Chapter 2Typing
Traits and Types
When we watch a film, our perception of the characters is organized by our prior
knowledge about people. We don’t start from scratch with each person, representational
or real, that we encounter. In art as in life, we sort people into categories, which can be
described in various ways: personality traits, prototypes, stereotypes, or character types.
For simplicity’s sake, I will call these character categories types, to distinguish this
concept from others that are closely related but not identical to it. Types may operate at
various degrees of generality, from highly specific genre types such as gangster’s moll
and mad scientist to very general social types such as man and child. Traditional literary
and film criticism has not been kind to types: as one influential study of narrative puts it,
“insofar as a character is a type, he is less a character.”1 But insofar as type means
category, we cannot help but think of characters as types—they permeate our whole
experience of character. This chapter will try to show the necessity of typing to
characterization, attempting to rescue the concept of the character type from the scrap
heap of critical concepts and show its indispensability both to theorizing narrative and to
understanding specific characters.
In chapter 3 I shall discuss characters’ personality traits which spectators infer on
the evidence of their behavior or, less commonly, which narration states explicitly in
dialogue. That discussion is complementary to this one: once a character’s personality
has been identified in relation to narrative events, we can begin to think about what
categories the character belongs in. However, while this description may have some
theoretical purchase or may be a useful procedure for doing film criticism, as a
descriptive chronology of film viewing, it is simply backwards. We don’t first perceive
traits, then types; categorization of people doesn’t work in such a linear fashion.
Types are actually more salient than traits because they are more richly
associative.2 We tend to think more in types than traits, even if types are defined to some
extent by their constituent traits. You don’t first see that a person is short, dressed for
school, speaking in a high pitch, and calling an adult “Mom,” then reason that he must be
a child. Rather, you see the traits and the type all at once, and the type offers you more
useful information than the traits. Traits function in networks of generalized information
about people; each type may be seen as a connection of associated trait nodes (and also of
associated behavior or emotion nodes).3
Types can be divided into levels, from the most to the least specific: low,
basic/middle, and high. Research has shown that middle-level types—confusingly called
“basic”—are more salient than high-level and low-level types. “Femme fatale” (middle-
level) is more likely than either “woman” (high-level) or “bookstore clerk” (low-level) to
offer a combination of rich, vivid, concrete and distinct associations that come quickly to
mind. This means that the mind prefers to think in basic-level categories; the first words
children learn tend to be basic-level terms. They are more likely to know “apple” than
“fruit” or “golden delicious.” Middle-level terms are useful for maximizing the
associations generated by the specific instance. This has the effect of easing the
cognitive demands on the social perceiver. If the world appeared to you largely in high-
level or low-level categories, it would overwhelm your ability to make efficient sense of
it.4
Already, it should be clear why types are a useful concept for characterization.
Categorization is something our minds do very well, and this process is of central
importance to narrative comprehension. Filmmakers, as intuitive psychologists,
recognize that our encounter with characters always involves typing. Sometimes, the type
becomes prominent because of its inability to include its expected traits, or because of its
incompatibility with other aspects of the character. When they are not trying to use types
as aesthetic devices in their own right, as when they create imaginative hybrid types (e.g.,
the nymphomaniac nun in Hal Hartley’s Amateur), caricatured exaggerations of types
(e.g., the broad supporting characters in Pee-Wee’s Big Adventure), or subversions of
defining type-traits (e.g., the maternal police chief in Fargo), filmmakers stick to
recognizable, middle-level types for the sake of clarity and efficiency. It follows that
types are an important aspect of characterization both because they are one of the
significant cognitive dimensions of person perception, and because filmmakers
manipulate typing for aesthetic effects.
How, then, does this process of typing work? What influences this
categorization? How do people sort instances—in this case, people or characters—into
types? Not according to necessary and sufficient conditions, as in the classical view of
categories and concepts. Character types cannot be defined formally by reference to
traits that are both necessary for the categorization and jointly sufficient for it. This
approach has been superceded not only in reference to person-categorization, but in many
other domains as well. Most philosophers and psychologists agree that categories are
typically fuzzy, especially around the edges. Membership is determined according to
similarities between instances (or as Wittgenstein termed them, family resemblances).
