A Guide to Narratological Film Analysis
Manfred Jahn

Full reference: Jahn, Manfred. 2003. A Guide to Narratological Film Analysis. Poems, Plays, and Prose: A Guide to the Theory of Literary Genres. English Department, University of Cologne.
Version: 1.7.
Date: August 2, 2003
This page: http://www.uni-koeln.de/~ame02/pppf.htm
Project introductory page: http://www.uni-koeln.de/~ame02/ppp.htm

To facilitate global indexing, all paragraphs in this section are prefixed 'F' for 'film'. If you quote from this document, use paragraph references (e.g., F2.1) rather than page numbers.

F1. Film as a narrative genre
F2. Moving pictures: the visual code
F3. Sound: the audio code
F4. Composition, Narration and Focalization
     F4.1. FCD: The Filmic Composition Device
     F4.2. Narration
     F4.3 Focalization
F5. Case studies
     F5.1. Fixed focalization: MASH, episode 154
     F5.2. Homodiegetic voice-over narration: Wonder Years 24
     F5.3. Verisimilitude and Goofs
F6. Film websites
F7. References

F1. Film as a narrative genre

F1.1. There are three common terms referring to our subject: cinema, motion picture (movie), and film. Because 'film studies' is the generally accepted name of the discipline I will prefer the term 'film' but reserve the other terms for occasional variation. In the following, I will approach film in the framework of the genre taxonomy presented on the project page. In this taxonomy, film, like drama, is listed both as a narrative and a performed genre. Film is mainly realized in the framework of a performance, and like drama, it is related to a textual form (a 'script'). Therefore, just as in drama analysis, film analysis can build on the interesting (but at times problematic) relationship between 'text' and 'performance'. Because film can be fruitfully compared to drama and other narrative genres, such as the novel, the following account will systematically borrow from the concepts defined within both 'narratology' (the structuralist theory of narrative, presented in this project's narratology doc) and the theory of drama (drama doc). Ideally, the reader should already be familiar with these sections. However, for convenience, all main definitions will be repeated here, and, if necessary, adapted to suit the purposes of film analysis.

Approaching film from a narratological angle is not a new idea, in fact the classic studies by Bordwell (1985), Kozloff (1988), Jost (1989), Chatman (1990), Deleyto (1996 [1991]), and Branigan (1992) show that this is a promising project whose synergetic potential is far from exhausted.

F1.2. Because there are strong commonalities between film and drama, our basic definition largely duplicates the definition of a play:

Beyond being a form designed to be shown in public performance, film is related to two kinds of paper media: the film script and the storyboard. To capture the relationship between these types of realizations, we will locally extend the tree diagram of genres presented on PPP's project page as follows:


F1.3. We are assuming that a film, like a play, is mainly a performative genre, that is, a genre designed to be performed, a genre that "comes to life" in a performance (cp. D1.5.2). Watching a film, like watching a play, is a collective public experience and a social occasion. (Watching a film or a theater performance on a television set is not quite the "real thing", rather, it has the status of a substitute performance.) Both drama and film are artifacts created in a process of collective and collaborative production, involving writers, producers, directors, actors, cinematographers, editors, and many more.

F1.4. The 'written' filmic forms can be defined as follows:

Alternatively, the film script is also called a 'blueprint', and the professional reader (the film practitioner) a 'blueprint reader'. See Sternberg (1997: 50n60) for seven sources of this term. Using the terms suggested by Roland Posner (1997), the 'recipe' script is a pre-transcript (preceding the final product or first performance), and the second is a post-transcript, written after (a transcription of) the finished product.

Like the film script, a storyboard can have either the status of a pre-transcript or a post-transcript (as defined above). In the following, I will freely use post-transcript storyboards (photographed from a TV screen) to substantiate definitions and sample analyses. For a storyboard-film script comparison of the famous cropduster scene of Hitchcock's North by Northwest see Giannetti (1993: 159-183; 353-357); for the rooftop chase scene of Vertigo, Auilner (2000: 39).

F1.5. This is not the place to tell anyone how to write a film script. There are many excellent sources both in print and on the net (Epstein 2002 is particularly recommended). A film typescript has a unique standard format which is functional rather than attractive to read. (One of its functional characteristics is that, as a rule of thumb, one page of text is approximately equal to one minute of performance time.) Consider the following excerpt from the screenplay of Rear Window, which introduces the main character's friend:


The excerpt begins with what is technically called a slugline. A slugline usually consists of up to four specifications:

  1. setting, usually either INT. ('interior') or EXT. ('exterior', out of doors),
  2. name of location (JEFF'S APARTMENT),
  3. time of day (lighting conditions),
  4. type of shot (CLOSEUP).

Dialogue is introduced by a speech prefix; manner of speaking may be characterized by a parenthetical (also called "wryly"). The action text, also scene text (Sternberg 1997: 65), contains the descriptions of characters and objects as well as the narrative report of the nonverbal action (this is the filmic equivalent of stage directions in drama). See D3.3 for a detailed account of the dramatic distinction between 'primary text' and 'secondary text'; see F4.3, below, for a storyboard version of the situation presented in the excerpt.

