Common Sense + Film Theory = Common-Sense Film Theory?

May 2011

Start with this question, which I think is one of the most fascinating we can ask: What enables us to understand films?

All films? Well, set aside some hard cases, like Brakhage abstractions and transmissions of the Crab Nebula from the Hubble telescope (above). Let’s start with a prototype: a film whose moving images present more or less recognizable persons, places, and things caught up in what we intuitively call stories. In other words, an ordinary movie shown in theatres and on video.

Catching a code

The Naked City.

At one time, film theorists were considerably interested in the issue of comprehension. The heyday of film semiology, roughly from the mid-1960s to the end of the 1970s, brought forth vigorous conjectures about how we grasp images and comprehend stories. One of the boldest proposals was the idea that understanding rests upon codes—rule-governed relations between the signifier (a material thing, like an image) and a signified (a concept). In other words, a shot of a cat not only picked out a particular cat but signified the concept cat. Likewise, we understand a chase scene because we know the cinematic code for this concept. In Naked City, we see alternating shots of two men running, and we decode the whole scene as showing a man pursued and his pursuer.

Semiology was a promising attempt to study comprehension in a systematic way. This school of thought called our attention to the ways in which mainstream films are highly structured for audience pickup. Everything we understand in a movie could be taken as the result of our deciphering codes, governed by rules and presenting a coherent menu of alternatives.

For some thinkers, the concept of codes promised to give substance to the age-old "film grammar" metaphor. Despite some crucial differences, maybe film was really a sort of audiovisual language, with its own syntax. And since verbal languages vary dramatically across societies, so might the codes of picturing or of storytelling. Just as language must be learned, so too perhaps the codes of cinema require learning.

Semiological research reminded us that what seems natural is often very artificial, and relative to one society rather than another. In another culture, the code of traffic signals might employ not red, yellow, and green lights, but any other colors. The notion of codes also suited an emerging view of what one influential book of the time called the "social construction of reality." Would people from cultures without cinema or television be able to recognize the blobs on the screen as people and settings? Do codes go all the way down to the very core of our perception? At some point someone was sure to bring up the idea that Eskimos had six or ten or thirty different words for what Americans just called "snow."

Today, classic semiologists are rare in film studies. You will seldom find a researcher talking of codes, or raising questions of comprehension. Nevertheless the idea that filmic expression is quite arbitrary, socially constructed, and learned remains in the ether. Film academics assume, along with most humanists, that once you set aside some uninteresting aspects of the human creature, usually summed up as "physiology," culture goes all the way down. Beyond cell division and digestion, let’s say, everything is cultural, and to invoke any other explanations risks rejection.

That 80s show

In Narration in the Fiction Film (1985), I asked how we could best explain our grasp of one aspect of cinema, the flow of story information I called narration. I argued that since most narrative films were made in order to be experienced by viewers, we ought to study the strategies filmmakers used to elicit understanding. Most of those strategies, it seemed to me, exploited rather general human perceptual and cognitive capacities.

Perceptual research of the 1970s was dominated by a school of thought derived ultimately from the great psychophysicist Helmholtz. "New Look" perceptual psychologists like Jerome Bruner and Richard Gregory held that the stimuli hitting our sense organs were noisy, incomplete, and ambiguous; we needed higher-level faculties to sort them out. Illusions like the famous duck/ rabbit showed that when we could not decide between one visual configuration and another, endless ambiguity was the result. The eye, as was commonly said, was part of the mind. Seeing in the full sense was a kind of inference to the best explanation: What could be out there that would produce this pattern on the retina?

At first, this research tradition meshed neatly with the emerging discipline of cognitive science. In the early 1980s, cognitive scientists were largely focused on matters of language, reasoning, applying categories, and making decisions about action. As with New Look thinking, cognitive science saw mental activity as a quasi-Kantian interplay of input stimuli and conceptual structures, sometimes called schemas, that made sense of the data. Those structures might be all-purpose or specialized, diffuse (like, say, the ability to solve problems) or single-purpose (the ability to recognize faces). Again, inference was the model, although some mental inferences, like those involved in vision, were held to be fast, automatic, and "informationally encapsulated" (i.e., ignorant of anything outside their dedicated domain). Eventually, the inferential approach would become the basis of a computational approach to both perception and cognition, and it probably remains the dominant view in psychological research.

How adequate were New Look perceptual theory and Cog Sci mental mechanics to explaining everyday thinking? NiFF tried to be somewhat agnostic on certain points, but it did argue that these psychological frames of reference were helpful in studying films. Perceptually, films are illusions, not reality; cognitively, they are not the blooming, buzzing confusion of life but rather simplified ensembles of elements, designed to be understood. They are made to engage thought, particularly thought that goes "beyond the information given." Film narratives, like narratives in all media, abstract and streamline their real-world components for smooth pickup and invite us to fill in what is left unshown and unsaid. What outline drawings are to the eye, narratives are to the mind.

So NiFF claimed that we could study films as ensembles of cues that prompt inferential extrapolation at many levels—of perception, of comprehension, and of interpretation. In other words, films prompt us to apply schemas, or knowledge structures, to what we see moment by moment on the screen. Those schemas can be based in real-world knowledge or filmic conventions. Each type posed problems for the concept of codes.

