1. An Invitation to Feel

Greg M. Smith

Film Structure and the Emotion System (Cambridge:  Cambridge University Press, 2004)

When nonacademics learn that I am writing a book, the polite conversationalist will ask what my work is about. After I reply that my book looks at film structures and emotion, inevitably their response is something like: "Isn't that an enormous subject? There must be so much written about film and emotion." Since emotions are so central to most people's cinematic experiences, they assume that film scholars must have placed the topic of emotion at the top of their research agenda. Most nonacademics are surprised to learn that there is relatively little written by cinema scholars on film and emotion per se.

But cinema studies is not unique in its neglect of emotion as a topic of study. From the Fifties to the Seventies, few academic disciplines gave precise attention to the topic of emotions. Cultural anthropologists had difficulty reporting such highly "subjective" states of mind using traditional methods of observation on other cultures . Instead they focused on more externally observable differences, such as those in language and ritual performances. Sociology's agenda led academics to areas where socialization was most clearly at work. These thinkers recognized that emotions were manipulated by society, and so they tended to view emotions in a purely instrumental fashion, as means to an end. Social forces relied upon fear or love to create prejudice or empathy, but few sociologists questioned the basic nature of these emotions.

In psychology, behaviorism's influence led theorists away from anything located within the "black box" of the human organism. When cognitive science arose to challenge behaviorism with a new emphasis on internal representations, researchers agreed not to consider emotion. Emotions, unlike memory or perceptual tasks, could not be simulated on computers, and so emotions were often considered to be "noise" unique to the human hardware, a possible source of interference with cognitive processes. Similarly, the emphasis on reason within the philosophy of mind kept many philosophers away from "messy" states such as emotions.

The past decade has produced a flurry of new research across academic disciplines in the neglected topic of emotions. New anthropological methodologies encourage researchers to examine more "subjective" states in their complexity. In many cases this emphasis on emotion grew out of earlier research problems. Studies of the effects of mood on memory helped open up cognitivists to considering emotion, as research on empathy did for sociologists. The research on emotion in these fields still carries traces of these originating questions.

Like these other disciplines, film theory has historically paid only spotty attention to emotional effects, though almost everyone agrees that eliciting emotions is a primary concern for most films. In the modern world's emotional landscape, the movie theater occupies a central place: it is one of the predominant spaces where many societies gather to express and experience emotion. The cinema offers complex and varied experiences; for most people, however, it is a place to feel something. The dependability of movies to provide emotional experiences for diverse audiences lies at the center of the medium's appeal and power.

Emotions are carefully packaged and sold, but they are rarely analyzed with much specificity by film scholars, particularly in the modern era. Some classical film theorists, particularly Sergei Eisenstein, foregrounded emotion as one of the primary goals for filmmakers, but Eisenstein's broad discussions of emotion did not give other theorists a specific foundation for discussing emotion (unlike his more concrete prescriptions about editing).(1) André Bazin also emphasized that filmmakers should evoke emotion, but he foregrounded a particular means of eliciting that emotion (realism). In the modern era, studying emotional responses to films became the task of quantitative communication researchers, resulting in a large body of work on topics such as the effects of media violence on children. Film theory devoted more of its energy toward issues of cinematic specificity, arguments over aesthetic valuations, and understanding of representation.

Contemporary film theory of the Seventies concentrated on issues of meaning and representation and their ideological implications. Some writers (Christian Metz being the most influential) attempted to arrive at a theory of "pleasure" and "desire." What pleasure does the cinema afford, and what desire motivates our viewing? Linking the Althusserian bent of ideologically based theory with theories of Freudian/Lacanian subject formation, Metz foregrounded identification as the principal emotive effect in film. Many film feminists including Laura Mulvey, Mary Ann Doane, Linda Williams, and Christine Gledhill have struggled with the phallocentric assumptions of these Freudian/Lacanian theories, refining but not reconstituting the central concepts of psychoanalysis to address feminine film pleasures in particular.

However, the concepts of pleasure, displeasure, and desire used in film studies are too broad to provide specific insight into how a particular film makes its emotional appeal at any given moment. If the range of emotion in the film theater is reduced to some point on the continuum between pleasure and displeasure, we lose the flavor of individual texts. Similarly, if we claim that all mainstream film viewing emerges from the same scenario of repressed desire, we ignore the diverse motivations driving the spectator's interest and emotion. Recent psychoanalytic theory has attempted revisions to correct its reductive, overly-broad approach. It has posited various positions of desire, rather than the former one or two positions. It has also articulated contradictory pleasures in an effort to make discussions of emotion more nuanced and specific. Nevertheless, I believe that, while investigations of cinematic pleasure and desire are certainly important, the ambiguity and spaciousness of these concepts, as currently used, compromises their usefulness. These general concepts are a poor basis for a specific theory of emotion, making them a poor foundation for an approach to filmic emotion.

