116 Chapter 3 Mindreading Action is character. Syd Field1 Mindreading: Character Psychology Inferences and Judgments A middleschool girl clutching a tray surveys the cafeteria. Several empty seats are quickly taken. She looks around, dreading her choice: where will she sit? This is the opening of Todd Solondz’s 1996 film Welcome to the Dollhouse and the beginning of the characterization of its protagonist, Dawn. In their most typical and ordinary scenes, movies dramatize human behavior for spectators to observe, and on the basis of their observation they make inferences and judgments about characters. As we watch we attribute various mental activities to Dawn: her thoughts and feelings, her place within her environment, and her personality traits. We also make some inferences about the setting and the other characters—they seem to be sources of suffering to Dawn. One indispensable aspect of characterization is the construction of character psychology. Some films accomplish this kind of characterization through direct means of subjective narration, explicitly representing thoughts, dreams, memories, or fantasies, or describing mental states explicitly in dialogue or titles. Yet some films, including Welcome to the Dollhouse, avoid subjectivity and verbal description, and their characters
117
are not consequently any less developed or complex. It might seem to be just common sense that we attribute thoughts to characters, and that this functions in a way similar to our attribution of thoughts to other people in real life. But this is actually more sophisticated than it seems for a number of reasons. First of all, social cognition research has produced a significant body of scholarship that is relevant to understanding how films tell stories. We should not ignore this research about how the mind understands other minds, which is a cognitive function that is crucial not only to social organization but also to understanding the appeal of narrative. Second, understanding how character psychology is constructed has enormous utility for film criticism, allowing us to replace vague terms such as “characterdriven” with more precise and informative notions of how narratives structure character engagement. It will be useful, I hope, in understanding how American independent cinema constructs characters. Third, social cognition research may help explain certain design features common to most narratives, especially canonical or prototypical narratives. For example, there are conventions of character introduction and narrative exposition common to narratives of all shapes and sizes. The way narratives introduce new people exploits certain cognitive structures, as we saw in Chapter 2, and is tailored to patterns of inference and judgment either to facilitate or, more interestingly, to frustrate or delay
118
comprehension. Fourth, an understanding of character psychology grounded in social cognition research is potentially applicable to other aspects of the study of film and other narrative media. For example, the social, moral and ideological rhetoric of narrative is typically a product of particular constructions of character.2 And fifth, an understanding of character based on social cognition research is implicitly opposed to a conception of narrative in film studies and related disciplines as working entirely on the basis of arbitrary conventions or codes of representation.3 This semiotic, structuralist or post structuralist conception of narrative is assumed by scholars who speak of characters in terms only of their intertextual connection to genre norms, star images, or discourses of identity. Of course characters are products of conventions of characterization, many of which are culturally and historically specific. But they are also products of natural processes of social cognition, or of nonarbitrary conventions, and to discount this is to miss a significant part of the story. There are several related components of character psychology. In the opening scene of Welcome to the Dollhouse, there are at least three kinds of inference or judgment that the film solicits from spectators: inferences about intentional states, which include beliefs, desires, plans, and goals (e.g., Dawn is looking for a place to sit); character traits (e.g., Dawn is an outsider); and emotions (e.g., Dawn is afraid). Loosely construed, these
119
inferences and judgments are all components of the faculty known variously as folk psychology, naïve psychology, theory of mind, and social intelligence, perception, and cognition.4 These different names reflect the philosophical and methodological differences of fields in which they are used, such as philosophy, psychology, and anthropology. It is a basic belief in many fields that human beings, as well (to some extent) as many of our close relations in the animal kingdom, are able to predict and explain the behavior of others with a degree of confidence and success better than chance. Many researchers use the term “mindreading” to describe our ability to attribute mental states to other people. For example, psychologist Nicholas Humphrey writes, “Small wonder human beings have evolved to be such remarkable psychological survivors, when for the last six million years their heavy task has been to read the minds of other human beings.”5 Simon BaronCohen puts it this way: “Now, you and I are mindreaders. I don’t mean that we have any special telepathy…I just mean that we have the capacity to imagine and represent states of mind that we or others might hold.”6 Stephen Pinker writes, “We mortals can’t read people’s minds directly. But we make good guesses from what they say, what we read between the lines, what they show in their faces and eyes, and what best explains their behavior. It is our species’ most remarkable talent.”7 None of these statements argues for psychic powers; as BaronCohen specifies,
120
this is an imaginative task, a matter of interpreting situations and behavior. This widespread usage is metaphorical but not really fanciful; it requires not just the recognition of symbols, as we recognize letters and words, but also the analysis of data. My mindreading might differ in some ways from yours, and the content of folk psychology is variable to an extent, though I will argue that there is often a baseline of inferences and interpretations warranted by the narration of a film, or unambiguously solicited through the rhetoric of narration, such that we expect all spectators to read some aspects of a narrative the same way. Films solicit inference and judgment just as real life does: by presenting us with people facing situations. Through a constellation of processes, combining a reliance on culturallyspecific knowledge with a dependence on universal human capacities, spectators are able to make sense of characters. The focused, constructed nature of film narratives makes possible a systematic exploitation of folk psychology as a principle of character construction. Because of this, in straightforward, canonical narratives we are often able to explain the behavior of characters even more assuredly and successfully than we make sense of our own realities, especially by the end of the story, when we have been given all of the pertinent information. The process of understanding is the same whether we are talking about characters or real people, but the outcome of the process is to an
121
extent fixed in cinema. Thus in considering the interaction of text and spectator, we must keep in mind each one’s contribution to comprehension: the patterning of narrative data in combination with the structures of thought produce the character. In order to account for character comprehension, we need to know about both of these sides of the story, the cognitive side and the textual side. Films in which character is made prominent, such as American independent films, rely on a host of techniques, many of which they share with less charactercentered narratives. Through the selective control of the flow of narrative information, films can maximize our interest in character by creating gaps that frustrate our attempts to read their behavior. Films can create ambiguous situations that demand that the spectator weigh competing explanations for behavior. And films can multiply the traits attributed to a character, which seems inevitably to lead to complexity and contradiction as the multiplication of types does, and makes simple snap judgments about characters much less likely. These possibilities illustrate that what makes a character interesting has as much to do with the temporal process of understanding the character as it does with the data of the character’s representation, a topic I revisit in Chapter 6. There are also cases of films which radically frustrate our attempts to understand character psychology, either by presenting insufficient data about characters or by presenting incomprehensibly
122
contradictory or inconsistent data. David Lynch’s Lost Highway might be the best example from independent cinema of a film with impenetrable characters. As limit cases, such films demand careful attention, but since they are so different from the vast majority of films and struggle to subvert their norms of character and characterization, I have decided to leave radical anticharacter films aside. This chapter identifies two kinds of mindreading. First I will discuss folk psychology in a strict sense of inferences about characters’ knowledge, beliefs, desires, and plans on the basis of observed action and context. Second, I will raise the question of whether characters may be said to have a personality to which spectators may attribute causality for narrative events. I also discuss a third topic (a related feature of social cognition): a process of social perception whereby perceivers evaluate persons they do not know well by making snap judgments about their traits and types. All three of these processes are fundamental to character construction and all are exploited by filmmakers. In the next chapter I will explore the way that facial, bodily, and vocal expressions have a characterizing function, especially as regards emotion. Despite the separate treatment each of these topics will be given, I see these process, in combination with typing, as mutually reinforcing and overlapping strategies for the construction of character. I take all of them, working in concert, to consist of “mindreading” in a loose sense that all
123
involve implicit rather than overt appeals to construct an inner life for characters.