Film scripts are notorious for their technical jargon, and they are often hard to read for the non-professional. Nevertheless, the technical terms that are used in them are very useful for film analysis, and the film script itself can serve as both a record and a reference. Another point of interest, inviting cognitive and linguistic analysis, is that the action text is characterized by a high incidence of specifically "visual sentences" (Epstein 2002). More recently, the practice of booksellers and publishers (Faber & Faber, Macmillan) to offer a range of classic screenplays, as well as an increased attention to the literary qualities of film scripts has led some commentators to elevate the film script to an autonomous literary genre (Sternberg 1997; Korte and Schneider 2000).

F2. Moving pictures: the visual code

F2.1. The units. The smallest unit on a film's visual plane is a frame or cell showing a single picture. If one projects a sequence of twenty-four frames per second on a screen the human eye is deceived into seeing a moving image. A shot is a sequence of frames filmed in a continuous (uninterrupted) take of a camera. A take stops when the camera stops rolling or goes offline. A sequence of shots makes up a scene. Some authors go beyond this and speak of acts (a sequence of scenes containing a major segment of the plot). Finally, a sequence of acts make up a film. (If you want, look up D7.5 in the drama section for a model of classical five-act tragedy.)

F2.2. The conventional system of shot types is based on two distinguishing features: 1) the camera's distance from the object, 2) the size of the object. The system works fine as long as the camera's focal length is normal (i.e., neither wide-angle nor telephoto) and the reference object is a person. The type of shot is much harder to determine when the object is not of standard human size or when the camera uses an unusual focal length. There is hardly any optical difference between the 'mountain range' of the surface of a small object as seen through an electronic microscope, i.e., an extreme close shot, and a true mountain range, i.e., an extreme long shot. Nevertheless, the following terms are good enough for professional use and make up a main part of the vocabulary of the filmic visual code. The four central categories are close-up, medium shot, full shot, and long shot (frames 2, 3, 5 and 6, respectively). Some common intermediate types are listed as well, as are the common technical abbreviations.


(Most of the graphics shown here were taken from CorelDraw libraries.)

F2.3. In the absence of further specification, the camera is assumed to be shooting from a stationary position. If the camera changes its position while filming we get the following types of 'dynamic shots':


F2.4. A cut marks the shift from one shot to another. It is identified by the type of transition which is produced. The two major kinds of cuts are 'direct' and 'transitional'. The direct cuts are as follows:

Transitional cuts, in contrast, are based on an optical effect and usually signal a change of scene (F2.1), i.e., a temporal and/or spatial re-orientation:

F2.5. Camera angles are a result of the camera's tilt (if any): upwards, downwards, or sideways (all to varying degrees). On the screen, the camera's tilt translates into the following principal angles:

F2.6. A fuller account of the elements of the filmic visual code would also include lighting, color values, lenses (standard, telephoto, wide-angle), filters, film stock and graininess, etc. A good account of this is given in Giannetti (1993: 74-76).

F3. Sound: the audio code

F3.1. A film's auditory sources of information are stored on a sound track (magnetic tape or digital medium). Unlike the visual track, the sound track is not a necessary element -- there are 'silent movies' but there are no films without pictures. However, although the visual channel contains a film's essential source of information, this must not be taken to mean that sound -- especially in the form of music and speech -- is in any way less important than the visual signs. See Kozloff (1988: 8-12) for a survey of controversial positions on picture vs. sound.

F3.2. Following Chatman (1990: 134), we will make a distinction between three main kinds of sounds (here treated as mainly self-explanatory terms): noise, speech, and music. Most technical terms correlate sound either to the current scene (F2.1) or, slightly more narrowly, to what is shown on screen.

The following terms relate sound to what is present in the current scene:

The following concepts relate sound to the current shot. Often, the meaning of the audio information is based on non-realistic but inconspicuous conventions, cp. Goffman (1986: 145).

F4. Composition, Narration and Focalization

F4.1. FCD: The Filmic Composition Device

F4.1.1. In order to tell a story, a film is a composition structuring a large amount of heterogeneous information flowing from different channels. Film analysis usually begins by identifying the different channels and sources of information in order to assess their individual contribution to, and function in, the filmic composition as a whole. Often, an important part of this exercise is to assess the relevance, validity, and reliability of the data. Seeing a film, we evidently cope with a flood of information; giving a theoretical account of how we do that is far more difficult. However, the baseline expectation that the film is going to be a functional and effective composition is already a crucial first step towards understanding and interpretation. Indeed, we will assume that a film involves its viewers in a "co-operative" undertaking similar to what is happening in an ordinary conversation involving speakers and hearers (Grice 1975). On this analogy, the viewer approaches the filmic data on the assumption of encountering a well-formed composition guided by maxims of giving efficient, sufficient, and relevant information. As a matter of fact, as film viewers we will actively exploit expectations in these matters, especially when we are facing difficult, incomprehensible, or illogical data (just like Grice's hearers do in a conversational setting, in order to derive interpretive 'implicatures').