We determine that a person is a child by perceiving similarities between her and
examples we have in mind of children. Our categorization is probabilistic in the sense
that we quickly sort people into types on the basis of what seems most likely. In ordinary
encounters with others, we don’t have time or interest in making absolutely sure that our
typing is correct.5
Some categories are understood by comparison with prototypes or exemplars—
the more similar to these prototypes or exemplars, the more likely they are to be included
1 Robert Scholes and Robert Kellogg, The Nature of Narrative (London: Oxford, 1966), 204.
2 Susan M. Anderson and Roberta L. Klatzky, “Traits and Social Stereotypes: Levels of Categorization in Person Perception” Journal of Personality and Social Psychology 53 (1987), 235-246
3 Ibid.
4 Elanor Rosch, et al., “Basic objects in natural categories” Cognitive Psychology 8 (1976), 382-439; Nancy Cantor and Walter Mischel, “Prototypes in Person Perception” in Leonard Berkowitz (ed.), Advances in Experimental Social Psychology vol. 12 (New York: Academic P, 1979), 3-52.
5Edward E. Smith, “Categorization,” in Daniel N. Osherson and Edward E. Smith (eds.), An Invitation to Cognitive Science: Thinking Vol. 3 (Cambridge: MIT P, 1990), 33-53;Kunda, 15-52.
in the category. These prototypes or exemplars have a large number of attributes
characteristic of the category, and these attributes are weighted according to their
salience. If an apple is an exemplar of the category fruit, it is more important that it is
edible than that it is round, because edibility is a more salient characteristic of fruits than
roundness.6
A prototype is an abstract (imaginary) instance of the category that has a full
complement of its traits, while an exemplar is a real instance of the category. A
prototype of a professional basketball player would have all of the characteristics that are
associated with the category: young, tall, runs and jumps well, and physically fit. An
exemplar of a professional basketball player would be Michael Jordan. There is some
debate about whether people use prototypes or exemplars in their formation and
activation of categories, and it could be that people use each one as necessary or
appropriate. Either way, we judge an instance to be a member of a category by its
similarity to other instances, real or abstract.7
Increased similarity also means increased typicality.8 The more typifying traits
the instance has, the more it seems to be a good example of the category. And salient
traits that are dissimilar to the prototype or to exemplars make the instance seem less
typical. Apples are more typical fruits than olives, partly because apples are sweet (a
6 Smith.
7 Kunda, 15-32.
8 Ibid.
prototype trait) and olives are bitter or salty (both dissimilar traits). Sam Spade of
Huston’s The Maltese Falcon is a more typical private eye than Philip Marlowe of
Altman’s Long Goodbye because Spade has more in common with the prototype of the
private eye than Marlowe does, and because Marlowe has some salient traits that clash
with the private eye prototype (e.g., he looks after a pet cat, he lacks Spade’s tough-guy
demeanor). Indeed, Spade functions as an exemplar of the private eye, so Marlowe of
The Long Goodbye is less typical because of his distinctness from Spade.
Relations of similarity help us identify many categories, yet there are other
relations that hold some categories together. Some are identified by applying an implicit
theory about the world, for example, a theory about causal relations among things. The
category “mother” can be understood in various ways, some of which identify a person as
having given birth to a child. Ruby and Amy may be incredibly similar—identical twins,
even—but if only Amy has no children, only Ruby is a mother. “Food” is understood as
giving nourishment when eaten. A wax hamburger might bear more similarity to a real
hamburger in size, shape, weight and color than it does to a wax museum statue of
Napoleon, but real hamburgers belong in the category “food” and wax ones do not. In
one experiment, when researchers asked subjects whether a three-inch disc is more
similar to a quarter or a pizza, most said a quarter. But when asked which it was more
likely to be, the answer was a pizza, because we believe that quarters can be only one
size, which is not three inches, but we hold no such beliefs about the size of pizzas. So
category membership is not defined exclusively by similarity, since theories about the
world (such as theories about the size of coins and pizzas and about the edibility of foods)
come into play.9
Yet objects can generally be classified in many different ways, sometimes by
relations of similarity (real and wax hamburgers may both belong to the category
“hamburger” on the basis of their similarity), other times by relations of causality (real
hamburgers and wax ones are not both food). It depends on the categories and on the
situation. There are dozens of traits associated with mothers, and if you didn’t know
whether or not a given woman had a child, you might categorize her as a mother (or,
more likely, regard her as a maternal type) based on her similarity to your prototype or to
an exemplar—perhaps your own mother. Part of what makes typing interesting—and
challenging to analyze—is that objects can be categorized in many different ways and
into many different categories.
