, meaning as a violation of a reading ideology. Ideology is defined as a 'largely coherent and logical system of images, ideas, values, feelings, and actions by which and through which persons experience their societies' and 'a system of representations through which the individual encounters the material conditions of existence'. Three notions are central to this definition: 'consciousness; system of representation (or text); social ground' (p. 145). In this sense, text production and the construction of meaning are always ideological.

In different ways, and Oshima more radically than Fellini, the two auteurs undermine outmoded notions of authorship as origin, images as mere objects of vision (Branigan here refers us to Brecht's social critique of the use of sentences), character as essentially humanist subjects, time or the relation of narration to narrative as an unambiguous realist construct, the traditional frame as innocent when it is actually a highly voyeuristic film practice, and the unit of mind as grounded in Human Nature instead of in contradictory functions of ideology.

Metatheory, the last chapter, extends the difference between the error theory of reading and the hypothesis theory of reading to the more general level of a distinction between empiricist and rationalist theorising. The tension between these two philosophical stances can be recognised, says Branigan, in practically all film readings. Where an empiricist position acknowledges the presence of a narrator only when it is marked, a rationalist view prefers to talk of a foregrounded or an effaced narrator. Where the former situates point of view in a person's consciousness, the latter favours a definition in terms of linguistic and logical relations as the 'correlation between two narrations' (p.177). An empiricist approach is inclined to argue subjectivity in psychological terms, whereas the rationalist position sees it as 'a paradigmatic function of character where the character represents a set of potential spaces' and, in the classical film, 'a certain reduplication of mise-en-scene confirming (modifying) the viewer's hypotheses about scenic space' (p.177). As before on the question of whether the error theory or the hypothesis theory is to be favoured, Branigan unequivocally supports what he terms the rationalist approach.

Branigan concludes his book by reiterating his belief in the active construction of film space by the viewer and leaves us with a number of issues beyond the scope of his study, such as the question of gaps or lacunae of indeterminacy' (p. 179) and the fundamental problem of 'truth' in the sense of the relationship between film signification and the world (p. 183).

Critical Commentary

I would like to stress that any misgivings about some of the book's theoretical positions do not undermine the fundamental value of Branigan's contribution to narratology and to theories of film narrative. With this proviso I feel I have earned the right to point to a few of the book's shortcomings. I shall address them in this order. I submit that (1) Branigan's extension of some of Roland Barthes's insights is accomplished at a cost; (2) in attacking essential meanings Branigan is flogging a dead horse, while the rejection of every kind of intentionality is rash; (3) his anxious disavowal of the phenomenological method is contradicted by his own procedure; (4) despite an otherwise rich reference apparatus there is a notable absence of a number of important studies on narration, point of view and time; and (5) Branigan's brief flirtation with the notion of 'truth' and the split between signified and referent is unproductive and predates the work of both Niels Bohr and Edmund Husserl.

(1) Branigan's units of origin, vision, time, frame, and mind like Roland Barthes' five codes in S/Z, address one side only of the narrational/narrative relation. But whereas Barthes' hermeneutic, proairetic, semic, symbolic, and referential codes focus on what is told, Branigan's schema is designed from the perspective of narrating. Significantly, Barthes' is a theoretical flaw not found in his actual reading of Balzac's story Sarrasine. There the codes are shown to operate at the same time at both the narrational and the narrative level. Unhappily, in summarising Barthes in Structuralist Poetics Jonathan Culler compounds Barthes' theoretical weakness by suggesting that he should have added a narrational code to the five he listed. This demonstrates not only that the schema misleads the reader, but also that Culler has failed to realise that the hermeneutic code, or the setting and resolving of a question or riddle, is a part both of the act of narrating and of what is told. Likewise with all other codes. Branigan appears to adopt Culler's stance, but conversely so, when he adds to his five narrational units a sixth, the object, or that which is narrated. 9

Instead, all of Branigan's five units (and one could add a number of more abstract ones) can be seen as having a share in the viewer's construction of the process of narrating and of the resulting narrative. His schema would then look something like this: Origin (source of narration and character as object), Vision (the act of seeing, gazing and what is seen, the image), Time (the temporal features of seeing, presenting and the time coordinates of the signified objects), Frame (the act of decoupage and that which is cut out), Mind (the way in which objects are made 'intelligible' and the state of mind as narrative object). This makes it unnecessary to add 'object' as a clumsy extra to Branigan's list of five units; instead, object is already a part of each of those 'elements' namely their narrative aspect.

(2) That there are no essential meanings in film viewing would not come as a surprise in a post-Derridean world, but why Branigan wishes to throw out altogether the notion of intentionality is not so easy to see. Intentionality does not produce essential meanings, but is simply the directional activity involved in such noetic acts as viewing. And at no point does the book suggests that film can exist without such activity. Perhaps Branigan should retain non essential intentionality for the active participation of the film audience. After all, Derrida who has probably offered the most incisive critique of meaning stability has not discarded intentionality, but merely robbed it of its meta physic of presence.' 10

(3) In rejecting phenomenology Branigan constructs a peculiar version of this branch of philosophy. He suggests that an empirical view easily slides toward a phenomenology where a word or image simply "is"' and that 'both phenomenology and realism are direct theories of experience in which there is no important mediation between viewer (viewing) and world' (pp. 169,183). This certainly is not the phenomenology of the mature Husserl, nor of Schutz or Ingarden whose insistence on the noetic modification of all experience is at the centre of their work. Except for the important question of ideology, Branigan's description of the activity of viewing a film has a good deal in common with the stance Ingarden took in the thirties when he began his analysis of the acts we perform when we confront language in literary works. Moreover, at one point in his argument towards the end of the book Branigan notes the need for an explanation of how we 'project new spaces from new angles thus "filling-in" gaps in the presentation of the scene' (p.179). This statement reads like a quotation from Ingarden elaborating on the schematic nature of the literary text - its lacunae of indeterminacy and the reader's task of imaginatively filling in the gaps. Ingarden's notion of 'concretisation' - which in turn is derived from Husserl's 'appresentation' - could be employed to talk about the process of 'filling-in' which Branigan mentions. 11 But Branigan does not seem to be aware of this research.