F4.1.2. In order to exploit the Cooperative Principle for filmic communication, we need a theoretical agent whom we can assume to abide by the rules and maxims of co-operative storytelling or story-showing. This is clearly a counterpart of the speaker in conversational discourse, or, more aptly, the narrator in novels and short stories. However, a film narrates not by speaking but by arranging and composing information from various sources, sometimes to the extent of including written narrative texts and actual narrative voices. For this reason, I will call a film's primary narrative instance a 'filmic composer' or, more neutrally, a 'filmic composition device':

Various alternative terms have been suggested on "Who Really Narrates" the filmic narrative: the camera (of course), a "grand image-maker" (Metz 1974), a filmic narrator, an implied filmic author, a shower-narrator (Chatman 1990), an implied director, etc. See Kozloff (1988: 44) for a survey of these terms (Kozloff herself settles on the term 'image-maker').

F4.1.3. A film's FCD, as defined above, is a theoretical device that need not be associated with any concrete person or character, particularly neither the director nor a filmic narrator. Of course, it is quite natural to assume that the film's director has exactly the responsibility and authority that we have projected on the FCD. People tend to say that it is Hitchcock who tells the story of Rear Window; that he is the central authority behind all of Rear Window's elements; that Rear Window is his film; that he alone composed and shaped it exactly as he wanted. He can be credited with all its good points, just as he can be blamed for all of its flaws.

But is this factually true? And true for all films and all directors? For the purposes of film analysis, the FCD = director equation does not really work out straightforwardly. After all, we began by defining film as a collaborative product (F1.3.), and on this basis it is evident that many parts of a film's overall information are contributed by sources other than the director. There is the author of the original narrative, the scriptwriter who created the dialogue, the cinematographer who selected the appropriate camera angles, the sound director who devised the sound effects, the composer who wrote the musical score, and the editor who put everything together on the editing table (and many more). Using the names of the real people who produced the film and identifying their individual contribution is often a hopelssly difficult task.


F4.1.4. Rather than refer to the multitude of professionals who actually produced the film, we are better off if we refer to theoretical entities such as the filmic composition device (FCD), the narrator, and the characters. The hierarchy that determines the relative authority of these agencies is illustrated in the graphic on the right. The model shows three levels, but the dotted line around the narrator's box indicates that this is an optional level that may be missing from some (or indeed many) films.

The boxes basically symbolize thresholds of control and knowledge. Any higher-level agent dominates and frames any lower-level agent. A lower-level agent is never aware of the existence of a higher-level agent. The characters do not know that they are characters in somebody's story, and they cannot complain if their acts, views, and motives are misrepresented (cp. N2.3.5 for the identical scenario in narrative prose texts). If a film has a voice-over narrator, that agent is also not aware that s/he is just a subordinate source of information among many (and possibly of questionable reliability), used by the FCD for purposes of its own. Because the FCD is the highest authority in the hierarchy, all filmic information ultimately flows from its mediation, choice, organization, and arrangement. In order to tell the film's story, the FCD can freely adopt, quote, and represent data from sources at its disposal -- it can quote a narratorial voice, for instance, or use a 'POV shot' (F4.3.8.) to present a scene as seen from the point of view of a character, and so on.

F4.2. Narration

F4.2.1. First, remember that not all films make use of narrators. If and when they are present, filmic narrators come in two kinds depending on whether they are visible on-screen or not. Both are speaking parts but only the on-screen narrator is a speaking as well as an acting part.

Obviously, a narrator can be temporarily off-screen or permanently off-screen. Or, looked at differently, a narrator can be temporarily on-screen or permanently on-screen (however, the latter is not a likely configuration). (For a borderline case consider a split-screen scenario, however.)

Not covered by the two definitions above are narrators of written texts, especially inserts, intertitles, introductory written background histories, or closing outlooks (prolepses) on the future fates of the characters, etc. However, these elements are clearly also part of the filmic code.

F4.2.2. In the TV series Wonder Years, the narrator is generally only an off-screen voice-over presence. Episode 21 ("Square Dance", first shown 2 May 1989) is an exception, however, because it opens with the following sequence:


Frames 1 and 2 show the narrator's hands leafing through the pages the Kennedy Junior High School Yearbook of 1969, and that is all we are ever going to see of him (in the whole series, as far as I know). Frame 3 zooms in on the picture of one particular girl, and frame 4 executes the flashback to the story's NOW (20 years ago) -- the moment when Kevin (Fred Savage), aged twelve, learns, much to his dismay, that he has been partnered with quaint (three-pigtails) Margaret Farquhar for square-dance lessons. (See F5.2.1. for a more detailed case study of another Wonder Years episode.)

F4.2.3. As in a novel or a short story, narrators are situated in a discourse-HERE and a discourse-NOW, which generally postdate story-HERE and story-NOW (see N6.3 and N5.1.4; on retrospective narration N5.1.4.). The first-person Wonder Years narrator narrates this particular episode in 1989, but the story itself is set in 1969. Narratologically speaking, the 'narrative distance' between discourse-NOW and story-NOW is exactly 20 years. Logically, the narrator's text is not spoken by the actor who plays young Kevin, but by a mature actor (Daniel Stern, born 1957).

Depending on whether narrators tell a story in which they were involved themselves, or a story about others, they are either 'homodiegetic' or 'heterodiegetic':

See N1.10. ff. in the narratology script for a discussion of the implications of these types when situated in specific 'narrative situations'.