Folk Psychology We are all psychologists. Steven Pinker8 Folk psychology is the everyday psychology of ordinary people. It is an inferential system for predicting and explaining the behavior of others. Based on observed actions and informed by prior knowledge, subjects infer many details of other people’s thoughts, and this is an essential component of narrative comprehension. Indeed, Jerome Bruner has argued that folk psychology is organized as narrative and identified as its constituent features several essential attributes of narratives: “Folk psychology is about human agents doing things on the basis of their beliefs and desires, striving for goals, meeting obstacles which they best or which best them, all of this extended over time.”9 Like most aspects of online cognition, folk psychology is “fast and dirty.” It is clearly a product of natural selection: our ancestors who were able to read minds using folk psychology had a distinct advantage.10 This means that it is not the product of careful thinking, though it is not always right. It is part of the social organization of primates that they are able to make these predictions and explanations with some reliability; this
124
ability has been called “Machiavellian intelligence” because of the possibility it offers for socially talented social subjects to compete with their conspecifics for social rewards.11 Folk psychology is generated in response to observed actions, and constitutes a set of inferences guided by expectations and prior knowledge. Principally, it is made up of desires and beliefs attributed to others.12 On the basis of these desires and beliefs, we generate further inferences about the intentions and goals underlying actions. In terms of film characters, the questions we might ask of them are first, what do they know, and second, what do they want? Sometimes we also have to ask, what are they doing? Following that, we can answer questions such as, why is she doing that or why isn’t she doing x? Sometimes spectators actually formulate these questions and even speak them aloud, but generally these questions are not posed verbally but are tacit. These folk psychology questions have the same status as the questions that readers or spectators pose as they follow a narrative described by Noël Carroll in his discussion of “erotetic narration.” 13 Basically, narratives prompt readers or spectators to see events as raising questions and answering them. For example, in Reservoir Dogs, one question is: Which guy is the cop?
125
Folk psychology is usually discussed in terms of intentional states, not affect states or personality dispositions. Intentionality is a philosophical term used to describe the relation between the mind and the world. Those thoughts that are about something external to the subject, such as beliefs, desires, plans, and goals are called “intentional;” philosophers use an “aboutness” criterion to establish intentionality. If I see Marie go out to the garden with her basket and infer that she is off to pick blueberries, I have attributed beliefs and desires to Marie about the garden, the basket, and the blueberries. The “aboutness” object may be another person: if you observe two people in a fistfight, you infer that each desires to beat the other. 14 For the purposes of this discussion, intentionality is distinguished from affect, though they may be mutually determinative. Marie may be happy about going blueberry picking if she likes to pick blueberries, or she may be angry about it if she is only doing it because she lost a bet to Jean; her emotions are not considered part of the intentionality inference. (As should be clear, intentionality has nothing to do with the intentions of the author or filmmaker.) When you observe Gene running to catch the bus, folk psychology explains that Gene believes that the bus will stop and open its door to let him in, and he desires to board. Perhaps Gene is late for work, perhaps he overslept, perhaps the bus is early. These are all possible inferences.
126
Folk psychology does not explain that Gene is lazy or scatterbrained, or that he is afraid his boss will be upset with him if he is late for work. It does not explain that he is anxious about making the bus or angry at the driver if the bus is early. These inferences are based on other aspects of social cognition. The classic example of folkpsychological mindreading is a test administered to young children to assess their cognitive development called the falsebelief test. The child is shown that Jimmy sees Sally put a candy in her pocket. While Jimmy is away, Sally moves it to her purse. When Jimmy returns, the child is asked, where will he say the candy is? At two years of age, children answer that Jimmy will look for the candy in Sally’s purse. At three or four years, most understand that Jimmy will look in her pocket because he believes it to be there. Passing the test is evidence that a child has acquired folk psychology, because he is able to attribute a false belief to someone else. In other words, he knows that Jimmy thinks the candy is in Sally’s pocket, even though it’s really in her purse. The child has to read Jimmy’s mind in order to pass the test. When older children fail to acquire this skill, this is usually seen as evidence of autism, a disease of severe social impairment.15 One important aspect of folk psychology implied by the falsebelief test is mental embedding (also called nesting or metacognition), i.e., attributing knowledge of another
127
person’s knowledge, and so on.16 For example, you know that I know that my mother knows that my father doesn’t know about his surprise party. In narrative, we are often required to go farther than just attributing beliefs and desires to characters. We also attribute beliefs and desires about other characters’ beliefs and desires, which may themselves be about other characters’ beliefs and desires, and so on. In Hard Eight, the spectator learns that Jimmy (Samuel L. Jackson) knows that John (John C. Reilly) doesn’t know that Sydney (Phillip Baker Hall) killed John’s father. At once, the spectator infers that Sydney does not want John to know that his father’s killer was Sydney (we don’t even know if John knows that his father was murdered), and that Jimmy knows that Sydney’s desire to avoid John finding this out will work to his own advantage. It follows that the significance of folk psychology to cinematic narration is that it communicates information about characters’ knowledge. This is information that can be conveyed through direct means of characterization, but is more typically left implicit; it is much more economical to communicate information about characters’ knowledge through indirect means. Seeing that a character sees or hears something makes the process automatic; so does seeing that she does not see or hear something. This kind of information is a baseline element of characterization, because more sophisticated kinds of
128
information about plans and goals, about characters’ personalities, or about their emotions, depends on what the spectators knows they know or do not know. Knowledge and vision are closely linked in our minds. We often speak of knowledge in visual metaphors such as “an observation” meaning “a thought,” “view” meaning “opinion,” and “to see” meaning “to know.” Cognitive linguists such as George Lakoff and Mark Johnson argue that visual metaphors structure thought such that language itself is a product of the experience of navigating space visually.17 The connection between seeing and knowing is literal when we are talking about observing human behavior: folk psychology often relies on following other people’s eyes. The psychologist Simon BaronCohen has identified three mechanisms which feed into the brain’s theory of mind module: an intentionality detector, which attributes beliefs and desires to others; an eyedirection detector, which recognizes eyes and attributes cognitions to the people whose eyes are looking at things; and a sharedattention mechanism, which allows two people to attend to the same object and to infer that the other sees what he sees.18 Clearly, in order for two people to share their attention, they must have detected each other’s eye directions, so it follows that the eyes are a crucial component of folk psychology. Eye contact is a powerful force that causes physiological arousal all by itself.19 We interpret the gaze to lead back to intentionality: when someone
129
looks at you, you wonder what they are thinking about you. But beyond attributing beliefs and desires on the basis of the direction of the gaze, BaronCohen identifies a “language of the eyes,” offering evidence from the history of Western poetry, as when Shakespeare describes aggression as “murder in my eye.”20 BaronCohen offers the following list of examples of interpretations of eye expressions, which can be observed from pictures of eyes cropped out of their surrounding context: i. Troubled/Unconcerned ii. Caring/Uncaring iii. Certain/Uncertain iv. Reflective/Unreflective v. Serious/Playful vi. Sad/Happy vii. Nearfocus/Farawayfocus viii. Noticing you/not noticing you ix. Dominant/Submissive x. Friendly/Hostile xi. Interested/Disinterested xii. Desire/Hate xiii. Trust/Distrust xiv. Alertness/Fatigue xv. Scheming/Sincere xvi. Surprise/Knowing xvii. Anger/Forgiveness21 Although some of these are expressions of emotion rather than beliefs and desires, some of these are clearly meanings about objects in the subject’s world. I will discuss facial expressions of emotion in Chapter 4, but for this discussion it should be noted that several
130
of these meanings are significant for folk psychology as it functions in character construction. The determination that someone is noticing or not noticing, looking at something near or something far, is alert or not, trusts or distrusts, could all be crucial to narrative comprehension. Many film theorists have made “the gaze” a central concern; considering the way folk psychology functions, this is not surprising. Laura Mulvey’s example of Scotty in Hitchcock’s Vertigo is an excellent case of a character whose eyes tell us an enormous amount about his intentional states, i.e., his beliefs about Madeleine/Judy and his desires in relation to her.22 In following his eyes, we infer many of the meanings in the list above, depending on which scene we are watching: troubled when he sees a woman he believes to be dead, serious when he insists that Judy become Madeleine, noticing the woman to the exclusion of all else, interested in and desiring her, and surprised by several developments of the plot. This is inferred from several redundant components of the characterization, but Jimmy Stewart’s eyes are one powerful means of establishing the character’s interiority. The difference between folk psychology as it functions ordinarily and as a means of understanding cinematic narrative is that the spectator does not interact with the characters in the same way that two or more people interact with each other. In particular,
131