A far better candidate, I believe, for a productive theoretical perspective on emotion is the interdisciplinary mix of psychology and philosophy called "cognitivism." My book can be understood as part of a burgeoning new area of film studies that asks how cognitive research into mental functions can help us better understand the film viewer's task. David Bordwell, Noël Carroll, Gregory Currie, Murray Smith, Joseph Anderson, Edward Branigan, Torben Grodal, and Ed Tan have all produced major works on film theory from a cognitivist perspective,(2) and the work you hold in your hands should be seen as an outgrowth of their previous efforts. Although my emphasis on embodied emotion leads me away from emphasizing "pure" cognition, it is important at the outset to position my own efforts against the assumption that all these theorists (myself included) share: that a well-founded knowledge of how mental processes work can provide a solid basis for film theory.

I begin my approach to the topic with the seemingly tautological assertion that film emotions are first and foremost emotions. Unfortunately this assertion is rarely made in critical literature on film. When film academics do address emotion, they generally proceed as if the concept of emotions were clearly and widely understood; therefore the task of the film scholar is to say what is specific about filmic emotions. I believe that we have relied too long on commonsense understandings of emotions in such discussions. We need to better understand what emotions are before we discuss any unique qualities of filmic emotions. And so this book will not discuss specifically filmic emotions. Nor will I examine the specific nature of emotional responses to fiction, which much recent philosophy has investigated. There may be specific responses that the cinema and no other medium is uniquely qualified to elicit, but this book will spend little time examining such questions. Films are objects that are well constructed to elicit a real emotional response from our already existing emotion systems. Given a better understanding of that system, what film structures are seem well suited toward activating that structure?

I lay out both a theory of emotion and an approach to filmic emotion. In chapter 2 I survey recent psychological research on emotion and synthesize a theory of how the emotion system is structured. Based on my psychologically-rooted theory of emotion, I then formulate an approach to analyzing a film's emotional appeals. Following Kristin Thompson's usage, an approach is "a set of assumptions about traits shared by different artworks, about procedures spectators go through in understanding all artworks, and about ways in which artworks relate to society." I am not outlining a method ("a set of procedures employed in the actual analytical process)."(3) I share Thompson's concern that preconceived methods tend toward predetermined outcomes, which narrows the analysis. The theory of emotion outlined in this book provides the grounding for the approach's assumptions about emotion, but a theory alone does not show a critic how to analyze particular texts. On the other hand, an approach without a theoretical foundation can provide innovative readings, but it needs to be rooted in cohesive, systematic whole to be convincing.

What would a good combination theory/approach to analyzing filmic emotion look like? How might someone determine if one approach were better than another? In order to evaluate the approach to filmic emotions in this book, we should agree on some desiderata for an approach:

1) A good approach to filmic emotions should provide specific explanations, not generalizations, for how particular films elicit emotions. A good approach should provide different explanations for how film A and film B elicit emotional responses. If the approach reduces different films to the same mechanisms over and over, then it is reductionist. Such tools lead critics away from specific consideration of individual films, and a good approach should lead them toward the particulars of a film.

2) A good approach to filmic emotions should provide terminology for discussing emotions and how they are evoked. This desideratum is an outgrowth of the previous one. If a theory of the emotions is to be productive, it needs to give us a language to talk about the "messy" world of emotions with specificity and particularity. We need two kinds of terminology: we need to be able to label emotional states with some measure of certainty, and we need terminology to discuss the film structures that encourage these responses.

3) A good approach to filmic emotions should be able to explain emotional phenomena at the global and local level. A single film can elicit a wide range of different emotions and yet still have a kind of emotional unity. We need an explanation for the broader processes of emotion that operate across entire films as well as the more minute processes that govern scenes, and we need a way to describe how these global and local processes cooperate.

4) A good approach to filmic emotions should not only label emotional states but be able to discuss how they change over time. How does our emotional reaction evolve, progress, wane? How does a film change from one emotion to another? The challenge is for a theory to explain both emotion as stable state and emotion as dynamic process.