F4.3. Focalization

F4.3.1. A great deal of a film's data can be described as the product of an efficient and relevant (hence cooperative) selection strategy called 'focalization':

The model sketched above (F4.1.) suggests that we can, in general, attribute focalization to one of the following agents: (1) the FCD, (2) the narrator, (3) a character. (There may also be imported external sources, such as clips, which come encoded with their mode of focalization.) When attributing focalization, it is prudent to test all options; often, a specific attribution needs careful interpretation and argument. Questions that need to be raised frequently are: Why is this the character's view and not the narrator's or the FCD's? Could the narrator really have access to the kind of information that we get here? What does the character know and not know at this particular point?

F4.3.2. Focalization theory is of special importance in narratology because it sharpens the more general but fuzzy term 'point of view'. Film analysis, in particular, can benefit from focalization theory because the camera (including the sound recorder associated with it) are quasi-perceptual devices, determining a point of view as well as a point of audition. As a matter of fact, the human eye has a lens just like a camera has. It even projects a picture onto an organic 'screen' called the retina. But: the picture on the retina has not an independent medium: the picture that we see is the result of our cognitive processing of the contents of the retinal array.

F4.3.3. The basic concept in focalization theory is focus, and this term refers to two intricately related things: 1) the position from which something is seen -- in narratological terms, this is the spatiotemporal position of the focalizer; and 2) the object seen 'in focus' -- this is the focalized object or 'center of attention'. Consequently, in film analysis, we will often ask two questions: 1) Who sees?, i.e. Who is (in the position of) the focalizer? And 2) What is the object (thing or human being) that the focalizer focuses on?

F4.3.4. There are special filmic techniques of drawing attention both to focalizers and to focalized objects. Focalized objects, in particular, are often marked by close-ups, zoom-ins, movement, centrality of position, sharpness of focus, shifting focus, increased contrast, spotlighting, etc. (see Giannetti 1993: 50-52 on ways of marking what he calls dominant contrast).

F4.3.5. Turning to an actual example, consider the following storyboard sequence condensing an early scene from Hitchcock's Rear Window. The scene introduces Jeff's friend Lisa and establishes a problematical relationship.


Lisa (Grace Kelly) is first shown in closeup, kissing Jeff (James Stewart) awake (frame 1; cp. also the script excerpt quoted in F1.5). As she moves around in Jeff's apartment, we gradually see more of her, until finally she is presented in a full shot (frame 2). Frames 3 to 5 are typical shots covering the ensuing conversation.

F4.3.6. Which focalization strategy is employed here? Neither Jeff nor Lisa nor a narrator (Rear Window doesn't have one) is in a position to see the couple kissing exactly as it is shown in frame 1. Hence frame 1 can only reflect a 'hypothetical observer' viewing position assumed by the FCD. In a sense, what is shown is an ironical inversion of the sleeping beauty motif: the princess kisses the prince awake, rather than the other way round. Indeed, the film's FCD is often ironic in subtle ways such as these, and from the point of view of plot analysis it is quite literally true that Jeff has to wake up to a full appreciation of Lisa.

Frame 2 presents Lisa from Jeff's POV (now Jeff is the focalizer, and the film presents his perception). Assuming Jeff's position, the camera shows us what Jeff sees (the focalized) -- Lisa, to all intents and purposes a beautiful woman in a fashionable dress. Frame 3 shows both protagonists in a medium shot, and in the background we can vaguely recognize the rear window of the film's title. The conversation itself is shown by the camera moving closely behind Jeff (frame 4), partly showing him, partly showing what he sees, and then using a 'reverse angle shot' shooting over Lisa's shoulder (frame 5). In frames 4 and 5, the camera gets close to, but does not quite move into, or assume, or fully adopt, the perceptual positions of the characters. (See Deleyto 1996 [1991]: 227-8 for additional discussion of this scene.)

F4.3.7. Here is a challenging question: Frame 2 is evidently a POV shot, but do we really see what Jeff sees? Or could it be that even though the FCD lets us see what Jeff sees, what we and the FCD see even in using Jeff's eyes is not the same as what Jeff sees? The film script, for instance, tells us that frame 2 shows us a picture of a woman who "is now full figure, beautifully groomed, and flawless", so that's what we are we supposed to see, and I take it that is what we see. In the film, however, we know that Jeff has strong objections to Lisa in general, and to her high-style evening wear in particular. "If only she was ordinary" (20), he is heard to complain earlier, expressing his conviction that she will never make a suitable partner. How is one to interpret these conflicting pieces of information?

F4.3.8. Let us review the filmic techniques of getting close to a character's POV, of 'adopting' or 'assuming' a character's POV:

F4.3.9. In the context of these subjectifying filmic devices, consider Chatman's (1978: 159) note on Hemingway's "The Killers", a text which is famous for its style of camera-like external views:

Contrasting the 'neutral' style of "outside views" normally attributed to "The Killers" (see N3.3.11), Chatman points out that there are filmic techniques which underline "a character's point of view" and heighten "our association with him". Consider, however: "George looked up at the clock. It was a quarter past six" ("The Killers", The Complete Short Stories of Ernest Hemingway, New York: Scribner's, 1987, p. 218.). To all intents and purposes, this seems to be the exact epic equivalent of an eyeline shot.