the characters and spectators cannot look at each other in any meaningful sense. No spectator of right mind interprets a look into the camera to be “why is he looking at me?” Spectators are like sideparticipants to a conversation who never become directly engaged in it.23 In a sense, by being free to think only about others, we have more attention to give to making inferences about their intentionality. We may put ourselves in their place without having to worry about our own, as we would in a reallife situation. In addition to characters’ knowledge, the other crucial dimension of folk psychology for narrative construction is its function in relation to character goals and sub goals. The pursuit of clear character goals is often taken to be a hallmark of canonical story form, and is especially important in classical Hollywood narration.24 Jonathan Culler writes that “all récits involve a character seeking something and encountering internal or external help and opposition.”25 In relation to cinematic character, Per Persson discusses the way that goals serve to focus folk psychology inferences and bring coherence to our understanding of narrative situations, and the ways they may be informed by character perceptions and beliefs.26 Overarching character goals are often stated directly and redundantly in dialogue, both by the character pursuing the goal and by others characters. In Meet Me In St. Louis, Esther’s pursuit of a relationship with John
132
Truitt is not something we need to figure out by reading anyone’s mind—she sings a song about it! But goals or intentions also occur on more micro levels, and belief, desire, and intention are typically all intertwined together. Goalinferences feed back into knowledge —pursuit of a particular goal must be predicated on beliefs about it—and into emotions based on appraisal of a situation, as when a character in flight is understood to be afraid of something. We must track many goals, large and small, throughout a narrative. A subgoal of Esther’s overarching goal of being John Truitt’s romantic partner is that she wants to “let him kiss” her at her brother’s goingaway party. (She tells this to her sister.) A subgoal of being kissed is getting him to stay after all of the other guests have gone home. In order to accomplish that, Esther hides John’s hat in the breadbox so that she will have to retrieve it for him later, and we infer this whole string of intentions on the basis of his hat being covered in crumbs. Thus the goal inference is embedded in an existing pattern of knowledge. Demanding such goal inferences of spectators can be a clever comic device of subtlety and economy in the hands of a skilled director such as Minnelli or especially Lubitsch, a master of this technique, but it can also be quite ordinary and obvious when combined with redundant emotion cues such as clear facial expressions of basic emotions such as anger and fear. This device is a staple of action and horror films.
133
In crafting stories, filmmakers use their own expectations as their guide in anticipating how the audience’s folk psychology will make sense of represented actions and events. In essence, they assume that folk psychology is a universal capacity that functions in roughly the same way in everyone capable of social cognition. They tacitly assume, usually correctly, that you and I will respond as they would. This does not mean that there are not failures of representation, in which the filmmakers err in their assumptions, or that there is not variation among spectators, whose diverse experiences and habits of thought create divergent readings of the same narrative. Cultural and historical specificity does not throw a wrench into these theoretical works. Variation among spectators in the way they understand narratives may explain how cultural and historical contexts affect the production of meaning. At the same time, agreement about a great many meanings across these divisions of time, place, and identity points to a universality of narrative, of narrative as a instrument of human social understanding that all of us share. Agreement about inferences basic to comprehension, which are the vast majority of meanings to be found in narratives, far outweighs divergences about matters of interpretation. For example, controversial films such as Do The Right Thing invite interpretation about the thematic significance of characters’ actions and narrative events; only rarely do
134
they invite interpretation about the very content of those actions and events, such as whether or not a given event takes place, as in Last Year at Marienbad. In the case of Do The Right Thing, a primary point of controversy is the motivation underlying Mookie’s act of throwing a trash can through the window of Sal’s pizzeria during the climactic street uprising scene. One character’s motivation for one action is unclear, and its obscurity is part of the point of the film. But the other main events of the narrative, and the motivation of each character underlying these events, are not ambiguous. It is only in a context of overwhelming narrative clarity that a kind of focused ambiguity can function as it does in Do The Right Thing. Spectators may disagree about the significance of Mookie’s action and even about the underlying beliefs, desires, and intentions of various characters, but they must agree on the facts of the narrative or else they would not be able to carry out their debates about the film’s significance—they would have no common frame of reference. The narrative clarity that can be the product of folk psychology allows such debate to take place. What does folk psychology offer filmmakers? Much obviously depends on the degree of knowledge we have about characters and situations. When characters’ intentional states have been established using other means, such as dialogue and facial
135
expressions, folk psychology may play only a minimal role. In general, for cinematic characterization, folk psychology is a kind of default procedure. Absent other more explicit data, folk psychology explains things we would otherwise not understand. Although psychologists might not agree with this crudeness of this distinction, for the sake of keeping this theoretical account focused and straightforward, I am using folk psychology to refer specifically to inferences about intentionality based on observed or described character actions. To clarify how folk psychology functions as a means of characterization, we return to the first scene of Welcome to the Dollhouse. Dawn carries her lunch on a tray and looks for a seat in the school cafeteria. Following eyelinematch cuts, we see several vacant seats being claimed by other kids. No one notices her or offers her a seat. Then she asks to sit at a table at which a sullen girl, Lolita, is eating alone. The scene goes on, but this much will suffice for the sake of this discussion. Note the significance of eye direction for our comprehension of this scene. We see that Dawn is looking for a seat; we see that she sees some spaces; we see that she sees the spaces get filled by others. Note also the intentional state communicated by her eyes: Dawn desires a seat in the cafeteria, preferably at a table where she won’t be alone, and is disappointed when the vacant seats are taken.
136
This short portion introduces several of the film’s central themes: Dawn as an outsider, the school as a site of her suffering, other people’s indifference to her feelings, and more generally the degradation and humiliation she regularly faces. We are introduced to Dawn and these motifs in various ways, including the cinematic style (framing, setting, costume, acting, editing and music). The casting of Heather Matarazzo as Dawn is one of the most significant elements of the overall effect of the character’s humiliation. Matarazzo has the nonprofessionalactor quality of performers in Italian Neorealism and some of the Dogme 95 films, which in her case is a function of her failure to meet conventional standards of female beauty as represented in the mass media. The actress conveys the awkwardness of puberty as a universal experience at the same that she also expresses implicitly that the character’s ostracism is a curse visited especially on her. All suffering is unique, her character seems to say, yet we all can relate to her suffering. One of the most commonly heard responses to the film was that it captured those inbetween years vividly and faithfully (though statistically, few spectators really could have shared the extreme degree of Dawn’s social outsider status). The opening moments of the film establish a clear image of the character and her world. This is accomplished merely by casting, but also via facial expressions, costume and makeup, and through the narrative scenario itself.
137
By “the narrative scenario itself” I mean not only the story as written in the screenplay but as brought to life onscreen using all of the various cinematic means of doing so. The distinctive look of that middle school cafeteria, and of those faces, are part of the narrative scenario of the opening scene of Welcome to the Dollhouse. More specifically, however, a narrative scenario is the combination of a character or characters and a situation (i.e., a setting in time and place, with the necessary salient details made clear, and some knowledge of the relationships among the characters.). folk psychology is activated in the presence of a character, in a given situation, whose intentional states are made relevant to us. The scenario can be described in varying degrees of generality. How important is it to the scenario in Welcome to the Dollhouse that we understand it to be an American middle school rather than some other kind of place? Our response would likely be quite similar if we were unable to ascertain any specific information about the age or nationality of the characters or the details of the setting. In essence, the opening scenario represents a character who doesn’t belong. No one admits her to their group. The desire to be part of a group is fundamental to human social psychology, and no one who is socially able could fail to recognize the pain Dawn suffers at the prospect of being alone
138
when she doesn’t want to be. In this scene, then, folk psychology establishes that Dawn has certain beliefs, desires, and goals. She believes there are seats available at tables where she would not be alone. She would rather have company that be by herself. She desires to fit in, not to be singled out. Children typically want to be part of the group, not excluded from it. In part we figure this out (instantly, really) by activating contextspecific schemas about school cafeterias. These schemas include the various forms of prior knowledge we bring to understanding our daily lives, including information about specific objects, processes, and events. But replace the cafeteria with any setting in which children of Dawn’s age are eating in groups and the results are basically the same. Indeed, replace the cafeteria scenario with a more abstract one: an individual faces several groupclusters of coevals who act with indifference or veiled hostility to her. The results in terms of the audience’s assessment of the character’s psychology are still basically the same. In this scene, then, folk psychology establishes that Dawn has certain beliefs, desires, and goals. She believes there are seats available at tables where she would not be alone. She would rather have company that be by herself. She desires to fit in, not to be singled out. The narrative scenario is one of social exclusion, and this experience is basic to group behavior—and not only in human beings.