Furthermore, just as there are different ways of categorizing, there are different
kinds of categories. There are naturally occurring categories, like fruit and mother, and
there are human-made categories, like private eye and hamburger. Both of these kinds of
categories can be understood by applying ideas about similarity and by activating causal
theories. A hamburger is designed to be eaten, and we are more likely to judge
something a hamburger if it seems edible (causal theory). A fruit is typically round and
sweet, and we are more likely to judge something a fruit if it looks and tastes like an
apple than if it looks and tastes like celery (similarity). People can be categorized into
9 Gregory L. Murphy and Douglas L. Medin, “The Role of Theories in Conceptual Coherence” Psychological Review 92 (1985), 289-316; Pinker 308-9.
natural and human-made categories as well. Infant, toddler, child, adolescent, adult,
elderly: all are categories of age based on natural distinctions according to relative age.
Even if some cultures and some languages carve the ages up differently and label them
accordingly, all have some notion of young and old and some articulation along the scale
from birth to death. But many of the markers of social identity are purely cultural or sub-
cultural, such as occupation categories like blue-collar worker and corporate executive.
So person-categories fall into two kinds, and are understood in two ways. When making
inferences about other people, much depends on which kind of category is activated in a
given situation and on whether categorization proceeds following from the application of
a theory or from a contextual judgment of similarity.
For example, Mr. Orange (Tim Roth) in Tarantino’s Reservoir Dogs is both a
police officer and a member of the criminal gang attempting a heist. The first type
judgment we make in relation to him is to recognize that he is one of the gang. He
dresses like them and is carrying out their collective goal. He has taken a bullet to the
abdomen in the course of the heist, so we are inclined to sympathize with him and see
him as just one of the guys.
Part of the puzzle of the film’s narration and characterization consists of
recognizing similarities and differences among the men and determining which will be
significant. They all wear the same dark suits and ties, all talk a similar tough street lingo
peppered with popular culture references, and are united by their task. Yet each of them
has several distinctive traits. For instance, aside from their color-names, there are
differences of temperament. Mr. Blonde (Michael Madsen) is psychopathically violent;
Mr. Pink (Steve Buscemi) is high-strung, hyper-verbal, and driven by professionalism;
Mr. White (Harvey Keitel) is empathetic but firm.
Mr. Orange is one of the quieter of the group at first; he seems to be studying the
others. This seems reasonable given the situation: the men don’t know each other—not
even their real names—and are brought together just for the one job. But after the
credits sequence, when we see that he has been wounded, he speaks in a high pitch, with
anguish and fear, writhing in pain in the back seat of the stolen car. He begs Mr. White
to drop him off at a hospital so that he doesn’t die. We are hesitant to attribute this
behavior to his individual disposition (see chapter 3 on trait attribution) and say that he
lacks courage, since he is clearly in critical condition and cannot be faulted for risking
being caught—or risking his fellow gang members being caught—to save his life. Yet
he does not seem to be cut from the same cloth as the other men, whose verbal bravado
and masculine aggressiveness seems more typical of street hustlers (or of their cinematic
representations).
Tarantino also gives us bits of casual business to flesh out the other main
characters. Mr. Brown (Tarantino) begins the film with a sexually charged interpretation
of “Like a Virgin;” Mr. Pink refuses to tip the waitress at the diner and makes a big deal
about how he doesn’t believe in tipping, but gives in when the boss, Joe, demands that he
pitch in his dollar; and many of the men use ethnic slurs in conversation at the table. Mr.
Orange, however, is more taciturn and his character is thus sketchier in the beginning.
By withholding some of Mr. Orange’s traits, Tarantino makes our typing of him more
tentative.