F4.3.10. Many definitions in the preceding para involve the word 'gaze' -- a concept which has acquired interesting ideological and psychological ramifications in film studies (Mulvey 1999 [1975]; Kaplan 2000 [1983]). As Giannetti (1993: 403) puts it,

Frame 2, above, is a typical example of a male gaze. As a matter of fact, Jeff is explicitly accused of voyeurism by friends and acquaintances. Question 1: Are there any mitigating circumstances to the male gaze presented in frame 2? Question 2 (difficult): does a POV shot like frame 2 exonerate the FCD of the male gaze? The spectator?

F4.3.11. Exercise: Analyze types of shot and composition of information in the following sequence (also from Rear Window):


Given the definitions listed above, it should be fairly straightforward to categorize frames 1 to 3. To help you along, compare the relevant section of the script:

F4.3.12. Specifically, consider frame 4, above, in conjunction with the following passage from the script:

If you are familiar with the film's plot you can probably explain the significance of this moment. Discuss, using the terms introduced in this section. How would a narrator in a novel present what is going on here? If Jeff were awake, how would he react? What happens to the spectators' awareness of the events? (The latter question is the really difficult one.)

F5. Case studies

F5.1. Fixed focalization: MASH 154 ("Point of View")

The CBS TV series MASH ran through eleven seasons, from 1972 to 1983. In 251 episodes, it portrayed life in a "Mobile Army Surgical Hospital" during the Korean War (1950-53).

F5.1.1. In episode 154, fittingly entitled "Point of View", the camera consistently assumes the point of view of a single character, repeating Robert Montgomery's famous experiment (The Lady in the Lake, 1946) in rather more convincing fashion. According to the Internet MASH guide, "in this unique episode, the camera becomes the eyes of a young wounded soldier. It records his sensory responses to being wounded, flown by helicopter to the 4077th , examined, operated on, and treated in post-operation."


As the storyboard sequence shows, all the shots in the episode are strictly POV shots (F4.3.8.). The perceiving subject or 'focalizer' is always the same character, Private Rich, of whose body we see nothing but an occasional boot, as he is wheeled in on a stretcher (frame 1), or a hand holding a clipboard (frame 9, below). Rich has a throat injury, and temporarily, the clipboard is his only means of communication. There are neither gaze shots nor reaction shots, and the mirror trick that might show us his face is not used (N3.3.10). In other words, the actor who plays him does not have to dress up. (As a matter of fact, that boot might well be an empty one.)

F5.1.2. In a novel, consistent focalization of this type is known as 'fixed focalization' (N3.2.4), and the most likely narrative situation to epxloit it is 'figural narration' (N1.18, N3.3.7). In the MASH episode, fixed focalization is supported and strengthened by a number of subjectifying features. For instance, scene-opening and scene-closing shots coincide with particular states of the focalizer's consciousness -- usually, a fade-in signals waking up, and a fade-out signals falling asleep or loss of consciousness. Pans and oblique angles reflect the movement of the focalizer's head and eyes (in frame 2 he watches the chief nurse at work, in frame 4 he is spoken to by a fellow patient). Other movements of the head (e.g., nodding/shaking yes or no) is also replicated by camera movement. (This is largely a failure, however: shake your head "no" and watch what happens to your area in focus; then consider what happens if this is imitated by a camera.) All diegetic sound is strictly aligned with the spatiotemporal position of the focalizer's head. The episode ends on a shot which shows some members of the camp seen through the frosted windscreen of a departing ambulance (frame 5) (cp. Gombrich 1980: 249 [on creating the effect of perspective by showing partially obstructed views]).

F5.1.3. Rich, the focalizer, knows very little about the characters of this MASH unit. This is rather significant because evidently many viewers of the series know them very well. Indeed, as in many episodes in which the camp is seen from a visitor's point of view, this one is largely a study in character. Consider the following shots:


On the whole, the MASH regulars -- Pierce (frame 6), Father Mulcahy (7), Pierce and Winchester (8), Colonel Potter, the camp commander (9), Pierce, Potter, and "Radar" O'Reilly (10) -- try to appear their best -- they introduce themselves, they explain how the camp is run, they crack the usual dead-pan jokes, and so on. In frame 6, Pierce (Alan Alda) introduces himself as follows (qtd. Kalter 2000: 144):

In Q5.2 I am using this speech for explaining the concept of 'implicit autocharacterization' (N7.1). The unusually high incidence of straight-on shots -- characters looking directly into the camera when talking to the focalizer -- is clearly well suited to the type of close-up characterization aimed at in this episode. In other respects, consistent fixed focalization in film is cumbersome and has some very obvious drawbacks (Peters 1989; see also Branigan 1992: ch. 5 for the topic of subjectivity in film in general, and a case study of The Lady in the Lake in particular).

F5.1.4. One does not have to be a confirmed MASH fanatic to appreciate the excellent quality of many of the episodes. In addition to no. 154, I particularly recommend nos. 92 and 191. In 92, "The Novocaine Mutiny" (1976), "Frank has Hawkeye up on charges of mutiny for various infractions when Potter was away on leave and Frank was the C.O." Frank Burns's account of the "mutiny" is presented as a Mitty-ish wish-fulfillment recollection, amounting to a splendid demonstration of unreliable narration. Episode 191, "Dreams" (1980), is one of the absolute highlights of the series -- "The 4077th can't escape the Korean War, even in its dreams. Exhausted after two days without sleep, members of the 4077th steal away for cat naps and experience dreams that reveal their fears, yearnings and frustrations". The haunting dream scenes are excellent material for a case study of what Kawin (1984: 41) calls 'mindscreen' technique.