Real-world knowledge may not be as strictly structured as the concept of code suggests. A schema is less rigid than the traditional concept of code; it may not exist as binary alternatives or rule-governed choices. Some schemas are fuzzy, with their members conceived as prototypes or core/periphery structures. So for us a robin is a prototypical bird, a penguin or ostrich is not. The latter might be prototypes for people in other cultures, but that doesn’t invalidate the point that some categories are organized by "best-instance" criteria rather than hard and fast boundaries.

Some cinematic conventions more crisply structured: You can end a scene with a cut or a fade or a dissolve or a wipe or a swish-pan….and that’s about it. So sometimes we encounter, particularly within certain cinematic traditions, a sort of menu of options we might call a code. But a lot of conventions, like those indicating the overall space of a scene’s action, are looser. There is no rule that requires a long-shot to be followed by a close-up, the way a preposition in language requires an object. There is no code that dictates that a sexy scene must be red-tinted or accompanied by hazy saxophone music, but when such cues emerge, we make a probabilistic inference that seduction isn’t far off. Not all conventions, it seems, are coded. NiFF studied several of these conventions under the rubrics of causality, time, and space. Those three categories, NiFF claimed, are basic to narrative and to human cognition, and so they ought to play roles in the process by which we understand stories.

Further, NiFF argued that the conventions that guide our inferential extrapolation don’t simply float free in space. There were recurring clusters of favored choices for presenting causality, time, and space. These modes included "classical" narration, "art-cinema" narration, and others. The historical layout still seems valid to me, and they seem to have proven useful to other researchers.

Theoretically, however, NiFF ran into problems in the role it assigned to inference. At the time of writing NiFF, I was aware of the writings of J. J. Gibson and his insistence that perception evolved in environments very different from the impoverished information that New Look theorists assumed triggered perception. In the three-dimensional world in which creatures like us live, the stimuli are not typically partial or degraded; they are in fact quite rich, even redundant. Moving through space, we register an optic flow that specifies the layout of surfaces quite precisely.

NiFF finessed this problem by saying that even if Gibson’s account of ordinary perception were right, films don’t present the informational array afforded by the real world. Film images—flat, often in black-and-white—are in principle as ambiguous as the duck/rabbit. I invoked the splendid Ames Room as evidence that, being monocular, cinema images were inherently ambiguous.

This now seems to me misguided. Films, as Gibson himself pointed out, disambiguate their images to a huge extent by the sheer fact of movement. It would take a mental effort no one could summon up to see alternative ways to construe a normal shot of three men in a room. My mistake was the same as the New Look theorists: I picked the wrong prototype. Just as illusions aren’t fair samples of perception in the wild, so the Ames Room is an extraordinary piece of filmmaking artifice, not a typical one. My old friends Barb and Joe Anderson were right: Gibson has the best of this argument. (That still won’t settle whether the inferential/computational approach or the ecological approach is the better explanation of natural vision. On that I retreat to the amateur’s agnosticism.)

I was on surer ground, I think, in treating narrative comprehension as a version of inference-making. But NiFF pushed it in a problematic direction. Considering narrative comprehension as inferential led me to bring in the Russian Formalist distinction between fabula and syuzhet. These two terms have been used in several ways, but the most plausible way, it seemed to me then and seems still, is to see fabula as the chronological-causal string of events that may be presented by the syuzhet, the configuration of events in the narrative text as we have it.

Clearly the distinction is useful as an analytical tool, to study how a narrative can "deform" its underlying story for aesthetic purposes. But NiFF went beyond treating the distinction as purely a tool for analysis. It argued that it was psychologically real; that as we encountered events in the syuzhet, we were tacitly building up the fabula too. The process is a bit like double-entry bookkeeping, with the viewer keeping track not only of what is happening each moment on the screen but also slotting that into the chronological pattern of fabula events. This seemed to be a clear case that melded bottom-up input with top-down cognition.

Unfortunately, some people argued, it’s psychologically implausible. Eventually I had to agree. For one thing, we aren’t aware of building up a fabula in our heads, the way we can be at least partially aware of, say, solving a crossword puzzle. For another, we can’t access it easily; try stopping a film on video and reciting the entire chain of events leading up to the moment of pause. Worse, try at the end of the movie to grasp mentally the entire fabula you’ve purportedly worked out. Chances are you can’t do it. Given that our memories are reconstructive rather than photographic, creating an accurate fabula is extremely difficult. More theoretically, Julian Hochberg and Virginia Brooks proposed some reasons that the viewer’s mental representation for the most part cannot reflect the underlying structure of the film.

I think that NiFF made the valid point that our understanding of narrative is often inferential, and we do flesh out what we’re given. But I now think that the inference-making takes place in a very narrow window of time, and it leaves few tangible traces. What is built up in our memory as we move through a film is something more approximate, more idiosyncratic, more distorted by strong moments, and more subject to error than the fabula that the analyst can draw up. Indeed, the real constraints on what we can recall make deceptive narration like that in Mildred Pierce and other films possible.