The center of the narrative’s dramatic conflict is the presence of the undercover
cop among the group. The enigma of the cop’s existence or identity is introduced early
in the film and structures the spectator’s process of discovery of the characters. When the
cop is revealed to be Mr. Orange, we are forced to rethink our typing. Mr. Orange’s
distinctiveness is hard to identify since from the early part of the film, he is constantly
writhing and moaning in pain. But as we learn about his preparation for the job, we learn
that his difference from the other characters is based not on similarities or lack of
similarities, because these features cease to be significant on a dramatic level. His
difference from the others is based on a causal distinction between criminals whose
purpose is to make money, and a cop whose purpose is to infiltrate their ranks and help
stop their crime. Tarantino’s careful withholding of information about Orange forces us
to type him using criteria of similarity at first; once the character’s goals are made clear,
however, our evaluation of him changes completely.
Stereotypes
One kind of person category is a stereotype, but stereotype is a somewhat confusing term.
Originally it referred to a printing plate much in use in the 19th century. In 1922 Walter
Lippmann appropriated the term to refer to many different kinds of preconceptions about
the world that impede people’s ability to see reality as it is. Its connotations were of
fixity and standardization. Lippmann writes, “For the most part we do not first see, and
then define, we define first and then see.”10 This is echoed in Ernst Gombrich’s theory of
the history of artistic representation in Art and Illusion, which identifies the idea of a
“mental set” that guides our vision, and which changes over time.11 The difference
between modes of representation in different historical eras is attributable to the artists’
differing mental sets. These are made up of stereotypes, the term Gombrich uses to refer
to the artist’s visual expectations, which he argues are the starting point of representation:
[The artist] begins not with his visual impression but with his idea or concept: the German artist with his concept of a castle that he applies as well as he can to that individual castle, Merian with his idea of a church, and the lithographer with his stereotype of a cathedral. The individual visual information, those distinctive features I have mentioned, are entered, as it were, upon a pre-existing blank or formulary. And, as often happens with blanks, if they have no provisions for certain kinds of information we consider essential, it is just too bad for the information.12
A stereotype, then, is a set of expectations about some aspect of the world, and since the
world is so dense with information, no set of expectations we have about it could possibly
account for it all. The point of using these expectations to make sense of the world is
simplification—it promotes cognitive efficiency and frees the mind for other tasks. But
in using these expectations, we sometimes miss important things that we might be better
off noticing.
10 Walter Lippmann, Public Opinion (New York: Harcourt, Brace, 1922), 81.
11 Ernst Gombrich, Art and Illusion: A Study in the Psychology of Pictorial Representation (London: Phaidon, 1959), 53.
12 Ibid, 62-63.
While Gombrich and Lippmann use the term stereotype to refer generally to
concepts, its most common usage today is in reference to persons. The Oxford English
Dictionary definition of stereotype reads: “A preconceived and oversimplified idea of the
characteristics which typify a person, situation, etc.; an attitude based on such a
preconception. Also, a person who appears to conform closely to the idea of a type.”
Neither the OED nor Lippmann specifies that stereotypes consist of traits associated with
a race, ethnicity, gender, or social identity in particular, though Lippmann does use the
term to refer—among various other things—to national types. According to these
conceptions, you could have a stereotype of your neighbor if you always see her as
virtuous but simple-minded, failing to recognize any characteristics that may not fit your
conception, such as her bad gambling habit and her proficiency at word games, or
dismissing them as exceptions to the general rule.
But in many academic fields—and colloquially as well—stereotypes usually refer
to characteristics of groups of persons or to specific roles rather than to individuals. And
the term’s connotations suggest that the effects of group categorization are pernicious.
As one oft-cited social cognition text puts it, stereotypes are “the cognitive culprits in
prejudice and discrimination.”13 The phrase “negative stereotypes” is common, but its
alternative is not “positive stereotypes,” but rather no stereotypes at all. In most
discussions, stereotype is assumed to refer to a constellation of traits that is harmful to the
group under description because it derogates them. It is also assumed to contain false
13 Fiske and Taylor,161.
traits, whether they be exaggerations of statistically valid generalizations or absolute
inventions. These include ethnic stereotypes of Jews being cheap, Irish being drunks, and
blacks being lazy; or occupation stereotypes of lawyers being rapacious and librarians
being homely. Much social cognition research is devoted to how stereotypes might be
changed or counteracted.14 We must assume that there would be no reason for
counteracting stereotypes that were oversimplifying—false, even—but not harmful, such
as stereotypes of firefighters as selfless heroes.