139
The pain communicated in Dawn’s expression, the soundtrack, the color, the editing, and the rest of the Solondz’s cinematic narrative devices are also part of the meaning of the scene. Folk psychology is the meaning of the scene abstracted from the specificity of its cinematic realization. It is the activation of inferences on the basis of the narrative scenario before, so to speak, the details are filled in. (In reality there is likely to be no temporal priority to folk psychology as opposed to other elements of narration. Comprehension presumably happens all at once, not in stages.)
Causal Attribution As a pattern of inferencemaking about people (real or fictional), folk psychology is a process of social cognition. Another important process that concerns the way we explain human behavior is the attribution of causality to social events. We seek to identify causes of behavior by attributing traits to others and inferring that they act consistently with their traits. Thus the character is perceived to be the cause of narrative events not only in the behavioral sense, as when Superman causes the world to spin backwards, but in the psychological sense, in terms of Superman’s internal characteristics of courage, determination, and fearlessness. We also attribute causal priority to various situations in
140
which people of any personality type might be expected to respond similarly. For example, we expect bereaved people to grieve regardless of whether they are sensitive, stoical, or aloof. Social psychologists who study these sorts of causal inferences call this body of research “attribution theory,” and it is concerned with the question of the extent to which social perceivers attribute behavior to individuals or to situations. In many narrative theories, the fundamental basis of character is the trait. Some theories of character take the trait to be the very substance of a character, and though I believe that a character is not just a cluster of traits, it is indisputable that people real and fictional are described if not defined by their traits, i.e., that traits explain characters, they are the language through which we understand them. A trait may be physical, such as weight or age, and may be a matter of group membership, such as religion or nationality. But the paradigmatic character trait in literary and film criticism is the personality trait. Social psychologists call personality traits dispositions, which they contrast against situations.27 Dispositions may be as numerous as the words we use to describe people’s personalities: friendly and shy, easygoing and difficult, upbeat and depressive, caring and callous, excitable and laconic, and so on. The significance of dispositions for narrative theory is that they function to define characters and to create expectations about
141
them. Character psychology is created in part by an understanding that he spectator arrives at about what sort of person the character is. Dispositions are not mere descriptions, since they are perceived to be causes of behavior. If something internal to a character causes her to act in certain ways, where could this something be but in her mind? Therefore, attributions of personality traits to characters may be considered a factor in the construction of their psychology. The basic assumption of attribution theory is that social perceivers attribute people’s behavior to various causes, which fall (crudely speaking) into two categories: the personality of the social agent, and the environment within which they act.28 (Although this may be the basic approach social perceivers follow in determining attribution, in reality of course the causes of behavior cannot so crudely be divided into binary categories.) According to the pioneer of the field of attribution theory, Fritz Heider, social perceivers act as “naïve scientists,” treating their worlds as experiments wherein they test their assumptions against the data they collect. Ultimately, their goal is to assess the causes of other people’s behavior.29 In a fictional narrative, the reader or spectator is regularly faced with questions about characters and their actions or lack of actions. One of the most common questions we ask ourselves as we watch a film is “Why is he doing that?” This is especially true in
142
narratives, such as detective stories and horror films, in which important expository information is strategically withheld from the narrative’s audience. It is also especially true of narratives like Go in which chronology is manipulated in such a way that narrative context is suspended. When we first see the two television actors, whom we believe to be undercover police officers (Jay Mohr and Scott Wolf), we ascribe a cause to their behavior (seeking to buy drugs to get high), which we subsequently revise (they are actually helping the police try to bust the dealer). And it is also true in narratives that follow a minimalist expository pattern of offering insufficient information to the audience, refusing to grant them a full picture of the relevant narrative world. In Hard Eight, asking “Why is Syd doing that?” dominates our response to the narrative. Our whole interpretation of the narrative is framed by that question until the point in the film when we learn the answer. This is a matter of Syd’s intentions. The standard framework for attribution theory stands on the notion that social perceivers infer causes for behavior by attributing intentions to a social agent. These inferences depend on the agent knowing that her actions will yield certain outcomes, and also on her being able to bring them about. Given this knowledge and ability, intentions are inferred, which become evidence to the
143
naïve scientist of the agent’s disposition. This is called the theory of “correspondent inferences” because the inference of an intention corresponds to the inference of a disposition.30 You observe a young man give up his seat on the bus; an elderly woman sits down. You infer that he knows his act will help her—that he is giving her his seat. And you infer that he is a helpful and selfless sort of person on the evidence of this action. In other words, the inference that he is trying to help someone and the inference that he is a helpful person correspond. As the attribution theorists Edward Jones and Keith Davis describe it, “Correspondence refers to the extent that the act and the underlying characteristic or attribute are similarly described by the inference.”31 Not all inferences are equal, however, in the social perceiver’s mind. Social scientists have observed that we tend to emphasize dispositional over situational causes when making inferences about other people’s behavior. In the above example, could it be that the young man believes that everyone on the bus expects him to give his seat up? That he is seated nearest to the old woman, and that he is the youngest and most able bodied person there? Could it be, then, that he is just doing what anyone in his situation might do, and that his actions tell us little about how helpful or selfless a person he actually is, especially absent information about what percentage of people in his situation would act the same way? Of course it could, which is why we call our attributions
144
inferences. The fact that it seems to make sense to ascribe his behavior to his personality traits is casual evidence of the fundamental attribution error, the tendency of naïve scientists to err on the side of dispositions at the expense of situations when they make their inferences about other people’s behavior.32 As Richard Gerrig and David Albritton have discussed, this tendency is tailormade for narrative fiction because it predisposes us to see the heroes of stories as causal agents.33 It works especially well with the protagonists of adventure or quest narratives, such as James Bond and his ilk, whose individual exploits are the central feature of their narrative worlds. You would never say of Bond that “anyone in his situation would have done the same.” The fundamental attribution error also works in the favor of storytellers working in less schematic forms than adventure stories. The heroes of American independent cinema are hardly heroes at all. They tend, rather, to be ordinary folks. Even so, they are typically represented as distinctive individuals whose traits correspond closely to their actions. Dawn in Welcome to the Dollhouse isn’t just any ordinary middleschooler; rather, she is the social outcast. The two girls in Ghost World are cynical commentators on everyday life often too busy observing to actually take part. Dazed and Confused portrays a crosssection of a Texas high school in the 1970s, with each character filling a
145
distinct role (bully, stoner, etc.) and none merely blending in. Spike Lee’s films present vividly distinct types who typically (though not always) act consistently with their established character traits. Only a truly experimental narrative would present characters whose actions are all better explained by their situations than their dispositions, because such characters would have limited individuality and thus would lack salience and be insufficiently individuated from the other characters.34 While theoretically this could work, in practice few feature films are so radical that they make all their characters functionally the same. Indeed, storytellers work hard to give each character defining traits that will make him stand out against the others, so that, for example, when one is brainy, the other is brawny. Independent films may avoid the more banal forms of character stereotyping, but they hardly avoid this basic approach to characterization. The fundamental attribution error functions to ensure that spectators will prefer characters’ personality traits to other explanations when determining the causality of events. Spectators and storytellers alike have an interest in seeing characters as the causal agents of a narrative because this squares with the way they interpret events in their world. We can go further in describing the ways in which social perceivers, including film spectators, evaluate causal information. Jones and Davis’s “correspondent inferences” theory distinguished between inferences about behavior and inferences about