F5.2. Homodiegetic voice-over narration: The Wonder Years, episode 24

The ABC series Wonder Years ran through seven seasons 1988-1993. In 115 25-minute episodes it portrayed the formative years of Kevin Arnold (Fred Savage) aged twelve to eighteen. On a more general scale, the series presents a miniature historical picture of life in America in the late nineteen-sixties and early seventies. As Katy Pearce puts it in one of the fine Wonder Years pages available on the net:

F5.2.1. In F4.2.2. above, I already referred to a Wonder Years episode in order to illustrate 'homodiegetic narration' and establish concepts like discourse-NOW, story-NOW and narrative distance. The narrative voice (spoken by Daniel Stern) is a distinctive feature of the series, and in this section, I will make an attempt to work out its special functions and effects.

The Wonder Years episodes are not only excellent examples of filmic homodiegetic narratives but everything that was said of typical first-person narrative situations in the narratology script (N3.3.2) is applicable to them as well. In particular, we can recognize what theorists call the I-I structure of first-person narration, i.e., the split of a narrating person into two "versions": (1) the individual who acts as a narrator (mature Kevin Arnold, aged 33, technically, the narrating I), and (2) a character on the level of action (young Kevin, aged 13, the experiencing I). Many effects are, in fact, direct consequence of the perceptible distance (temporal and psychological) between present and past self. Moreover, the episodes usually enact one of the standard story types described in N3.3.4.: the story of initiation -- a story about a young person's introduction into a new sphere of experience. Indeed, Pearce's quote, above, aptly exemplifies what initiation stories are all about.

F5.2.2. Episode 24 of Wonder Years is entitled "Summer Song" and was first shown on 3 October 1989. Here is a brief sequence from the beginning of the episode showing Kevin reading a letter from his girl friend Winnie. Because the storyboard gives a good impression of the visuals, I have made drastic cuts to the detailed post-transcript available at http://home.t-online.de/home/reynders/wy/episod24.htm .


In the beginning, the narrator executes one of the most common narratorial functions -- he presents a block exposition mentioning dates and historical events (accompanied by appropriate visual clips). Then he situates Winnie's letter and begins to read it out (parts of it are actually legible, see frame 2). In other words, the narrator lends his voice to articulate the perceptions and thoughts of young Kevin. A multitude of different forms and channels is used for presenting the various bits of information and also the point-of-view indicators that tie them to different originating sources. For further tools of analysis in these matters see the narratology page's section on forms of speech and thought presentation (N8). For instance, saying "She'd met somebody", the narrator quotes Winnie by using the technique of 'free indirect discourse'. Interestingly, the original ('direct') version of this free indirect quotation is present as well, both in the immediate context and as visual data.

Incidentally, some narratologists have claimed (Banfield 1982), still claim, actually, that free indirect speech is an "unspeakable form", meaning that it cannot occur in ordinary speech. As Toolan (2001: 135) puts it, "FID sentences are [...] unspeakable [...] since they are impossibly divided between two distinct speakers (narrator and character) and anchorages". What do you say? [Toolan hasn't seen Wonder Years.]

F5.2.3. In the foregoing passage, the audio track that broadcasts young Kevin's voice has no other output than a single sigh. Later scenes, however, present a more complex interaction between the voices of the experiencing I and the narrating I.


Not very surprisingly, perhaps, the main story line of episode 24 is about our boy meeting a girl. Strolling along the beach, Kevin picks up a stray straw hat. Somebody addresses him (frame 1), asking him to give it back. A POV shot shows us Teri ("with an RI"), the owner of the voice (Holly Sampson, frame 2). Nondiegetic music ("Good Vibrations" by the Beach Boys) emotionally asserts how very taken Kevin is with her appearance (the song includes a line about "the way the sunlight plays upon her hair"). But can there be any future in this? Consider the following passage from the script:

Again we see that the narrating I's voice expresses the experiencing I's thoughts. But there is much more to it. Seeing and hearing this scene everybody I know has a grin on his or her face. Explaining comic effects is more difficult than experiencing them, however. But it can be done -- with reference to channels of information, narrative situation, speech presentation techniques, points of view, and focalization.

F5.3. Verisimilitude and Goofs

Film viewers come with a large number of (mostly unconscious) expectations about how the filmic medium presents a real or fictional story. Above all, one generally assumes that the film creates a verisimilar or at least likely world, a world that runs on laws of nature and logic and is, by and large, compatible with what might count as a fact or a possible experience in our own world. The less this principle is disturbed the greater is the film's 'reality effect' (Barthes 1982).

These assumptions are very helpful because we can actively exploit them when we are facing difficult, incomprehensible, or illogical data. The most common strategy in this case is to 'naturalize' the information so that it becomes interpretable according to common patterns of experience after all (Culler 1975). If necessary, we will also locally suspend or modify normal assumptions and expectations for the purpose of dipping into the stranger worlds of dreams, visions, cartoons, fairy-tales, science fiction, alternate histories, and so on. However, these, too, ultimately work on laws of consistency and logic. Having said that ...