5) A good approach to filmic emotions should be able to explain why films are able to elicit such dependable emotional reactions across a broad range of audiences without denying the variations among individual emotional reactions. Again, the problem is dual. Films do get remarkably similar emotional reactions when shown to a variety of viewing audiences, and this continuity of response needs explanation. However, the approach should not endorse a single emotional response as being the only valid one because of the incredible range of reactions among individual viewers. Although no theory can explain every individual's emotional response, it should be able to explain how such a range of response exists.

6) A good approach to filmic emotions should be able to explain the emotion in a wide range of films. If an approach works for melodrama or contemporary cinema but not for action-adventure or silent films, then it is of limited use value. If an approach is prejudiced towards the kinds of emotional appeals made by the classical Hollywood cinema, for instance, it will tend to reshape other films to fit that mold. Although it may be true that the classical cinema may require a different explanation of emotion than the art cinema, an approach that explains both would clearly be more powerful.

7) Similarly, a good approach to filmic emotions should be able to discuss a wide range of cinematic signification. Films use an enormous set of mechanisms to elicit emotion: lighting, camera, acting, sound, music, mise-en-scene, character, narrative, genre conventions, etc. If an approach to filmic emotions concentrates too heavily on one of these mechanisms, then the approach is likely to miss much of the other emotion cuing in the film. For instance, we will see that many theories emphasize character identification as the pivotal mechanism in filmic emotions. My approach opens up a discussion of the emotional significance of a broader range of cinematic cues, emphasizing the importance of cinematic style in encouraging emotional responses.

7) A good approach to filmic emotions should be able to explain not only why a film succeeds in eliciting emotions but also why another film fails to do so. To show how a film cues emotions is only half the battle. If a theory is to have explanatory power, it must also be able to explain how some films fail to generate emotions. If the approach cannot explain why certain narrative structures are less effective in cuing emotion, then it is too broad to provide specific insight into effective film structure.

8) A good theory of filmic emotions should generate specific research questions for future research. If a theory seems to explain everything so totally that it precludes further investigation, then the theory is too totalizing to be useful to researchers. A theory that generates a number of interesting, investigatable questions for research is of more use to scholars than a grand theory.

9) Finally, a good approach to filmic emotions should be rooted in a body of theory and empirical research on the emotions. Film studies should take advantage of the explosion of research that has been done on the emotions in various fields during the last decade. Our conception of filmic emotions should be consistent with the best available models of how the emotions work. In particular, the researchers who are using empirical methodologies (particularly psychology, anthropology, and sociology) have an edge in discovering new insights into the emotions. All too often discussions of the emotions perpetuate generally held beliefs about the emotions that may or may not be true. Of course empirical researchers are susceptible to the same self-perpetuating ideas, but at least they have the opportunity to encounter data that refute these ideas. Empirical research is particularly useful in dealing with the emotions to create a solid foundation that is rooted in real world processes.

Film Structure and the Emotion System proposes to bring to film theory a more nuanced understanding of what emotions are and how they function, based on current researches in experimental psychology. Advances in neuropsychology have opened up new and more complicated understandings of the brain's interconnectivity, making it more difficult to separate "reason" from "emotion" in any strong sense. Current researchers in neuropsychology are adding more finely tuned tools to investigate emotion (in addition to those developed by traditional experimental psychology), giving psychology a more precise understanding of the basic nature of emotions. This burst of new attention to emotion in psychology has altered that field's basic understanding of what emotions are. One of this book's basic assertions is that film studies' understanding of emotions should be consonant with (or at least not directly contradict) the best current understanding of emotions in psychology.

There is much that psychology is not well suited to tell us about emotions. For instance, anthropology is more likely to produce work sensitive to cross-cultural differences in concepts of emotion, and sociology is more geared toward examining how socialization shapes people's emotions. The burgeoning research on emotion in these fields should help film studies' continuing effort to explore the importance of cultural difference in film viewing.

However, I believe that we should begin this close attention to filmic emotion using the insights provided by experimental psychology. The insights provided by sociology and anthropology should not contradict the psychological ones, since culture and socialization shape the individual but do not fundamentally rework the basic structures of the human. Sociocultural experiences help us define what particular emotions are, but the shape of the emotion system itself and the basic mechanisms by which it operates are best articulated by close attention to the individual.

Film Structure and the Emotion System is concerned with this emotion system's structure, rather than with particular emotions themselves. This is not a book about sadness or joy; instead it deals with the foundational structures that make such emotions possible. Culturally nuanced work on particular emotions certainly needs to be done, but we should make sure that we first understand the basic principles of how the emotion system is constructed. Although the subject of this book may initially seem too all-encompassing, in actuality its aims are quite humble. I am not attempting to explain all of emotional experience. I do assert, however, that an understanding of the basic nature of the emotion system should provide a firm (if limited) foundation for more fully understanding filmic emotions.