Yet it is in our nature to categorize, and the categorization of people has a
significant social use value.15 Group categorization tends to have a self-affirming and
communal function, promoting the in-group at the expense of the out-group.16 This point
is complementary to one in Chapter 4 about in-group and out-group differences in the
recognition of emotion expressions, such as the facial expression of fear or anger. In-
groups recognize their own members differently from out-group members, and these
relations are highly significant in explaining social and cultural variability of response to
film and other media.17 People are more likely to activate a stereotype based on an out-
14 Walter G. Stephan, “Intergroup Relations” in Gardner Lindzey and Elliot Aronson (Eds.), The Handbook of Social Psychology 3rd ed., Vol. II (New York: Random House, 1985), 599-658.
15 Pinker,126-129.
16 Stephan.
17 Hillary Anger Elfenbein and Nalini Ambady, “On the Universality and Cultural Specificity of Emotion Recognition: A Meta-Analysis” Psychological Bulletin 128 (2002), 203-235.
group categorization (such as a racial stereotype) when the member of that group has
wronged or harmed them, or evaluated them negatively. We naturally know more about
our group than about other ones, so our ideas about out-group characteristics are likely to
be wanting in some respects. People tend to see greater variability among members of
the in-group than among members the out-group; the idea that the members of the out-
group all look alike has a basis in our cognitive structure. These facets of human
psychology explain why stereotype categories tend to simplify or distort the
characteristics of out-groups.18
Stated crudely as above, gross generalizations about ethnic groups are obviously
false and those who subscribe to them are bigots. But stereotypes may function in more
benign or neutral ways. For example, we might consider a stereotype of Muslims to
include avoiding pork, a stereotype of lawyers to include working long hours, and a
stereotype of women to include caring for children. Of course, some Muslims love
bacon, some lawyers go home at 5 o’clock every day, and many women never care for
children and have little interest in doing so. But stated as probabilities—that it is more
likely for Muslims than non-Muslims to avoid pork, more likely for lawyers than non-
lawyers to work long hours, and more likely for women than men to care for children—
these are just truisms. Our use of stereotypes is typically probabilistic: you encounter
18 Fiske and Taylor, 159-167; David L. Hamilton and Jeffrey W. Sherman, “Stereotypes” in Robert S. Wyer, Jr., and Thomas K. Srull (eds.), Handbook of Social Cognition 2nd ed. (Hilldale, NJ: Lawrence Erlbaum Associates, 1994), 1-68 (8-9).
someone new and figure that on the basis of belonging to category X they are likely to
possess traits Y and Z but not A and B.
The problem with social stereotypes is that they generally contain some truth and
some falsehood, and although some may be invented out of whole cloth, many can be
explained by causal factors and may be statistically valid generalizations.19 For example,
the stereotyped association between African Americans and criminality has a causal basis
in the economic disadvantages faced by American blacks and is supported by statistical
rates of violent crime. The social problems of the African American community have
multiple causes and are products of centuries of oppression; generalizations about
African Americans that arise out of these problems are not necessarily false. But of
course, it is only a fraction of any population that commits crimes, so associating
criminality with all African Americans or all members of any group is simply wrong. As
Steven Pinker argues, the problem with stereotypes is not that they are necessarily false,
but that they deny people their individuality. No one should be judged on the basis of
their group membership alone and assumed to share all of the characteristics
stereotypically associated with the group.20
Yet this very thing happens, often with amazing cognitive efficiency. The social
cognition research on stereotype activation and application makes for depressing
reading.21 According to this research tradition, stereotypes are not merely clusters of
19 Fiske and Taylor, 166.
20 Pinker, 313.
traits, but are cognitive structures applied in understanding others. They are made of
patterned networks of associated links. Stereotypes can be activated automatically,
without the perceiver intending to do so or even being aware of it.22 Although there are
individual differences in the activation of negative associations, even among people who
consider themselves to be non-prejudiced, full-blown stereotypes come to mind
automatically when primed with negative cues about a group.23 Once activated,
stereotypes function to bias future information about a person so that subjects are more
likely to consider evidence that confirms the stereotype and ignore contradictory cues.