146
traits. Social psychologist Harold Kelly goes beyond the correspondent inferences to describe three aspects of covariation between social events and effects.35 Kelly shows that there are different kinds of information that social perceivers can use in assigning causality to social events. Let’s say that Tom the postal carrier meets Phil the pit bull on the street and Tom reacts with fear. The social perceiver asks: is Tom afraid of dogs? There are several further questions to ask to come to a conclusion, each addressing one kind of covariation between the effect (Tom’s fear) and the putative cause (Phil). Distinctiveness refers to whether an effect occurs only in the presence of a given cause. Is Tom afraid of Phil and only Phil, or is he also afraid of Angie the Lab and Champ the Sheltie? Consistency refers to whether the cause and effect always go together. Is Tom always afraid Phil, or it just this time, when Phil barks and growls at him and claws at the fence, that Tom is afraid? Finally, consensus refers to whether everyone reacts the same way to the same event. Is Tom the only one afraid of Phil or is everyone in the neighborhood afraid of him? Clearly, in a case of low distinctiveness, high consistency, and low consensus, we infer that the cause of Tom’s behavior is a fear of dogs (in this case, Tom is always afraid of Phil, but no one else is). If distinctiveness and consistency were low (Tom usually loves dogs, including Phil) and consensus were high
147
(everyone else is afraid of Phil today), we would infer the opposite: that Phil is striking fear in the neighborhood today. In a film narrative it may often be difficult to assess all of the aspects of covariation because we are typically not given enough information to make judgments about distinctiveness, consistency, and consensus. There may not be enough instances of a character’s behavior to judge if they “always” behave the same way, for instance, though expository sequences often give a sense of the characters’ ordinary routines, Anyhow, the character is fictional and to ask whether they “always” behave a certain way may be to mistake a fictional creation for a real person. It may be easier to determine whether everyone else reacts the same way as the character in question (consensus), as narratives often present characters who stand out from the crowd. But we can never say how others in the same situation would have behaved, as the situation itself is fictional and no others would ever find themselves in it. Social scientists can run an experiment to determine how many people out of 100 respond in a given way to a situation, and then determine what the baserate consensus is in a population. But it would be meaningless to ask how many people in Rick’s situation would have let Ilsa board the plane with Laszlo, and then gone off to start a beautiful friendship with Captain Renault fighting in the resistance.
148
We are tempted by the romanticism of Casablanca’s fictional world to say that no one except Rick would have done that, but to go as far as to say that is to entertain the delusion that Rick is a real person who can be judged in comparison to others. There is also a builtin bias against using consensus information which is common to social perceivers (including spectators) and filmmakers alike. This bias underrates the extent to which situations explain social behavior. Researchers have found that social perceivers underutilize consensus information when attributing causality.36 They are more likely to base their judgments on distinctiveness and consistency information, so that if they observe Tom react in fear to Phil on more than one occasion, they will judge Tom to be afraid of dogs without paying any attention to whether or not everyone else in the neighborhood is afraid of Phil. Partly this is because social perceivers may not have ready access to consensus information such as population baserates. Which is statistically more likely: that a person has a general fear of dogs, or an isolated fearful reaction to a menacing pit bull? This is precisely the kind of information that we do not have at our fingertips when witnessing an interaction between a postal carrier and a dog. Yet researchers found that even after giving research subjects baserate data, they underestimated the importance of consensus information when making social judgments.37 This is especially true of people’s estimation of their own behavior. Even
149
when we know that most people would react in p fashion to situation c, we still predict that we will act in q fashion, defying the odds, because we think we know ourselves well enough. Social psychologists have shown that this selfknowledge is often mistaken. This finding is similar to those of Kahneman and Tversky in the field of human judgment, in which they argued that the heuristics on which we commonly rely to made momentby moment decisions are often logically and empirically flawed. We underuse consensus in life, and it follows that in cinematic narratives, the emphasis on individual characters as causal agents makes distinctiveness and consistency information more relevant to the spectator. This goes along with the importance of the fundamental attribution error for understanding characterization. In both cases, information about situations is given less weight than information about dispositions. Situations are seen as secondary explanations, individual characters’ personalities as primary. It makes sense intuitively to say that characterization is based more on traits that are distinctive (specific) and consistent, than on traits that many characters (i.e., people) share. Consensus information is precisely the kind of information that would not be useful for narrative characterization because it works against differentiating personalities from one another.
150
Narrative, as Aristotle wrote, is made up of agents executing actions.38 When we ask why did the character do that, our answer tends to come back to the character, who explains himself. Rick acts as he does because of Rick’s traits x, y, and z, in combination with his beliefs and desires. Narrative situations are presented as obstacles for the character to overcome, rather than as factors explaining and determining actions. Characters are typically portrayed as acting in spite or defiance of their situations than in keeping with them. Maybe everyone else would do the expected thing, but not our hero. Partly this is because characters must be distinctively individuated to be comprehensible as characters, and as such must be marked off from the crowd. Partly this is because characters must be shown to have causal agency to be relevant to the audience. If the character’s actions are ultimately determined by something other than their own will, what is the interest in their stories? The idea of a situation governing human behavior makes it sound as though we are more predictable than we would imagine ourselves to be, and such predictability runs against our preformed ideas of what makes a story good, or for that matter, a story. Narrativity itself may be seen to require characters whose actions are selfdetermined (individually or as a group). Selfdirection means choice or freedom, and since our (modern, Western) conception of the self is such as it must be free to act as
151
it wishes, we prefer fictional persons to live up to this ideal. We might go farther yet and say that it is difficult for us to comprehend a fictional self (or a real one) whose choices are less significant in determining his actions than impersonal forces. Our egos demand a certain kind of story, representing a certain conception of character, and so stories as such are formed in the image of this egoistic conception of human action. (This is not to say that we do not believe that situations have any explanatory power, only that we give individuals priority over them.) This point is well illustrated by stories that present characters whose lives seem governed by a factor beyond their control such as history, chance, coincidence, God, or fate. These impersonal forces are typically portrayed as being stronger than any individual’s free will. Many examples come to mind: in romantic comedies and disaster films, which would seem to have little in common, a foreordained outcome hovers over the plot. The audience senses that the characters’ lives are governed by the destined ending (happy in comedy, tragic in disaster). Yet the significant actions in such films are still typically characterdriven. Even if the situation determines the ultimate outcome of narrative events, the most significant narrative actions still tend to be represented as arising from within the characters who are constrained by their situation, but not defined and determined by it.