Still, I think the error was a productive one. In assigning to the spectator the task of ongoing fabula construction, NiFF harmonized with one premise I consider central: a holistic sense of form. Even if we scan the entire narrative through a narrow slit, it’s important for the analyst and theorist to consider the overall design of the work, the more or less coherent principles that govern the unfolding tale. I’m thinking of such matters as smoothly cascading character goals, psychological motives and personality change, gradual development of knowledge, shifts in viewpoint, repeated and varied motifs, and finer-grained patterns of visual and sonic presentation. In an analysis of Jerry Maguire, for instance, I tried to show how such features were operating at many scales, creating a considerable formal richness.

Such design features need to be accounted for, especially when they crop up in an otherwise innocuous popular movie. Why are many movies more tightly organized than they need to be, given the drastic limits on viewer attention and memory? Clearly, goals, motifs, and the rest aim to shape the spectator’s experience in some respect, and we may well register many of them at some level of awareness. NiFF posited a too-sapient viewer, but methodologically at least, it’s better to point up many things that a spectator could respond to, even if no real spectator grasps or recalls all of them. Indeed, some narrative traditions seem to try to pack things tightly, so that readers or viewers can return to the book or film and notice things that escaped them on a first pass. Here, as elsewhere, NiFF’s desire to mix formal analysis with an account of spectator response created some gaps in the theory, but in some respects it’s better to have more to explain (about the architecture and detail of the film) than less. It’s a dynamic I’m still trying to refine many years later.

For the most part, NiFF explicitly left aside the emotional dimensions of narration. That was done on the assumption that comprehension as such was relatively insulated from affective response. You can follow a story, I claimed, without being moved by it. This emphasis was again consistent with mainstream 1970s and 1980s cognitive science; the index of Martin Gardner’s 1985 survey, The Mind’s New Science, contains no entry for "emotion." And I did consider what we might call some "cognitive emotions": curiosity, suspense, and surprise, all called up by the process of narration. In the decades since the book was written, however, the relation of emotion to cognition has become central to cognitive science, and it has been explored by several film scholars working in the cognitivist paradigm. It’s still not something I focus on, but it’s obviously of great importance.

Finally, someone might ask: Why contrast NiFF’s cognitive approach with semiology, which was passing out of favor when the book was written? Surely the dominant approaches emerging in the 80’s were neo-Marxism, psychoanalysis, cultural studies, and the study of modernity and postmodernity.

Here’s my answer. These perspectives don’t play a role in NiFF, or in this essay, because their proponents weren’t asking about how films are understood. These writers focused on questions of how social, cultural, and psychodynamic processes were represented in films. Typically those questions were answered by interpreting individual films, reading them for traces of the larger processes made salient by the given theory. My concern was explaining, not explicating; I wanted functional and causal-historical accounts of why films in various traditions displayed certain regularities in their narrational strategies. That was, I thought, most pertinent to the semiological line of inquiry.

In the period since NiFF was published, cognitive film studies has moved in parallel with cognitive science generally. We have had neurological studies of film viewing; we have seen appeals to evolutionary psychology; we have seen studies of suprapersonal patterns of emergence. These all seem to me fruitful. In what follows, I want to sketch out some ideas that I’d develop in a new and improved version of NiFF. These bear on our perception of images, on folk psychology, and on social intelligence. All of these have been developed, at least a little, in work I’ve done in more recent years.

Dog!

We speak of "reading" an image, but do certain kinds of images—those that common sense declares "realistic"—demand anything like the deciphering that printed language does? How much does grasping an image depend on learned conventions of representation?

In NiFF I waffled on the question too much. Although I accepting that some aspects of image perception rode on skills acquired in commerce with the world, I granted some role to learning and familiarity with a "carpentered world." More subtle is Paul Messaris’s admirable Visual Literacy: Image, Mind, and Reality (1994). Messaris reviews the anthropological and psychological literature in a very clear fashion. He points out that some conventions for representing depth in still images may not be widely understandable; the classic example is the drawing above, which was interpreted by viewers in some African cultures as a hunter pointing his spear at a very tiny elephant. This suggested that some pictorial depth cues require repeated exposure or training. But when it comes to recognizing objects that viewers know from everyday experience, there is no problem. The African viewers recognized the tiny elephant as an elephant.

When it comes to moving pictures, the issue is even clearer. Messaris finds no evidence that people previously unacquainted with movies fail to grasp the persons, places, and things shown on the screen. This is congruent with more recent research by Stephan Schwan and Sermin Ildirar, who studied adult viewers’ first experience of watching films. Indeed, all three researchers offer evidence that even some editing techniques are immediately understood by first-time viewers.

On the "film as language" question Messaris’s conclusions are clear:

What distinguishes images (including motion pictures) from language and from other modes of communication is the fact that images reproduce many of the informational cues that people make us of in their perception of physical and social reality. Our ability to infer what is represented in an image is based largely on this property, rather than on familiarity with arbitrary conventions (whereas the latter play a primary role in the interpretation of language, mathematics, and so on).

Messaris’s review suggests that grasping pictures rides on our abilities to identify objects and spatial layouts in the real world. Some intriguing research on infants reinforces the point.