Subjects are also more likely to attribute traits or behavior inconsistent with the
stereotype to situational causes and consistent information to dispositional ones, thus
reconfirming the stereotype.24 Social perceivers are also likely to misremember ideas
they generated based on observed behavior as observed facts rather than inferences,
confusing what they inferred with what they actually saw.25 This is one of the reasons
that eyewitness testimony is so notoriously unreliable: Gombrich and Lippmann were
correct that people often see what they expect to see, even if it is not what is actually in
front of their eyes.26 These cognitive functions of stereotyping are crucial to narrative
comprehension because they efficiently orient and focus the spectator’s attention on the
character’s most textually salient traits, which are especially important for the attribution
of character causality to narrative events.
Of course, sometimes we notice unexpected things that do not conform to a given
stereotype. Stereotype inconsistency may arise when an individual has attributes that
contradict those expected according to the stereotype. If you believe that Jews are rich, a
poor Jew might challenge your view. But we often reconcile such contradictions by
actively seeking to maintain, rather than revise or reject, the stereotype. We slot
exceptions into subtypes, effectively removing them from the stereotype category and
dismissing them from consideration.27 A poor Jew might be a recent immigrant from
Russia driving a New York City taxi, and thus a subcategory of “recent Russian
immigrant Jews” who are seen as different from the rest. Our cognitive structures are not
easily transformed, though over the course of several generations in 20th century
America, most ethnic stereotypes changed to some extent, so that Jews were more likely
to be stereotyped as ambitious in the second half of the century, whereas earlier they were
more likely to be seen as mercenary.28
Although stereotyping is a universal, the content of an individual’s stereotypes is
a product of experience and is variable according to the multiple influences of history,
21 One influential study is Patricia G. Devine, “Stereotypes and Prejudice: Their Automatic and Controlled Components” Journal of Personality and Social Psychology 56 (1989), 5-18.
22 Kunda, 317-325.
23 Ibid, 335.
24 Hamilton and Sherman.
25 Ibid.
26 A clearinghouse of information on this topic can be found at Gary Wells’ Eyewitness Identification Home Page (Wells is a leading researcher in this field): http://www.psychology.iastate.edu/faculty/gwells/homepage.htm.
culture, and sub-culture. In addition to changing over time, stereotypes also vary from
person to person and group to group. Although ethnic, racial, national, gender, and age
stereotypes are generally well known among members of a given population, they are not
all applied to the same degree or in the same fashion among all members. Moreover, the
stereotypes themselves vary among different groups, so that conceptions of male and
female may be quite different in one place than in another, or among members of various
sub-cultures, ethnicities, nationalities, ages, etc. Of course, cultural approaches to cinema
make much of this variation, assuming that members of different audience formations
interact with texts according to their differences. I don’t deny this facet of culturally-
determined difference in responses to narratives at all. This is clearly an instance in
which there is a natural explanation for cultural variation: it is in the nature of groups to
see themselves differently from how they see others, and it is not hard to imagine that
each in-group’s conception of various out-groups is a response to specific conditions of
the in-group’s environment. Among groups A, B, and C, it would make sense if groups
A and B hold differing stereotyped conceptions of C. If A and C are enemies, A might
see C as cruel. If B and C are rivals, B might see C as unfair.
The movies have furnished dozens of vivid stereotypes that have perpetuated
negative images of ethnic and racial minorities, such as Italian mobsters and Native
American savages. These are combinations of two kinds of types: social stereotypes that
27 Kunda, 384-391.
28 Ibid, 391.
pre-exist the movies and exist apart from narrative representation, and genre types that
are developed in cinema and the other media. In practice, these two types of types are
mutually reinforcing and interdependent, and it is widely believed that the stereotypes
that circulate in society are products of the media. But typing also functions in more
benign and ordinary ways, and is undeniably useful. Because stereotype is such a loaded
term, I prefer to discuss person categorization in cinema as typing and reserve the term
stereotype to refer to clear instances of prejudiced simplification or distortion.