152
In Titanic, the love between Jack and Rose displaces the sinking of the ship as the most significant narrative event. One especially charged moment of the ending in terms of the film’s characters comes when Rose casts her diamond into the ocean. When asking why she does that, the answer is in terms of Rose’s disposition. You wouldn’t say that anyone faced with the situation of a sinking ship and a drowned lover would throw the jewel away. Rose does what she does as an expression of her love for Jack. In Cast Away, the hero’s determination to live saves him more than anything else, though it is a chance event that brings him back to civilization. Hollywood demands protagonistcentered plots tailored to the images of its stars, but narrative itself is tailored to a kind of characterization that makes the disposition or personality trait primary. In offHollywood narrative traditions, such as European art cinema and American independent cinema, the character is clearly just as central. The difference is in what kind of characters and situations these different traditions prefer, not in the relation of character to situation. Welcome to the Dollhouse is a good example of a film in which a character is clearly constrained by her situation at the same time that she is shown to be the author of her own fate. As one watches it, the relevant question for this discussion of personality traits as causal explanations is to what extent the main character, Dawn, is a victim of her
153
unfortunate circumstances, and to what extent the causes of her suffering is her dispositions. There are powerful suggestions that lead in both directions. On one hand, Dawn’s ostracism is portrayed as unique in her social milieu. No one but Dawn suffers the indignities and injustices visited on her. On the other, she behaves very similarly to the other children and, indeed, frequently mimics their most hurtful acts. If she is unique in her victim status and in her humiliation, we suspect that it must be because of who she is. In some ways, the film makes clear that Dawn is no different from the rest of the kids at her school. As a running motif, she is often shown treating others in ways she personally finds hurtful. The cheerleaders in the cafeteria call her “lesbo,” and at dinner that night she calls her younger sister Missy “lesbo.” Brandon, the bully who later becomes her friend, mouths “fuck you” at her during detention, and at dinner that night she mouths “fuck you” at Missy. Late in the film she also insults her only friend, Ralphie, repeating the other kids’ taunts of “faggot.” Dawn’s brother, Mark, treats her brusquely, ordering her out of his room. In turn, Dawn tries to treat Missy the same way. Watching the film, every time you start to sympathize with Dawn’s suffering, she turns and tries to inflict the same treatment on Missy or Ralphie, the only two characters who are smaller and weaker than she is. By showing that everyone is doing it, it makes it seem
154
that the social situation is a better explanation for the characters’ behavior than their individual dispositions. Namecalling, taunting, insulting, and teasing are behaviors that are normal for middle school students. If you or I were to be transported back to middle school, it is not unlikely that we would take part in one of these unpleasant activities. They are certainly more likely to be found among adolescents than adults. However, the situations in which Dawn teases and bullies others are different from those in which she is teased and bullied. She never does any of these things in school. Furthermore, despite this behavior, Dawn is still more of a victim than any other character. Only Dawn has her locker vandalized and only she is given an insulting nickname, “Weiner dog” that even the other social outcasts use against her. She is the only one taunted by the cheerleaders and the only one victimized by spitballs during the class assembly. She is also portrayed as uniquely hapless. She is given a detention when Brandon tries to cheat by looking at her test, and then gets a Dminus because, she says, of being upset by the incident. She asks the teacher if she can retake the test, and for that is assigned as punishment a 100word essay on the subject of dignity, which in a scene of painful indignity she is forced to recite before the class. She is showered with spitballs during the assembly, but the one spitball she fires back is the only errant one, finding its
155
way into a teacher’s eye. When the family watches the video of the parents’ anniversary party, everyone is impressed by Mark’s musical performance (which is actually laughable) and at Missy’s cuteness (which we might take as cloying), but they all laugh at the one image of Dawn: a bit of cruel physical comedy in which she is shoved into a wading pool and splashes in on her backside. When Missy is kidnapped everyone is made sick with worry but when Dawn disappears for a night to search for her sister no one seems to miss her or care because Missy has been found. And finally in the last scene of the film, with Dawn on the bus with her singing group en route to Disney World, they all seem delighted and excited as they sing their theme song, while Dawn sings along halfheartedly. She hardly wants to go on the trip at all. In unraveling what we perceive to be the causes of Dawn’s behavior, we must weigh several factors. Does she behave as she does because she is a social outcast, or does the causation work the other way around? That is, is her social status the cause or effect of her behavior? Can her actions be explained with reference to a disposition or personality—or at least to a number of traits that partially comprise a personality—that the spectator infers on the basis of the narrative events? I think the answer is clearly Yes. The narrative events function partly to establish, via inference, aspects of the character’s personality that are relevant to the scenario. Following patterns of realworld social
156
cognition, consumers of narrative are always seeking to infer personality information on the basis of narrative events. So is personality is a cause of events or are events the cause of a personality? While this may be a fascinating question, ultimately it need not be answered, as either way a personalityevents covariation is established and a personality is assumed. Narratives always offer both events and personalityinferences, and the “which came first?” question need not be answered definitively. On the basis of the events, we infer the personality, which is then seen as the cause of the events. Yet the personality itself can always been seen as a cumulative effect of many events, and the causal issue may in the end be a mise en abyme. Dawn has a strong desire to be popular, to fit in, to be liked, to have friends, and to have a romantic relationship with a boy (in particular with her brother’s bandmate, Steve, and possibly also with the bully, Brandon). But she lacks selfknowledge in certain important respects. She fails to see that her “special people’s club” is pathetic and that she would not make a suitable romantic partner for the much older and more mature Steve. When she tries to dress like the promiscuous Ginger on the day of the party, to try to attract Steve, she seems to have debased herself in a way that she doesn’t recognize at all. She also doesn’t seem to recognize that her battles against her parents are doomed to fail, as when she opposes having her backyard clubhouse removed for the party and when
157
she refuses to tell her sister she loves her. In the first case she is deprived of dessert, forced to watch as her siblings share her piece of chocolate cake, and in the latter she is forced to remain at the dinner table hours after everyone else has left. Facing these punishments seems to strengthen her resolve to defy her parents’ authority, while as spectators we wish that she would just capitulate to make her own life easier. As one who suffers constantly, she often fails to take opportunities to ease her own suffering. So to sum up her dispositional traits, Dawn is friendly and outgoing while socially awkward, selfdeluded and stubborn. Add to this her situational traits, the exterior factors that define her character: she is an outcast in her community, pickedon and bullied and almost without friends, and her parents prefer her siblings to her. Clearly, these dispositional and situational traits are mutually reinforcing. Being socially awkward and being a social outcast obviously feed into each other. Dawn’s lack of selfknowledge and selfawareness also feed into her social status and vice versa. At the same time, however, some of her personality traits seem to defy her situation. Why is she hopeful of having a romance with the older and more mature Steve when she has never had a romantic relationship with any boy, when the boys in her school either ignore or taunt and tease her? Why isn’t she happy to be part of a class trip to Disney World that gets her away from her family and her school environment? On one level, Dawn’s lack of
158
selfawareness is a facet of her social outsider identity—if she were more aware, she would be able to fit in better. On another level, this trait is a hindrance to her ever escaping herself, which would be her only possibility for change. Finally, to return to a question raised at the beginning of this section, when attributing a social event to the personality of a social agent (character), are we constructing a mental state for them? Is it the same thing as when we infer knowledge, beliefs, desires, and the like? In some respects, it is not the same thing. There is a difference in kind between an inference that Jack is afraid of snakes and an inference that Jack is a fearful sort of person. Yet the question of causality that arose above, the which camefirst question, demonstrates that both inferences are necessary to the construction of the character because each ends up explaining the other. It is in part by having constructed a profile of the character’s dispositions, through causal attribution of their behavior, that we are able to infer their psychological states at all. In some ways, the constructed personality precedes the inferences. This makes for yet another way in which the narrative action or scenario primes our construction of psychological states and has a characterizing function. Every narrative action is subject to our causal interpretation, and every such interpretation is biased in favor of character traits as explanations. Thus every action we observe in a narrative can be read as another bit of evidence in our ongoing
159
project of understanding characters’ inner lives.
160
Heuristics of Social Judgment
161
A final area of social cognition research of relevance to this discussion deals with patterns of inferences people make in reasoning about matters of social judgment. Daniel Kahneman and Amos Tversky’s influential research found that inferences we make about ourselves and others are predictably and systematically flawed.39 We have seen that in causal attribution our inferences are systematically flawed in our overemphasis on dispositions; Kahneman and Tversky’s research adds other ways in which social perceivers systematically err. Their research does not demonstrate that on the whole we do badly at making sense of other people, but that we don’t do as well as we might if our intuition were better at making predictions based on accurate statistical data and logical thought. Matters of social judgment that involve assessments of people we do not know well, such as evaluating candidates for a job, require fast social reasoning using shortcuts or rules of thumb to guide us. Kahneman and Tversky call these patterns of thought heuristics, and despite their biased, errorprone nature, they function remarkably well as shortcuts, by freeing our minds for other tasks. Filmmakers use their intuitive sense of heuristics in crafting narratives that will appeal to social perceivers already skilled at just these patterns of thought. That is, systematic errors in judgment are built into stories to allow for straightforward comprehension and a sense of familiarity of dramatic material.