Nor am I attempting to provide a particular theory for how each cinematic component evokes emotion. I will not present a theory of music, followed by a theory of facial expression, followed by a separate theory of camera framing, etc. Instead, I present a theory of how the emotion system is designed to coordinate information from these subsystems. Certainly I believe that more work needs to be done on the specific relationship between emotion and camera framing, emotion and film music, etc. In applying my approach to films in the case studies, I frequently return to music (for example) as an important factor in cuing filmic emotions, thus emphasizing the need for a more systematic approach to film, music, and emotion.(4) However, a complete theory of the interrelationship among music, film, and emotion is outside the scope of this study. I do not believe that we must wait for a complete theory of emotion in each cinema component before we can make valuable assertions about the nature of filmic emotions.

As the title of this work indicates, this book is about both emotion and film structures. Once we have a better understanding of how the emotion system operates, we can then take note of how certain film structures are particularly well designed to manipulate emotions.

The primary research questions addressed by this work are: 1) how is the emotion system structured? and 2) what filmic structures seem particularly well suited to take advantage of the properties of this structure?

For the purposes of this work, "film" will be assumed to be a highly coordinated visual (and usually audial) medium that has developed a number of conventional strategies (shared by producers and consumers) for storytelling in uninterrupted real time. This book deals only with emotion and the structure of narrative film. Admittedly audiences respond to nonnarrative/avant-garde films, and such responses call on the same emotion system in its audience members as do narrative films. However, the structures in avant-garde film are considerably different from those in films trying to tell a story, and so these structures are outside the realm of this work.

A crucial assumption for this book is that film takes place over uninterrupted time. Temporal limits for emotionality are built into the emotion system, and so the temporal unfolding of the filmic stimulus is crucial to the way it appeals to this system. A similar audio-visual medium such as American broadcast television is more interruptive, and so it is structured differently in the way it tells stories. It deserves separate consideration of how its structures appeal to the emotions.(5)

Also important to this understanding of film is that audiences and producers share a certain set of narrational conventions that allows the story to be told and understood. When I examine film structure in Eisenstein's Strike or Vidor's Stella Dallas, I will assume an "educated viewer," one familiar with the necessary basic conventions for making sense out of the film. This viewer is not necessarily a "spectator" (with all the implications of being ideally "positioned" by the cinema to receive pleasure), nor is he/she an actual person with specific experiences at the cinema. The "educated viewer" simply has the (not necessarily conscious) knowledge to allow him/her to understand the story and to respond accordingly emotionally.

Of course not all actual individuals who view a film may have the required knowledge to understand and respond to the film. I am not trying to specify the emotional responses of all individuals sitting in front of a screen watching Casablanca. There will necessarily be individual differences in emotional responses, depending on the ways that people's personal experiences have shaped their particular emotion networks. However, I do not want awareness of this variation in people's emotion networks to paralyze film criticism. It is all too easy to conceive of emotions as so messy and individualized that one cannot talk about them with any specificity.

Films do not "make" people feel. A better way to think of filmic emotions is that films extend an invitation to feel in particular ways. Individuals can accept or reject the invitation. Those who accept the invitation can accept in a variety of ways, just as those people invited to a party can be doing very different activities. Although there is much variety among what partygoers are doing, there are implicit conventions that set limits on the ways that most people accept the invitation (for instance, showing up to the party with no clothes on would in most situations be considered a violation of convention). One can acknowledge the range of partygoers' responses and simultaneously conceptualize a hypothetical "educated partygoer" who knows the rules for party behavior.

Films offer invitations to feel. Film audiences can accept the invitation and experience some of the range of feelings proffered by the text, or they can reject the film's invitation. In order to accept the invitation, one must be an "educated viewer" who has the prerequisite skills required to read the emotion cues. Not all "educated viewers" accept the invitation, of course. You can properly recognize how a film is cuing you to feel and still reject the invitation by not feeling. Film complicates the "invitation" metaphor, because film does not extend a single invitation but a succession of invitations across time. Film continually keeps extending invitations us to feel, and we can accept or reject any one of them.