As with all of the natural processes that drive characterization, typing works two
ways, in two directions: first are the bases of characterization that are our everyday
processes of social cognition, which we apply to the comprehension of the narrative;
second are art-, narrative-, cinema-, or genre-specific patterns of textual convention and
constraint that guide our experience. Each informs the other and depends on it.
The typing process is one of both confirming and disconfirming judgments. It is
often quite likely, especially in the highly artificial environment of a narrative, that a
person will defy some expectations as well as meet most of them. People who always act
as predicted may appear boring, and characters who always act as predicted may seem
much more so. But truly unpredictable people, and characters, are literally beyond
comprehension.
Social and Genre Types
I am proposing that characters are typed in two complementary ways, designating two
points of origin for person categories. One is in reality, the other in narrative
representations. Social types are person categories that precede narrative representation
and could exist independently of it. Genre types are person categories that are familiar
from a particular kind of story or from several kinds. An example of a social type would
be a woman; an example of a genre type would be a femme fatale.
Characters generally conform to social and genre types, but they also defy them.
Characters are rarely—if ever—nothing but type. Nor are they unique creations unlike
anyone you have ever met. Some categories of social identity inevitably become part of
the character’s dossier, such as their age, gender, sexual orientation, nationality, ethnicity,
occupation. In cases that a character’s gender, for example, cannot be identified, this
becomes a crucial bit of characterization, and this very thing can be considered a category
or type. In Boys Don’t Cry, the main character’s gender ambiguity is a defining trait. He
is characterized by membership in the category of transgendered. This distinguishes him,
for example, from the main character in Hedwig and the Angry Inch, whose gender
ambiguity is also a defining trait, but who is transsexual. It is very hard to defy
categorization altogether, even in films like Boys and Hedwig that thematize gender as an
unstable category that resists a binary understanding. We all belong in categories.
It might be tempting to read this defense of categorization as an implicit defense
of all categories as fixed, objective entities, or as a denial of the commonly held view (in
the humanities, at least) that categorization is an ongoing process that changes according
to history and culture. I am not claiming that categories are always stable and objective,
that they are immutable, air-tight, or easily defined. Some people disagree about the
existence of certain categories or about the membership of specific instances in particular
categories. Perhaps you consider Hedwig to be transgendered and I consider him
transsexual. This disagreement hardly invalidates the use value or efficiency of
categorization. I am merely claiming that it is in our nature to categorize, and that we do
so all the time, especially when encountering new people. Carving person categories in
two is a practical means of making sense of how this process works in cinematic
representation, but even this categorization could be done differently if the purposes or
the theoretical orientation were different. My designation of these two person categories
in cinematic characterization is not arbitrary, but it is a product of theorizing.
Genre types are characters specific to a certain kind of narrative. In some
discussions of genre in film criticism and theory, it is understood that a genre type is a
lesser form of character than a non-genre type.29 Genre films are contrasted against non-
genre films, with judgments of aesthetic value favoring the latter. The characters in genre
films are highly formulaic and predictable, and implicit in this judgment is the idea that
the filmmakers deserve little credit for any imaginative work in making the characters.
Western heroes are gunslinging cowboys on the side of right and law. Genre critics
29 This scale of valuation may be passé in academic film studies, but it lives on in the cinephile community surrounding festivals, art houses, and cinematheques, where saying that a character transcends genre is a kind of compliment.
might celebrate those directors whose western heroes rise above the ordinary level of
typicality (Ford, Hawks, Mann, Leone) but they do so at the expense of the ordinary
western and its characters, who are seen to make up an undifferentiated mass of genre
stuff. The general idea is actually correct: genre types share many traits in common, so to
some extent creating genre characters does not require much imagination. But the
characters in Ford et al. are not exempt from this description. The catch is that no
character is a carbon copy of any other, or of a prototype; all films, from B-movies to
Oscar-season prestige pictures, are made of a combination of conventional and original
ingredients.
This notion of genre types assumes all films to be generic to some degree, in some
fashion. There are highly specific cinematic genres, such as slasher films and screwball
comedies; there are genres that films share with novels and plays, such as thrillers and
farces; and there are genres that are so big as to include any conceivable narrative form,
such as the two super-genres identified by Patrick Colm Hogan, romantic and heroic
tragi-comedy, which he argues are universal narrative structures present in prototypical or
canonical stories.30 Types of narrative dictate expectations for types of character: the
typical protagonist in a love story automatically carries the trait of “potential romantic
partner;” the typical protagonist in a thriller does not. All of these levels of generality
may come into play in any given narrative.