162
One in particular that is relevant to understanding narrative characterization is called representativeness.40 The representativeness heuristic allows us to produce an estimate of probability about another person. It allows us to categorize a person by making inferences based on traits we know them to have. If those traits seem to be particularly representative of a given category, such as their occupation, we may be more likely to predict that they belong to that occupation than would be warranted according to a statistical sampling of the population. In other words, representativeness allows us to make “person A is a member of category B” predictions even when categories C, D, or E are the empirically more likely choices for A. We might, for example, believe that a tall, muscular, 30 year old man is more likely to be a professional athlete than a psychiatrist, but statistically there are so many more psychiatrists than pro athletes that our prediction would likely be wrong. We would make this prediction based on the man’s appearance being representative of what we think a professional athlete’s should be. Furthermore, given the choice between categorizing him as a psychiatrist or a psychiatrist who was once an Olympic swimmer, we might guess the latter even though the probability of one proposition (psychiatrist) being true is always greater than the
163
probability of two independent propositions (psychiatrist and Olympian) both being true. That is, the probability of someone being a member of category A is always greater than the probability of being a member of categories A and B, yet social perceivers are often drawn to the latter judgment because it seems to fit their ideal notion, or stereotype, of those categories more accurately. In this way, social perceivers regularly err in their social judgments. Representativeness, like other heuristics, functions as a cognitive shortcut, producing quick assessments of others. In reality, we are able to correct our initial judgments by further exposure to data about people. The reason we are able to function successfully in social contexts is that our heuristics are only the beginning of our encounters. We need them to make quick assessments of our surroundings because we are unable to generate scientifically accurate data about the plethora of phenomena we observe and with which we interact. It is for this reason that some cognitive scientists take the position that social perceivers are “cognitive misers,” seeking to negotiate their surroundings with a minimum of conscious effort.41 In presenting narrative scenarios, filmmakers can harness our reliance on heuristics such as representativeness for expository purposes, recognizing implicitly that film viewers are cognitive misers too. They want to understand narrative scenarios without struggling too much to asses who
164
the characters are and what their environment consists of. Narrative form can be viewed as a process of raising certain expectations in an audience, and then elaborating those expectations, revising them, challenging or undermining them, and ultimately confirming some—often many—of them. Representativeness is a robust means of creating expectations about characters whom we do not yet know well or as well as we would like to. Essentially, viewers make tacit predictions about the likelihood of character P belonging to categories Q, R, or S, where these categories could be any relevant designation applicable to a person, such as a personality type or disposition (e.g., affable), group membership (e.g., devout Christian), or social or economic class (e.g., yuppie). Presented with that tall, muscular, 30 yearold man at the beginning of a film, we would follow the same basic patterns of thought as we would in reality, expecting him to conform to groups he would seem to be representative of and making our inferences about him accordingly. It might seem that such inferences are most likely to be activated early in a film, however, exposition is often an ongoing process in cinematic narratives, being in various ways delayed and distributed throughout the length of a feature film.42 There are significant differences, of course, between reality and narrative. One of these is that the “statistical likelihood” of a prediction being true cannot be determined in
165
a work of fiction in any meaningful way, as it can in the real world. Perhaps one criterion of realism is the degree to which the statistical likelihood of events is represented (implicitly of course) to be the same as those of an historically and geographically specific reality. In a narrative set in Paris in 1789, any statistical probabilities applying to the characters are those that applied to people in Paris in 1789. The characters would be just as likely to be loyal to the king as real Parisians were. Yet even in the case of a narrative that is realistic by these terms, we are talking about a fictional statistical likelihood—to which the authors of the narrative are not bound. Historical fictions abound in which events with a statistical probability of zero occur, such as those mystery novels in which famous personages solve crimes, or those sequences in Forrest Gump and many other narratives in which fictional and nonfictional characters meet. Even realism has its share of unlikely occurrences and atypical characters, just as reality does. As Aristotle writes in chapter 9 of the Poetics, narratives are about possible events being made to seem probable, not the other way around. He contrasts what he calls poetry (what we would call narrative fiction) to history, stating that history is bound by the facts of what happened, whereas poetry is made of what could happen.43 Heuristics such as representativeness are our usual means of determining what seems probable. So in
166
fictional narratives such as feature films, spectators are primed to respond using representativeness as a guiding heuristic. The data of the narrative are organized specifically to appeal to the audience’s unexamined notions of what is and is not likely. To take just one brief example, the first half of Hard Eight unfolds using delayed exposition (i.e., withholding important information from the audience). Sydney (Philip Baker Hall) meets John (John C. Reilly) sitting outside a roadside diner and offers him a cup of coffee. We learn that John is in need of money and Sydney offers help. In subsequent scenes, we see how Sydney helps John get a complimentary room in a Las Vegas hotel by making the casino think he is a big gambler. Then after a fade out, the plot jumps ahead two years. Throughout this first portion of the film, Sydney is a mysterious character. We know nothing of his backstory until later in the film. Part of our engagement with the film consists of trying to figure Syd out. What is his interest in John? How does he know so much about Las Vegas? What does he do for a living? Who is his family? Is his help motivated by altruism or is there something he hopes to benefit from John’s friendship? Above all, is Sydney being good to John or is he taking advantage of him? These are matters of characterization, ultimately leading to an assessment of what kind of person
167
Sydney is. Most characterizations are much more forthcoming, as we are given enough information about the character within the first few scenes of their appearance in the plot to not have questions about the basic details of identity. But Syd is the proverbial “mysterious figure.” By a combination of his late middle age, dressy attire, brusque speaking style, stoical facial expressions, and assured actions and words, Sydney is characterized to appeal to the audience as a someone purposeful and determined, a character with a clear sense of his own goals and desires. He is representative of a Las Vegas player, an old hand at gaining every available advantage. His characterization contrasts sharply with the helpless waywardness of John, setting up an archetypal mentorprotégé/fatherson relationship. The most significant question to the audience regards Syd’s motivation, his intentions towards John. We must determine what is or is not likely to be true of each one of them on the basis of fragmentary evidence. Using Kahneman and Tversky as a guide, we would say that we infer details of their characters on the basis of what kind of social type each one of them seems best to represent. This process is the same as the one described in reference to the opening of Welcome to the Dollhouse. We make snap judgments about characters, slotting them into categories and assessing the effects of their situations on their psychology. In doing so,
168
we generate fairly elaborate mental dossiers for the characters, inferring much information about their intentional states, including blanks that need filling in, such as motivation, or a constellation of possibilities, as with Syd in Hard Eight who may be acting in John’s interest and may not, who may be honest and may not, and so on. Folk psychology, dispositional attribution, and the representativeness heuristic all give us guidance in making our inferences, all bring preexisting structures of knowledge about goaldriven behavior, charactercentered causality, and typicality to the task of making sense of characters.