In a famous experiment, Julian Hochberg and Virginia Brooks kept their son away from pictures during his first eighteen months. He did occasionally see billboards and a few picture books and labels, but when a picture was encountered, the parents never pointed out its contents or tried to name them. At nineteen months, when the boy was starting to spontaneously call out names of things he spotted in accidental images, "It was evident that some form of parental response to such identification would soon become unavoidable." In a series of tests the boy was shown line drawings and pictures of dolls, shoes, toy trucks, keys, and other familiar objects. He named them to a high degree of accuracy. Hochberg and Brooks concluded:

It seems clear from the results that at least one human child is capable of recognizing pictorial representations of solid objects (including bare outline drawings) without specific training or instruction. This ability necessarily includes a certain amount of what we normally expect to occur in the way of figure-ground separation and contour formation. At the very least, we must infer that there is an unlearned propensity to respond to certain formal features of lines-on-paper in the same way as one has learned to respond to the same features when displayed by the edges of surfaces….

The complete absence of instruction in the present case…points to some irreducible minimum of native ability for pictorial recognition. If it is true also that there are cultures in which this ability is absent, such deficiency will require special explanation; we cannot assert that it is simply a matter of having not yet learned the "language of pictures."

Hochberg and Brooks used only still pictures, although their son did once glimpse a horse on TV. (He cried, "Dog!") What about moving images?

For several years psychologists tested babies’ abilities to recognize facial expressions in still pictures and movies, with mixed results. Babies’ attention can be captured by external stimuli at an early age, and they start to control their focus and attention in the second month. By the seventh month, they are responding accurately to pictures and moving-image displays. Yet it’s possible that recognition starts much earlier. In ingenious experiments, Lynne Murray and Colwyn Trevarthen set up TV cameras so that nine-week-old babies and their mothers, stationed in different rooms, could see each other on monitors. The experimenters wanted to record the interactions between them, as well as to vary the timing of responses through pauses and replays.

Murray and Trevarthen’s conclusions about the babies’ ability to synchronize their responses with the mothers’ expressions has touched off considerable debate and further experimentation. That’s not what matters for us as students of cinema. What’s relevant for us is that the babies evidently did, in both real time and in tape delay, recognize the moving images of their mothers.

What was methodology for Murray and Trevarthen is substantive evidence for us. Very young babies could grasp the video image, at least to some extent, as a representation of the most familiar person in their lives. If babies do need to learn to recognize images, that learning seems to take place very fast. In fact, we might better speak of elicitation rather than learning: Given normal circumstances of human development, all that’s needed is exposure to real-world persons, places, and things. Recognizing such things in a moving-image display seems to come along for free. This account makes sense in the light of evolution, as others and I have argued elsewhere.

Folk psychology: Success stories

The Big Clock.

There is a lot more to be said and studied about grasping moving images as representations of real-world items, most saliently people, but let me turn now to some matters of narrative that I’ve rethought since NiFF was published.

Recognizing the contents of realistic images, I’ve suggested, depends heavily upon our everyday perceptual abilities. Similarly, filmic storytelling relies upon cognitive dispositions and habits we’ve developed in a real-world context. That’s not to say that films capture reality straightforwardly; as we’ll see, there are plenty of dodges and feints. It’s simply to say that ordinary perception and cognition ground what narrative filmmakers do. On that foundation quite various, even fantastic, edifices can be built.

Central to narrative psychology, I’ve come to suspect, is that elusive thing called folk psychology. Folk psychology calls on "common sense"—our everyday habits of attributing qualities, beliefs, desires, intentions, and the like to ourselves and to people around us. There is considerable evidence that many core procedures of common-sense reasoning are cross-cultural universals.

Consider "person perception." We tend to arrive at quick impressions about those around us. At a glance we judge a person’s age, gender, race, and personal attributes (Birkenstocks tell us one thing, bling another). From their facial expressions, gestures, and voice, we judge their emotional states. Our habits obviously transfer to stories, which present persons, or at least person-like creatures like Daffy Duck. To follow the story we have to assign the characters certain qualities. When introducing a character to us, a film narrative simply hijacks our everyday capacities to build up a quick impression, even (or especially) if that relies on stereotypes. That impression may be confirmed, tested, or repudiated as the story develops, but our quick and dirty habits of person perception provide a point of departure.

We also indulge in mind-reading. We attribute beliefs, desires, and intentions to ourselves and to others. You want a burger; you stop at a burger joint to get one. You act on your desire based on beliefs about the world, most notably the belief that you can get a burger at that joint. Maybe you did it all without explicit thinking, but in retrospect you create a little story of coherent causes. We interpret others’ actions the same way. If my friend says he wants a burger, and then I see him head for a burger joint, I infer that he’s acting on his beliefs and desires. Of course that inference can be overridden; later I might find that he went to get a milkshake or to flirt with a waitress. But even revising the inference requires the same schema. (Aha, he really wanted a shake, or a date for tonight.)

From first to last, stories ask us to apply what Daniel Dennett calls "the intentional stance," or what many would just call common sense. At the start of The Big Clock (1947), we see George Stroud slinking along a corridor and avoiding a guard. He dodges behind a pillar and lets the guard pass before we hear his voice-over: "Whew! That was close." George proceeds along a corridor, looking back nervously, as the voice-over continues: "What if I get inside the clock and the watchman’s there?"

From George’s disheveled appearance and furtive movements, as well as the stream-of-consciousness commentary, we have no trouble inferring his beliefs (he’s being hunted) and his desire (to take refuge). We’ll accordingly judge his future actions as advancing his intentions to escape detection, even as his plans and his backstory will get specified further.