(4) Since Branigan discusses at some length the contribution to point of view by Wayne C. Booth and other American critics, it is a pity that he has left unmentioned the early work by Franz Stanzel, Narrative Situations in the Novel (1955), in which the author offers a cyclical typology linking first-per son, third-person and figural (free indirect style) narration to a circular arrangement of the older generic categories of the lyrical, epic, and dramatic. Branigan does mention, but brushes aside Boris Uspenski's A Poetics of Composition which is surely more than merely 'a grand synthesis' of the tradition of point of view studies (p.222). Perhaps Branigan could also have acknowledged my own The Reader's Construction of Narrative (1981) in which the relation between 'presentational process' (his 'narration' and 'presented world' ('narrative') is the crucial point of the argument. Lastly, since Branigan is bound in his study to pay attention to temporal functions, Genette's distinctions and his sources, especially G. Mueller's 'Narrating Time and Narrated Time' (1948) as well as A.A. Mendilow's Time and the Novel (1952) would have been useful. 12

(5) A final observation concerns Branigan's somewhat spurious introduction of the notion of 'truth' as the actual relation between a 'system of codes and an ideology' and the 'world' (p.183). It is not consistent with the rest of Branigan's text to distinguish between what can be grasped by means of codes on the one hand and an actual 'world' on the other. Rather, one would have expected the author to suggest that the viewing which we perform by means of the constraints of ideologically informed codes is all the 'world' there is. Branigan's retreat to a Kantian position which separates appearances from a 'world as such' is a little disappointing. After all, Branigan seemed to have persuaded the reader throughout his book that film can be understood as an analogue, albeit perhaps the most complex one, of all perception. 13


1. Edward R. Branigan, Point of View in the Cinema: a Theory of Narration and subjectivity in Classical Film (Amsterdam, Berlin, New York: Mouton, 1984).

2. Percy Lubbock, The Craft of Fiction (London: Cape, 1965); Franz Stanzel Narrative Situations in the Novel (1955; rept. Bloomington: Indiana University Press, 1971); Norman Friedman "Point of View in Fiction: the Development of a Critical Concept", PMLA, no. 70 (1955), pp. 1160-84; Wayne C. Booth, The Rhetoric of Fiction (Chicago: Chicago University Press, 1961); Seymour Chatman, Story and Discourse (Ithaca. Cornell University Press, 1978); Boris Uspensky A Poetics of Composition Berkeley: University of California Press, 1973).

3. Seymour Chatman, 'Point of View in the Cinema: A Theory of Narration and Subjectivity in Classical Film by Edward R Branigan,' Film Quarterly, v. XL, no. 1 (1986), p. 45f.

4. Mark Nash, 'Vampyr and the Fantastic', Screen, v. 17, no. 3 (1976), 29-67.

5. Roland Barthes, S/Z (New York: Hill and Wang, 1974).

6. E. D. Hirsch, Validity in Interpretation (New Haven: Yale University Press, 1967) and The Aims of Interpretation (Chicago: University of Chicago Press, 1976).

7. Edward Branigan, 'What is a Camera?' in Cinema Histories, Cinema Practices Patricia Mellencamp and Philip Rosen eds. (Los Angeles: University Publications of America, 1984), pp. 87-107.

8. Perhaps it is worth emphasising at this point that the distinction has no proper parallel in the act of reading/viewing. There is no 'foregrounded' as against an 'effaced' kind of reading, though we may wish to speak of a highly engaged form of viewing as distinguished from a cursory attention to the text.

9. Roland Barthes, pp. 18ff. and Jonathan Culler, Structural Poetics (London: Routledge and Kegan Paul 1975), p. 203.

10. Eg. in 'Limited inc abc..." Glyph, Johns Hopkins Textual Studies (Baltimore: Johns Hopkins University Press, 1977), p. 249.

11. Roman Ingarden, The Literary Work of Art and The Cognition of the Literary Work of Art (both Evanston: Northwestern University Press, 1973), passim.; see also Edmund Husserl, Cartesian Meditations The Hague: Martinus Nijhoff, 1973), esp. The Fifth Meditation, pp. 50-54.

12. Horst Ruthrof, The Readers' Construction of Narrative (London: Routledge and Kegan Paul, 1981); Gerard Genette, Narrative Discourse (Oxford: Blackwell, 1980); G. Mueller, ' Erzaehlzeit und erzaehlte Zeit', in Festschrift fuer P Kluckhohn und H Schneider, 1948, pp. 195-212; A.A. Mendilow, Time and the Novel (New York. Humanities Press, 1972).

13. The changing views during this century of the relation between sign systems and the 'world' is perhaps best traced by comparing the positions of the Vienna and Oxford positivists, early and more recent phenomenology, and that of Derrida or the stance of the Wittgenstein of the Tractatus as against that of the late 'semiotic' Wittgenstein in the Philosophical Investigations.

New: 5 January, 1996 | Now: 8 March, 2015