As long as the film critic is dealing with the film text alone (as I am doing in this work), there are limits to the emotion analysis that can be done. I can discuss the range of emotional responses that a film encourages, but I cannot specify a particular emotional response without having access to a particular audience member's reaction. This is a book that examines the ways that particular films cue emotion, not a book about actual human emotional responses. Since my method deals with the text, I can only elucidate the primary invitation that the film offers. In textual analysis it is also difficult to examine highly complex emotional states (for example, existential angst), at least until more work on the sociocultural bases of higher-level emotions is done. For now my textual analyses will concentrate on relatively less complex emotional states because they can be more dependably discussed, at least until further work by sociologists and anthropologists of emotion allows us to understand these more complicated emotions in more detail.

Such textual study is a necessary prerequisite for a film scholar dealing with emotions. We have much to learn about what emotions are and how films cue emotions before we venture into the significantly more complicated laboratory of the real world. By knowing more about the emotion system, we can better articulate what texts are doing and potentially discuss audience responses in more particularity.

In chapter 2 I synthesize a range of research from experimental psychology and neuropsychology to create a cohesive, empirically based understanding of the emotion system. In chapter 3 I discuss the implications of this research on our concept of film structure and demonstrate the basic concepts involved in analyzing a film's emotional appeals using four short case studies (Raiders of the Lost Ark, Ghostbusters, Local Hero, and Stranger than Paradise). Chapter 3 outlines a basic approach for the critic analyzing a film's emotional appeals, which I call the "mood-cue approach." In chapter 4, I situate the mood-cue approach in the context of the growing body of literature on film, cognition, and emotion, differentiating my own approach from the assumptions of my predecessors Noël Carroll, Ed Tan, and Torben Grodal.

Chapters 6 through 9 provide more full case studies of how my conception of the emotion system can provide insights into texts. Each case study is designed to test a different capability of the critical approach. Can it explain how one text successfully shifts from one emotional appeal? Can it explain how a film might exhaust the audience's emotion system? Can the mood-cue approach explain how a film's emotional appeal fails? Can it provide a more insightful explanation than previous critics have done of the timeless appeal of a film like Casablanca? Can it mine new insights into a film that has been exhaustively studied by psychoanalytic film scholars (Stella Dallas)? Chapter 10 then provides a brief conclusion, suggesting future avenues of research open to scholarship. For those interested in such matters, an appendix follows that examines the Freudian assumptions about the nature of emotion that are the underpinnings of psychoanalytic film theory. In that appendix I argue that these assumptions make Freudian-based theory a poor choice to explain the nature of emotions.

The analytic approach defined here is not prejudiced toward one particular cinematic narrational paradigm in a way that promotes "cookie cutter criticism." The mood-cue approach demands the critic pay close attention. The goal of this approach is to help critics to be able to see and articulate the cinematic structures that appeal to audience emotions.


1. For more information about Eisenstein's very prescient intuitions about the nature of emotion, see Greg M. Smith, Movie Moods: The Emotion System and Film Structure (diss., U of Wisconsin-Madison, 1998), chapter 3.

2. David Bordwell, Narration in the Fiction Film (Madison: U of Wisconsin P, 1985); Noël Carroll, The Philosophy of Horror, or Paradoxes of the Heart (New York: Routledge, 1990); Gregory Currie, Image and Mind: Film, Philosophv. and Cognitive Science (Cambridge: Cambridge UP, 1995); Murray Smith, Engaging Characters: Fiction, Emotion, and the Cinema (Oxford: Oxford UP, 1995); Joseph Anderson, The Reality of Illusion: An Ecological Approach to Cognitive Film Theory (Carbondale: Southern Illinois UP, 1996); Edward Branigan, Narrative Comprehension and Film (New York: Routledge, 1992); Torben Grodal, Moving Pictures: A New Theory of Film Genres, Feelings, and Cognition (Oxford: Oxford UP, 1997); Ed S. Tan, Emotion and the Structure of Narrative Film: Film as an Emotion Machine, trans. Barbara Fasting (Mahwah, NJ: Lawrence Erlbaum, 1996).

3. Kristin Thompson, Breaking the Glass Armor: Neoformalist Film Analysis (Princeton: Princeton UP, 1988) 3.

4. For an intriguing introduction to issues around cognition, emotion, and film music, see Jeff Smith, "Movie Music as Moving Music: Emotion, Cognition, and the Film Score," Passionate Views: Film, Cognition, and Emotion, ed. Carl Plantinga and Greg M. Smith (Johns Hopkins UP, 1999).

5. It is possible that the portions of the discussion in this book may also apply to other audio-visual narrative media that unfold uninterrupted in real time (for instance, certain kinds of theater). This is not to say that there are no differences between the two media (for instance, the theatrical convention of a greater emphasis on language) and their emotional appeals, but that their narrational structure may call on very similar principles.