The centrality of characters’ goals in classical filmmaking, of which Kristin and I have made much on this site and in our research, fits our folk-psychological tendency to pick out actions that fulfill desires in the light of characters’ beliefs. The web of intentions can get very complicated—think of all the beliefs and desires at play in The Godfather—but we’re very good at tracking them because we expect that social situations exhibit what people are planning to achieve. To bring babies back in, it seems that they too can do mind-reading. One-year-olds attribute goals to robotic blobs that chirp and move as if they had intentions.

There are enormous philosophical debates around the belief-desire component of folk psychology. Is it truly explanatory, or just vacuous? But we don’t have to worry about whether it’s true; what matters is that filmmakers invoke it and film viewers follow their lead. Storytellers are practical psychologists, preying (usually in a good sense) on our habits of mind in order to produce experiences.

Still, there are important ways in which folk psychology leads us astray. Film exploits those too.

Folk psychology: The downside

Daisy Kenyon.

In Everything Is Obvious* (*Once You Know the Answer), Duncan J. Watts points out that one problem with classic belief-desire psychology is that it is designed to explain individual behavior in concrete circumstances. It doesn’t scale up well to explain large-scale trends. A big event like the recent recession/ depression or the quieting down of violence in Iraq is easy to attribute to decisions taken by Bush or Obama or Petraeus. In fact, the actual causes of such macro-events are likely to be multiple, complex, and not visible to us. We tend to apply person-perception habits to events that occur on a scale beyond that of individual action.

Watts’ book is a contribution to "Wrongology," the study of our tendencies to overestimate our abilities, make simple logical errors, and act inconsistently. The research area has its roots in the studies of heuristics and biases conducted by Daniel Kahneman and Amos Tversky. Rationality, as postulated by philosophers and economists, seems to be a rare gift. To take a now-classic example:

Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.

Now, which is more likely?

Linda is a bank clerk.
Linda is a bank clerk and a feminist.

Most people say the latter is more probable, although it can’t be. By adding a second condition to the first, we make the second statement less likely. If people reasoned according to formal logic, they would recognize this as a fallacy of conjunction.

Likewise, the rational agent beloved of economists turns out to be motivated by more than gain, as shown in the so-called Ultimatum Game. Veronica has $100, while Betty has no money. According to the rules, Veronica must offer Betty some of the money, and if Betty refuses the split, neither player gets anything. Now if Veronica is a rational agent, she ought to offer Betty as little as possible, maximizing her own gain. And Betty, who starts off with zero dollars, should take even a measly $1, since that leaves her better off than before. But in experiments conducted around the world, the players in Veronica’s position tend to offer a fifty-fifty split. More surprisingly, players in Betty’s position tend to reject offers of less than $30, leaving both players with nothing.

Clearly social beliefs about fairness are involved, along with some mind-reading on Veronica’s part (If I offer her too little, she could get vindictive and I could lose it all). Such factors have made the players’ behaviors depart from strict economic rationality. Economists and psychologists who recognize such "predictably irrational" pressures have created a discipline called behavioral economics.

So folk psychology has its own biases. Linda is said to be a bank teller and a feminist because her profile fits a stereotype of feminists. This is sometimes called the availability heuristic, the tendency to apply the handiest schema to a situation. There is as well confirmation bias, the habit of looking only for evidence that supports the idea you’re leaning toward. Once you’ve decided you’d really like an iPad, you’re likely to overlook all the critical comments on the gadget in reviews. If you believe in astrology, you’ll tend to remember the times that your horoscope seemed to predict what happened to you and forget the more numerous times when it failed to do so. Watts points out the reconstructive nature of memory as another biasing effect. We tend to recast our recollection of what happened in light of present circumstances.

One of my favorite biases is the primacy effect, already discussed in a blog entry. Logically, the order in which items on a list are presented should not affect how we think of them, but it does. Take Hong Kong supermogul Leonard Ho’s three daughters. Which of their names doesn’t belong with the other two?

Maisy, Daisy, and Pansy

Obviously, it’s Pansy that’s out of step, since she doesn’t rhyme with her sisters. But present them in reverse order:

Pansy, Daisy, and Maisy

…and the outlier is Maisy, who isn’t named after a flower. The first item in a series tends to serve as a benchmark against which we measure the ones that follow. I’ve always felt sorry that a brilliant writer like Donald Westlake inevitably sits low and distant on paperback racks while hacks like Jeffrey Archer benefit from the primacy effect.

Again and again, narratives manipulate our psychological biases. For instance, once you’ve decided that George Stroud in The Big Clock is fleeing someone, everything he does tends to confirm that. The filmmakers exploit confirmation bias. Likewise, the syuzhet layout relies on the primacy effect. The film starts at a point of crisis, with George fleeing his boss’s goons. The narration then flashes back to the beginning of the action, when George tried to escape from Janoth’s overweening control by taking a long-promised vacation with his family. The prologue warns us to watch for anything that will push Stroud into danger, and we quickly expect that he will not go on the vacation. Had the film begun with the more prosaic events that come earliest in the fabula, we would not have been on the alert for Stroud’s plunge into a critical situation. The prologue also signals the importance of the clock as a sinister force and time as a motif through the film.