Conclusion: Mindreading in American Independent Cinema Implicit in my discussions of Hard Eight and Welcome to the Dollhouse in this chapter have been ideas about some general tendencies of American independent cinema that I would like now to bring to the foreground. In comparison with those of mainstream cinema, the characters of independent films often seem psychologically richer, more multidimensional, deeper, more lifelike, or more surprising. As I discussed in Chapter 2, this is partly a product of how these films handle typing. It is also a product of how they structure the inferential patterns activated through the cognitive processes described
169
in this chapter. Hard Eight and Welcome to the Dollhouse are both typical in their adherence to the viewing strategies described in Chapter 1. Both films are rooted in a strong sense of place, both dramatize the experiences of social outsiders, and both have a modesty of scale and ambition and an emphasis on character that might be seen as the antidote to the excesses of the mainstream commercial cinema. Furthermore, both films activate a rich pattern of inferential activity concerning the interiority of their central characters, Syd and John in Hard Eight and Dawn in Welcome to the Dollhouse. Neither film has significant character narration, subjective sequences, voiceover, or other devices that give more direct access to character—they rely on the dramatization of action, on the narrative scenario for their effects of psychological richness. They achieve these effects by exploiting basic principles of social cognition to both frustrate and maximize the spectator’s inferential interaction with the text. Hard Eight does this by providing insufficient narrative information to make mindreading efficient. In Hard Eight, the effect is that our inferential activity is increased with the active questioning of characters’ beliefs and desires. Syd’s goal of helping John is clear enough, but it is equally clear that it is itself the subgoal of an overarching goal that is
170
suppressed until late in the film. This complicates the process of understanding Sydney. Welcome to the Dollhouse maximizes inferential activity chiefly by complicating the attribution process, making the designation of causes and effects difficult and forcing us to speculate about Dawn’s role in her own situation. In both instances, the spectator must actively search out meanings, whereas a more conventional presentation of character might leave less room for judgment and interpretation. In both cases, representativeness is exploited during the initial exposition and cues some clear judgments about minor characters, such as the catty popular girl who doesn’t invite Dawn to her party. In theory representativeness could force us to jump to conclusions that we might later disconfirm given the extensive data the films eventually introduce about main characters, and indeed Dawn and Syd do turn out to be more multidimensional than their initial characterization might lead us to expect. More mainstream films might rely more straightforwardly on the efficiency of representativeness as a cognitive heuristic. But by making Sydney and Dawn more highly individuated and multidimensional, Anderson and Solondz make them the sort of characters who resist being so easily categorized using a mental shortcut. By the end of each film, the character seems much richer than our initial traittype inferences would suggest. The two films have many differences, the most significant of which is a matter of
171
exposition: Hard Eight depends on delayed exposition regarding Sydney’s relationship to John, setting up a gap, whereas Welcome to the Dollhouse is straightforward and linear, with few events of significance occurring before the first scene of the film. But they have in common that they maximize inferential activities of various sorts and close off the option of easy heuristics to make sense of their main characters. By considering how these socialpsychological processes function ordinarily, we can see how filmmakers may present characters who place extra demands on us to work through them. When critics admire independent films for creating characters of depth and complexity, a topic to be revisited in Chapter 6, they are attesting to the powerful effects of mindreading and the ability of filmmakers to manipulate this cognitive function for aesthetic ends.
1
Syd Field, The Screenwriter’s Workbook (New York: Dell, 1984), 55.
2
Carroll, A Philosophy of Mass Art , 291359; Murray Smith, Engaging Characters, 187227.
3
For a critique of the equation of convention with arbitrariness, see Bordwell, “Convention, Construction, and Cinematic Vision.” 4
On theory of mind, see Daniel C. Dennett, The Intentional Stance (Cambridge: MIT P, 1995); Stephen Stich, From Folk Psychology to Cognitive Science: The Case Against Belief (Cambridge: MIT P, 1983); Janet W. Astington, Paul L. Harris, and David R. Olson (eds.), Developing Theories of Mind (Cambridge: Cambridge UP, 1988); Andrew Whiten (ed.), Natural Theories of Mind: Evolution, Development and Simulation of Everyday Mindreading (Oxford: Blackwell, 1991); Nicholas Humphries, The Inner Eye: Social Intelligence in Evolution (Oxford: Oxford UP, 1986); Simon Baron Cohen, Mindblindness: An Essay on Autism and Theory of Mind (Cambridge: MIT P, 1995); Richard W. Byrne and Andrew Whiten, Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes and Humans (Oxford: Oxford UP, 1988). For an application to literary narrative, see Zunshine. For an application to cinematic narrative, see Persson. 5
Humphries, 50.
6
BaronCohen, 2.
7
Pinker, 330.
8
Ibid, 329.
9
Bruner, Acts of Meaning, 423.
10
Byrne and Whiten; Pinker; Barrett, Dunbar and Lycett.
11
Byrne and Whiten.
12
A discussion of how children acquire this capacity is Henry M Wellman, “From Beliefs to Desires: Acquisition of a Theory of Mind” in Whiten, 1938. 13
Noël Carroll, The Philosophy of Horror or Paradoxes of the Heart (New York: Routledge, 1990), 130136. 14
Dennett; John Searle, Intentionality: An Essay in the Philosophy of Mind (Cambridge: Cambridge UP, 1983).
15
The falsebelief test is described in Pinker, 330331.
16
Andrew Whiten and Josef Perner, “Fundamental Issues in the Multidisciplinary Study of Mindreading,” in Whiten, 117. George Lakoff and Mark Johnson, Philosophy in the Flesh: The Embodied Mind and its Challenge to Western Thought (New York: Basic Books, 1999), 238240. 17
18
BaronCohen, 3158.
19
Ibid, 98.
20
William Shakespeare, As You Like It (Act III, Scene V).
21
Ibid, 114.
Laura Mulvey, “Visual Pleasure and Narrative Cinema” in Philip Rosen (ed.) Narrative Apparatus Ideology: A Film Theory Reader (New York: Columbia, 1986), 198209. Of course, I mean “desire” in its psychological, rather than psychoanalytic, sense. 22
23
This idea of viewing as sideparticipants is from Richard A. Gerrig and Deborah A. Prentice, “Notes on Audience Response” in David Bordwell and Noël Carroll (eds.), PostTheory: Reconstructing Film Studies (Madison: U of Wisconsin P, 1996), 388403. 24
CHC.
25
Jonathan Culler, Structural Poetics (Ithaca: Cornell UP, 1975), 233.
26
Persson, 164169;186201.
27
Lee Ross and Richard Nisbett, The Person and the Situation: Perspectives of Social Psychology (Philadelphia: Temple UP, 1991). 28
Ibid, 220. For an application of attribution theory to literary criticism, see Jonathan Culpepper, “Inferring character from texts: Attribution theory and foregrounding theory” Poetics 23 (1996), 335 361. 29
Fritz Heider, The Psychology of Interpersonal Relations (New York: John Wiley, 1958). For overviews of attribution theory see Fiske and Taylor, 2045; Perry R. Hinton, The Psychology of
Interpersonal Perception (London: Routledge, 1993), 138155; Michael Roth and Garth J. O. Fletcher, “Attribution and Social Perception” in Gardner Lindzey and Elliot Aronson (eds.), Handbook of Social Psychology Vol. 2, 3rd ed. (New York: Random House, 1985), 73122. 30
Edward E. Jones and Keith E. Davis, “From Acts to Dispositions: The Attribution Process in Person Perception” in Leonard Berkowitz (ed.), Advances in Experimental Social Psychology Vol. 2 (New York: Academic P, 1965), 220266. 31
Ibid, 223.
32
Ross and Nisbett, 125139.
33
Gerrig and Allbritton, 380391.
34
This may be the case in Soviet Montage films of the 1920s, in which social and historical forces rather than psychologized characters are the main causes of narrative events. See Kristin Thompson and David Bordwell, Film History: An Introduction 2nd ed. (Boston: McGrawHill, 2003), 119143. Harold H. Kelley, “Attribution Theory in Social Psychology” in David Levine (ed.), Nebraska Symposium on Motivation (Lincoln: U of Nebraska P, 1967), 192238. See also Fiske and Taylor, 29 35, and Kunda, 124137. 35
36
Kunda, 63.
37
Ibid, 64.
38
Aristotle, Poetics, trans. W. Rhys Roberts (New York: Modern Library, 1954), 224.
39
Daniel Kahneman and Amos Tversky, “On the Psychology of Prediction” Psychological Review 80 (1973), 237251; Amos Tversky and Daniel Kahneman, “Judgment Under Uncertainty: Heuristics and Biases” Science 185 (1974), 11241131; Daniel Kahneman, Paul Slovic and Amos Tversky (eds.), Judgment Under Uncertainty: Heuristics and Biases (New York: Cambridge UP, 1982. See also Richard Nisbett and Lee Ross, Human Inference: Strategies and Shorcomings of Social Judgment (Englewood Cliffs, NJ: PrenticeHall, 1980); Kunda, 53110. 40
The representativeness heuristic originally discussed by Kahneman and Tversky is explained in the context of more recent research in Fiske and Taylor, 268270 and Kunda, 5662. 41
Fiske and Taylor, 1112.
42
Bordwell, 56.
43
Aristotle, 236237.