F5.3.1. Occasionally, especially if looked at closely, a filmic element may simply not seem natural, plausible, or possible, even allowing for special circumstances. In fact, if ultimately this comes across as a fault for which a professional must bear the blame then we have found a 'goof':

Note that what may look like a goof to someone who has no understanding of the rules and practices of the medium is not a goof when it occurs: (1) in the context of a standard filmic convention (for instance, in 'interior monologue voice over' [F3.2] a character is visible and we hear her/his voice, but her/his lips do not move -- but that is a convention, not a goof); (2) embedded in a representational medium which is part of the fictional world itself (e.g., a fault in a "quoted" home movie clip); (3) as a 'rhetorical' figure serving an ulterior purpose (e.g., an alienation effect [D6.3] intentionally revealing the representation to be an artifact).

F5.3.2. Many goofs will simply be ignored by an ordinary audience. Consider one of the most obvious goofs, the 'heavy-suitcase goof' (not a technical term). How often can one see characters lugging around "heavy" suitcases which are evidently empty and thus not heavy at all. Of course, watching a film, nobody really cares. As viewers we focus on the verisimilar world that we expect to see rather than the distractive or non-conforming detail that might undermine the reality effect. While this holds true for ordinary viewers, whose primary interest lies in immersing themselves in the fictional world, it is less valid for professional and enthusiast viewers, who have special obligations, interests, and viewing habits. (I recently saw an advert for a DVD player whose main selling point was that it allowed easy detection of goofs.) Indeed, for many peripheral audiences, especially on the net, goofs have become collectibles. For two excellent internet sources on goofs see http://moviegoofs.cjb.net and http://www.movie-mistakes.com

Because goofs tend to go unnoticed by general viewers, they throw considerable light on the interpretive impact of media expectations. As will be demonstrated below (F5.3.5), the cognitive shock that comes with the recognition of a goof can also be exploited for entertainment purposes. Increasingly, goofs are also shown paratextually (at the closing margins of the film or show), usually for comic effect.

F5.3.3. Goofs can be categorized according to verisimilarity violation (factual, logic, chronology), area of responsibility (e.g., editing), material causes (mirrors, shadows, etc.), and shooting conditions. In the following, no attempt has been made to create watertight (exclusive) categories, hence a 'chronology goof' can imply a 'logic goof', a 'logic goof' a 'factual goof', and so on. The examples were culled from enthusiasts' web pages, particularly http://www.mash4077.co.uk/goofs.html and http://home.t-online.de/home/reynders/wy/homepage.htm .

F5.3.4. In the following sequence from Vertigo, Scottie and Judy are shown driving towards San Juan Bautista, the missionary settlement where Judy will meet her fate :((. Both characters know this stretch of road because they have traveled it before, in a different car. But one doesn't even have to know these details in order to identify the goof that surreptitiously undermines the verisimilarity of the sequence.


Without any cause or reason, Scottie is driving on the left in frames 2 and 3.

As Auiler (2000: 91) points out, Hitchcock used back projections for most of his car shots. All backgrounds for Vertigo's many driving scenes were shot separately, then added later with the characters sitting in the sound studio's car mock-up. "Though somehow car work is always obvious," Auiler says (2000: 110), "the projection shots in Vertigo are of the highest quality -- and a slight difference in quality can make all the difference in preserving the audience's suspension of disbelief" (2000: 110). Nevertheless, as the example shows, goofs rule okay.

Here is an attempt to explain what really happened. When the back projection was filmed (this was in the very early stages of the production), it was shot from the passenger's (not the driver's) seat. You can verify this by printing out this page and looking at it, against the light, from behind. The view presented in frame 2 could be the result of (inadvertently?) projecting the reverse side of the background footage. At any rate, this would shift the car to the wrong side of the road, as we see it now, and the POV would be the driver's (Scottie's). Perhaps the editing team assumed that nobody would notice because the focus of attention is on the characters, one of them doomed, the other relentlessly pursuing his horrible suspicion. If you have a different explanation, let me know.


F5.3.5. For an interesting case of a pseudo-goof, consider a shot which is part of a TV commercial advertising the "Bayern Alpha" culture channel in Germany. In the frame shown on the right, three channels/sources of information are concurrently open. Two of them are visual, one is auditory. The first visual channel shows a suburban apartment block. The second visual channel shows an insert title saying "Concertgebouw, Amsterdam". The auditory channel broadcasts a piece of classical music.