Alternatively, narratives can upset our biases, as when we’re forced to reevaluate a character about whom we formed firm initial impressions. We’re obliged to do this with Danny Ciello in the closing moments of Prince of the City. Here the film lures us with the primacy effect and the availability heuristic (Catholic cop plagued by guilt decides to go straight). Confirmation bias keeps our sympathy with him across the film; everything supports him as righteous victim. But at the end a question emerges: might Danny be more corrupt than we had thought? Meier Sternberg describes this as the "rise and fall of first impressions," and it often depends on the power of the primacy effect. At the limit, a narrative film can try to avoid setting up any clear-cut first impressions, as happens in "art films" like In the City of Sylvia.

Sometimes one aspect of folk psychology can rescue another. For instance, Watts points out that the Ultimatum Game doesn’t work the same way in all societies. In experiments with the Machiguenga tribe of Peru, the moneyed partner tended to offer only about a quarter of the total, and nearly all offers were accepted. Both parties were being more "rational" by the economists’ standards. But now belief-desire psychology kicks in. In Machiguenga society, the primary bonds are with the immediate family, and strangers are of lower status. So the moneyed partner felt little obligation to make a fair split, and the recipient was happy with whatever was offered. Once we know this, the Machiguenga strategy makes sense. A similar sort of thing happens in fantasy and science-fiction films. Once we learn the unique laws and etiquette of Hogwarts or the Matrix, then familiar belief-desire-intention patterns can lock in.

Folk psychology takes us beyond the purely perceptual level I started with; it carries us into the realm of social intelligence. Mind-reading requires us to detect, sometimes on very faint cues, what people are expressing or signaling through their behavior. Elsewhere I’ve talked about this in cases involving eye behavior—blinking and eyebrows, in particular. But there’s much more to be done with the ways in which cinema mobilizes our social intelligence in order to track a narrative. Sometimes the narrative eases our task by making things redundant and clear; sometimes the film throws up problems, making it hard to understand characters’ intentions or reactions, as in the enigmatic veteran played by Henry Fonda in Daisy Kenyon.

I found the concept of social intelligence especially useful in explaining a form of cinematic storytelling that has become prominent since the 1990s, what I called the network narrative. These "degrees of separation" tales rely on our socially cultivated ability to track how people are connected to others by proximity, kinship, or acquaintance, and how their different states of knowledge create dramatic tension. We might as well call it the soap-opera effect. Again, however, such films are likely to streamline the vast complexity of real-world social networks: the networks in movies like Love, Actually and Sunshine State tend to be simple and redundantly stated. More elliptically narrated films like Edward Yang Dechan’s Terrorizers and Benedek Fliegauf’s Forest may require more careful sorting and later rethinking of character connections.

Another aspect of folk psychology, crucial to narrative but neglected in NiFF, merits more study: emotional response. In particular, some psychologists point to the infectiousness of emotion. Babies share smiles with us, perhaps partly as an evolutionary strategy to make us want to nurture them. (Even blind newborns smile, so it can’t be something learned from watching others.) Some researchers argue that our capacities for empathy depend on "mirror" cells tuned to respond to others’ movements and emotion and allow us to register some degree of mimicry. Macaque monkeys’ mirror neurons fire not only when they watch a mate grasping a cup but also when they watch a film of a mate doing it. (More evidence that film images require no special code-learning.) V. S. Ramachandran suggests that mirror neurons could explain the fact that a mother sticking out her tongue provokes her newborn baby to do the same, a presumably innate response. If assumptions about mirror neurons in humans turn out to be well-founded, we may find that cinema, with its ability to capture gesture and the flickers of facial expressions, is an ideal medium for triggering involuntary reactions (kinesthetic, emotional), which can in turn can be recruited for narrative purposes.

Uncommon sources of common sense
Someone might suggest that this general line of thinking leads to observations that are quite superficial. What viewer doesn’t see that George Stroud in The Big Clock is trying to avoid the guards? I’d reply that once we move beyond the moment to look at strategies of patterning at different scales, we find things aren’t so obvious; that was the primary task of NiFF. But I grant that our point of departure will seem very commonsensical. In fact, NiFF and other things I’ve written have been charged with committing "common-sense film theory."

In one way that’s true. The humanities have in general suffered from straining for the most far-fetched accounts of how art, literature, and music work. In the literary humanities in particular, ingenious interpretations—often relying on free-association, wordplay, and talking points lifted from favored penseurs—get more notice than plausible explanations do. In various places I’ve argued for naturalistic and empirical explanations as the best option we have in answering middle-range questions, and even bigger ones like "How do we comprehend movies?" Sometimes our answers will not be counterintuitive. To say that looking at images recruits our skills of looking at the world will not surprise many people; but it is likely to be true. What’s likely to be counterintuitive are the discoveries of mechanisms that undergird perception. Would common sense predict that an object’s form, color, movement, and spatial location are analyzed along distinct pathways in the visual system? Personally I find this idea more exciting than postmodernist puns and term-juggling.

More important, we can embrace common sense at a meta-level. Recognizing that it is in play in narrative comprehension makes it something we need to analyze. We can understand filmic understanding better if we recognize what’s intuitively obvious, and then go on to ask what in the film, and in our psychological and social make-up, makes something obvious. And those factors may not be obvious in themselves. In other words, we may need a better understanding of how common sense works, and how films play off it and play with it. That understanding may in turn oblige us to accept empirical experiment, evolutionary thinking, and neurological research—all of which most literary humanists find worrisome.

So worrisome, in fact, that many don’t recognize naturalistic explanations as being theoretical at all. For them, the only theories that exist are Big Theories, and so efforts like the one I just mentioned are condemned as expressing a disdain for or suspicion of theorizing tout court. But that objection, feeble to start with, was blocked back in 1996 by the opening sentences Noël Carroll and I wrote in our Post-Theory: Reconstructing Film Studies:

Our title risks misleading you. Is this book about the end of film theory? No. It’s about the end of Theory, and what can and should come after.

That introduction and many of the pieces included in the volume float arguments for theorizing as an activity that asks researchable questions and comes up with more or less plausible answers—some commonsensical, some not, and some probing what counts as common sense.

Ironically, just as filmic interpretation is amenable to task analysis from a cognitive standpoint, a surprising amount of Grand Theory seems to me to rely on the sort of folk-psychological schemas and shortcuts that we find in ordinary life. But that’s a whole other essay.

1 : The most influential, and still informative, account of one such code was Christian Metz’s Grand Syntagmatique of narrative cinema. See Metz, "Problems of Denotation in the Fiction Film," in Film Language: A Semiotics of the Cinema, trans. Michael Taylor (New York: Oxford University Press, 1974), 108–146. Metz’s more general consideration of cinematic codes is to be found in his Language and Cinema, trans. Donna Jean Umiker-Sebeok (The Hague: Mouton, 1974).

2 : Peter L. Berger and Thomas Luckmann, The Social Construction of Reality: A Treatise in the Sociology of Knowledge (New York: Anchor, 1967).

3 : On this Golden Oldie of humanities lore, see Geoffrey Pullum, The Great Eskimo Vocabulary Hoax and Other Irreverent Essays on the Study of Language (Chicago: University of Chicago Press, 1991), 159–175.

4 : See Howard Gardner, The Mind’s New Science: A History of the Cognitive Revolution (New York: Basic Books, 1985).

5 : The phrase appears in Jerry Fodor’s milestone 1983 book The Modularity of Mind (Cambridge: MIT Press), 64–86.

6 : The phrase became something of a slogan for the New Look school. See Jerome S. Bruner, Beyond the Information Given: Studies in the Psychology of Knowing, ed. Jeremy M. Anglin (New York: Norton, 1973).

7 : J. J. Gibson, Perception of the Visual World (Westport, CT: Greenwood, 1974 [orig. 1950]; J. J. Gibson, The Senses Considered as Perceptual Systems (Boston: Houghton Mifflin, 1966).

8 : On the Ames Room, see William H. Ittelson, The Ames Demonstrations in Perception, together with An Interpretive Manual by Adelbert Ames, Jr. (NewYork: Hafner, 1968). Go here for many videos employing the principles of the Ames Room. Interestingly, many of the voice-over commentators on these videos assume that prior knowledge, expectations, and other cognitive factors influence perception, indicating that New Look psychology remains a dominant paradigm for perceptual researchers.

9 : Gibson made his arguments about movies in The Ecological Approach to Visual Perception (Boston: Houghton Mifflin, 1979), 292–302. See Joseph D. Anderson, The Reality of Illusion: An Ecological Approach to Cognitive Film Theory (Carbondale: University of Southern Illinois Press, 1996) and the articles collected in Moving Image Theory: Ecological Coniderations, ed. Joseph D. Anderson and Barbara Fisher Anderson (Carbondale: University of Southern Illinois Press, 2005). Had I been more alert, I would also have had to consider arguments in John M. Kennedy’s A Psychology of Picture Perception (San Francisco: Jossey-Bass, 1974).

10 : Julian Hochberg and Virginia Brooks, "Movies in the Mind’s Eye," in In the Mind’s Eye: Julian Hochberg on the Perception of Pictures, Films, and the World, ed. Mary A. Peterson, Barbara Gillam, and H. A. Sedgwick, (New York: Oxford University Press, 2007), 387–395.

11 : See my "Cognition and Comprehension: Viewing and Forgetting in Mildred Pierce," Poetics of Cinema (New York: Routledge, 2008), 135–150. The chart in the essay was printed inaccurately; an accurate one is at http://www.davidbordwell.net/books/poetics.php.

12 : See The Way Hollywood Tells It: Story and Style in Modern Movies (Berkeley: University of California Press, 2006), 63–71.

13 : Major examples include Ed Tan, Emotion and the Structure of Narrative Film: Film as an Emotion Machine (Mahwah, NJ: Erlbaum, 1996); Torben Grodal, Moving Pictures: A New Theory of Genres, Feelings, and Cognition (Oxford: Oxford University Press, 1997), Carl Plantinga and Greg M. Smith, eds., Passionate Views: Film, Cognition, and Emotion (Baltimore: Johns Hopkins University Press, 1999), Greg M. Smith, Film Structure and the Emotion System (Cambridge: Cambridge University Press, 2007), and Carl Plantinga, Moving Viewers: American Film and the Spectator’s Experience (Berkeley: University of California Press, 2009).

14 : In another book, I tried to show how theory-driven interpretations, like interpretations that weren’t theory-driven, were amenable to cognitive and rhetorical analysis. See Making Meaning: Inference and Rhetoric in the Interpretation of Cinema (Cambridge: Harvard University Press, 1989).