If, as is natural, one understands the title as identifying the building, there is a jarring contradiction -- the Concertgebouw is a Philharmonic Hall, not an apartment block. On the face of it, this is the result of a factual goof, a fault apparently perpetrated at the editing table. At the same time, one feels a strong impulse to come to grips, to make sense of, the conflicting data -- just as one tries to make sense of a seemingly nonsensical phrase such as Wordsworth's "The child is father of the man" (P3.8). Indeed, there is one aspect in particular that encourages us to make the best of the goof: it is too obvious. Hence one repair strategy is to impute an intention to it. The jarring effect created by it may be there for a purpose. Could the apartment block actually be called "Concertgebouw" (and if not actually, then ironically)? Well, it is a naturalization of sorts, but not a satisfactory one because it does not get us anywhere. Pursuing a different tack, one notes, however, that the 'wrong' text is actually quite compatible with (or, one might say, isotopically linked to) the classical music coming from the audio channel. On this basis, and with a bit of effort, we can construe the message that ultimately makes pretty good sense of what is after all a carefully composed complex of information -- that the Bayern Alpha channel brings the music played by the Concertgebouw to your home, whereever your home may be (i.e., even if you live in one of those ugly apartment blocks). Hence the pseudo-goof effectively overcomes one's protective instinct to ignore TV commercials.

F6. Film websites

F7. References

Auiler, Dan. 2000.
Vertigo: The Making of a Hitchcock Classic. New York: St. Martin's Griffin.
Barthes, Roland. 1982.
"The Reality Effect". In: Todorov, Tzvetan, ed. French Literary Theory Today. Cambridge: CUP. 11-17
Beaver, Frank. 1994.
Dictionary of Film Terms. New York: Twayne.
Bordwell, David. 1985.
Narration in the Fiction Film. Madison: U. of Wisconsin P.
Branigan, Edward. 1992.
Narrative Comprehension and Film. London: Routledge.
Bruner, Jerome. 1986.
Actual Minds, Possible Worlds. Cambridge: Harvard UP.
Chatman, Seymour. 1978.
Story and Discourse: Narrative Structure in Fiction and Film. Ithaca: Cornell UP.
---. 1990.
Coming to Terms: The Rhetoric of Narrative in Fiction and Film. Ithaca: Cornell UP.
---. 1999.
"New Directions in Voice-Narrated Cinema". In: Herman, David, ed., Narratologies. Columbus: Ohio State UP. 315-339.
Dana. 2000.
"The Daily Script" Glossary Page. www.dailyscript.com/glossary.html .
Deleyto, Celestino. 1996 [1991].
"Focalisation in Film Narrative". In: Onega, Susana; Garcia Landa, José Angel, eds. Narratology. London: Longman. 217-233.
Epstein, Alex. 2002.
Crafty Screenwriting. New York: Holt. Internet: http://www.craftyscreenwriting.com .
Giannetti, Louis. 1993.
Understanding Movies. 6th ed. Englewood Cliffs: Prentice Hall.
Goffman, Erving. 1986.
Frame Analysis. Boston: Northeastern UP.
Gombrich, E.H. 1980.
"Standards of Truth: The Arrested Image and the Moving Eye". Critical Inquiry 7: 237-273.
Jost, Francois. 1989.
L'oeil-caméra: Entre film et roman. 2nd ed. Lyon: Presses Universitaires.
Kaplan, E. Ann. 2000 [1983].
"Is the Gaze Male?". In Kaplan, E. Ann, ed., Feminism and Film. Oxford. Oxford UP. 119-138.
Kalter, Suzy. 2000.
The Complete Book of M*A*S*H. Abradale Press.
Kawin, Bruce. 1984.
"An Outline of Film Voices". Film Quarterly 38.2: 38-46.
Korte, Barbara; Schneider, Ralf. 2000.
"The Published Screenplay -- A New Literary Genre?". Arbeiten aus Anglistik und Amerikanistik 25.1: 89-105.
Kozloff, Sarah. 1988.
Invisible Storytellers: Voice-over Narration in American Fiction Film. Berkeley: U of California P.
MASH, episode 154, "Point of View".
Directed by Charles S. Dubin, written by Ken Levine and David Isaacs. November 20, 1978.
McInernay, Jay.
Bright Lights, Big City. [1984.] London: Penguin, 1993.
Metz, Christian. 1974.
Film Language: A Semiotics of the Cinema. New York: Oxford UP.
Mulvey, Laura. 1999 [1975].
"Visual Pleasure and Narrative Cinema". In Thornham, Sue, ed. Feminist Film Theory: A Reader. New York: New York UP. 58-69.
Peters, Jan Marie. 1989.
"The Lady in the Lake und das Problem der Ich-Erzählung in der Filmkunst". In Albersmeier, Franz-Josef; Roloff, Volker, eds. Literaturverfilmungen. Frankfurt: Suhrkamp. 245-258.
Posner, Roland. 1997.
"Performance and Transcripts: Towards a Theory of the Media". Plenary lecture. Second IALS Conference, U of Freiburg.
Rear Window. 1954.
Dir. Alfred Hitchcock. Perf. James Stewart, Grace Kelly, Raymond Burr. Adapted from "It Had to be Murder", by Cornell Woolrich. Paramount 1954. Film script by John Michael Hayes (Classic Movie Scripts http://www.geocities.com/classicmoviescripts/ ).
Reynders, Peter. 2002.
Peter's Wonder Years Guide. http://home.t-online.de/home/reynders/wy/homepage.htm
Sternberg, Claudia. 1997.
Written for the Screen: The American Motion-Picture Screenplay as Text. Tübingen: Stauffenberg.
Wonder Years, episode 24, "Summer Song".
Written by Mark B. Perry. Directed by Michael Dinner. The Black/Marlens Company. 1989. Transcript available at Reynders (2002).