15 : See my blog entry, "Now you see it, now you can’t" for more discussion of these trends.

16 : See William Hudson, "Pictorial Depth perception in Subcultural Groups in Africa," Journal of Social Psychology 52 (1960), 183–208, and "The Study of the Problem of Pictorial Perception among Unacculturated Groups," International Journal of Psychology 2 (1967), 89–107.

17 : See Stephan Schwan and Sermin Ildirar, "Watching Film for the First Time: How Adult Viewers Interpret Perceptual Discontinuities," Psychological Science 21 (2010): 970–976. Online access is at http://pss.sagepub.com/content/21/7/970.abstract.

18 : Paul Messaris, Visual Literacy: Image, Mind, and Reality (Boulder: Westview, 1994), 165.

19 : Julian Hochberg and Virginia Brooks, "Pictorial Recognition as an Unlearned Activity: A Study of One Child’s Performance," in In the Mind’s Eye, 64.

20 : For a review of early experiments, see Charles A. Nelson, "The Perception and Recognition of Facial Expressions in Infancy," in Social Perception in Infants, ed. Tiffany M. Field and Nathan A. Fox (Norwood, NJ: Ablex, 1985), 101–125. Later experiments in infant cognition are considered in the light of "folk" theories of mind, physics, and the like, in Steven Pinker, How the Mind Works (New York: Norton, 1997), Chapter 5.

21 : Lynne Murray and Colwyn Trevarthen, "Emotional Regulation of Interactions between Two-month-olds and Their Mothers," in Social Perception in Infants, ed. Field and Fox, 177–197. Ellen Dissayanake has proposed a fascinating theory of the origins of art based in mother-child interactions; see Art and Intimacy: How the Arts Began (Seattle: University of Washington Press, 2000).

22 : See, for example, Phillipe Rochat, Ulric Neisser, and Viorica Marlan, "Are Young Infants Sensitive to Interpersonal Contingency?" Infant Behavior and Development 21, 2 (1998): 355–366.

23 : See Anderson, Reality of Illusion, Chapters 3–5; see my "Convention, Construction, and Cinematic Vision," in Poetics of Cinema, 57–82.

24 : See for a summary Donald E. Brown, Human Universals (Philadelphia: Temple University Press, 1991), 130–140.

25 : Daniel C. Dennett, The Intentional Stance (Cambridge: MIT Press, 1987).

26 : Alison Gopnik, The Philosophical Baby: What Children’s Minds Tell Us about Truth, Love, and the Meaning of Life (New York: Farrar, Straus and Giroux, 2009), 98.

27 : To get the flavor of some of the debates, see Barry Loewer and Georges Rey, ed., Meaning in Mind: Fodor and His Critics (Cambridge: Blackwell, 1991), especially Daniel C. Dennett, "Granny’s Campaign for Safe Science," 87–94 and Fodor’s reply: "The enormous practical success of belief/desire psychology makes a prima facie case for its approximate truth" (277). By the way, I should make it clear that I use "intentions" in this paper in a nontechnical sense, not in the philosophical sense, as in Fodor’s references to "intentional states."

28 : See Daniel Kahneman, Paul Slovic, and Amos Tversky, eds., Judgment under Uncertainty: Heuristics and Biases (Cambridge: Cambridge University Press, 1982).

29 Meir Sternberg, Expositional Modes and Temporal Ordering in Fiction (Bloomington: Indiana University Press, 1978), 99–102. More generally, a great deal of Narration in the Fiction Film, including its focus on curiosity, suspense, and surprise, is indebted to Sternberg’s book, a trailblazing work of modern narratology.

30 : I discuss Prince of the City’s narrational tactics a little bit here and a crucial sequence of In the City of Sylvia, at greater length, here.

31 : Duncan J. Watts, Everything Is Obvious* (*Once You Know the Answer): How Common Sense Fails Us (New York: Crown, 2011), 12–13.

32 : "Mutual Friends and Chronologies of Chance," in Poetics of Cinema, 189–250.

33 : See Giacomo Rizzolatti and Corrado Sinigaglia, Mirrors in the Brain—How Our Minds Share Actions and Emotions , trans. Frances Anderson (New York: Oxford University Press, 2008) and Marco Iacoboni, Mirroring People: The New Science of How We Connect with Others (New York: Farrar, Straus and Giroux, 2008)

34 : V. S. Ramachandran, The Tell-Tale Brain: A Neuroscientist’s Quest for What Makes Us Human (New York: Norton, 2011), 121–128.

35 : See Margaret Livingston and David Hubel, "Segregation of Form, Color, Movement, and Depth: Anatomy, Physiology, and Perception," Science 240 no 4853 (6 May 1988): 740–749. Available at http://www.sciencemag.org/content/240/4853/740.abstract.

36 : "Introduction," Post-Theory: Reconstructing Film Studies, ed. David Bordwell and Noël Carroll (Madison: University of Wisconsin Press, 1996), xiii. More generally, much of what I’ve said in this online essay was said more pointedly in Carroll’s pioneering 1985 article, "The Power of Movies." See Carroll, Theorizing the Moving Image (Cambridge: Cambridge University Press, 1996), 78–93.