30Patrick Colm Hogan, The Mind and Its Stories: Narrative Universals and Human Emotion (Cambridge: Cambridge UP, 2003), 98-121.
Just as all characters are to some extent types, so are most films and narratives,
and certainly the vast majority of independent films. Boys Don’t Cry is a love story too,
though it may not seem like an exemplar of the category. Films may belong to more than
one genre and may contain types of several different genre categories. This is not just a
matter of genre-hybrid experiments like musical-westerns or science-fiction-noir, but of
many ordinary films that are composites of genre material, including types. Indie films
are a genre if by genre we mean category of narrative texts. In film studies genre
typically means something more specific within an industrial and critical context,
demanding a particular iconography or emotional rhetoric, but in other media such as
literature genre has much looser connotations—for example, the novel is sometimes
referred to as a genre to distinguish it from the romance. There are types specific to
independent film too and actors who have become associated with them: Parker Posey’s
clever single girl (e.g., Clockwatchers, The Daytrippers), John Turturro’s beleaguered
everyman (e.g., Box of Moonlight, Barton Fink), James Spader’s sexually repressed loner
(e.g., sex, lies and videotape, Secretary). These types might not have the same range of
associations or the iconic status of the western gunslinger or the femme fatale, but that
just means that some types are richer and more vivid than others.
Of course, many indie films are also genre films in the more conventional sense.
Barton Fink is a movie about the movies, a genre that includes Sunset Boulevard, The
Bad and the Beautiful, Singin’ in the Rain, 8 1/2, Day for Night, and The Player. As
such, it has some of the characters typical of the genre: the writer, the producer, the actor,
etc. Of course, these are social types as well as genre types, but the image of the
bombastic, cigar-chomping Hollywood producer is a construction of Hollywood just as
the Native American savages and the Italian mobsters are. Indie films may be romantic
comedies, family melodramas, or thrillers, and have the characters typical of these
genres: romantic partners in comedies (e.g., Next Stop Wonderland, Kissing Jessica
Stein), parents, siblings, and children at odds with each other in family melodramas (e.g.,
Happiness, Pieces of April), and violent criminals in thrillers (e.g., Blood Simple,
Reservoir Dogs).
Some character types belong to more than one genre or to genres at very high
levels of generality. There is a type of minor character often seen in recent mainstream
American films and television dramas who might be called the Black authority figure.
She may be a judge or doctor or school principle. Her function is to place a racial or
ethnic minority as the face of respected power and authority. This is presumably a
gesture promoting multiculturalism, especially in films or on programs with few non-
white characters. The Black authority figure is typically reasonable and intelligent, but
also tough and uncompromising, especially in matters of morality. Whether the narrative
genre is courtroom suspense, soap opera, or fantasy, the type is consistent. (As a minor
character, of course, there is less opportunity for exploring other dimensions of the type
or of ways of defying it.)
A given character may be an example of more than one type. In go-for-it sports
movies like Rocky, The Karate Kid, and Blue Crush, the main character is both the
competitor type (driven and hard-working, not the most talented, often working class)
and the romantic lead type. The idea of a competitor type and a romantic lead type are
complimentary but not mutually inclusive; nor are they exclusive to one kind of narrative.
The go-for-it protagonist type is a combination of these two.
Genre types are closely related to social types, with which they sometimes
overlap and from which they must often be derived. For example, many male or female
genre types are based on male or female social types. This is not necessarily to say that
they are based on real individual men or women, because social types may be no more
veridical than genre types. Social types—including, of course, stereotypes—are
themselves typically partial fictions, and genre types piggyback on them. There may be
no real femmes fatales, but the women in film noir are based to some extent on social
types of women that existed in the 1940s and 1950s, which themselves were products of
partially accurate and partially inaccurate generalizations about real women.
This also works the other way around. Many social types would be much poorer,
in terms of the range of associated traits and the expectations generated by them, without
types from narrative representations, i.e, without genre types. This is especially true of
occupation types, like police officer, doctor, car salesperson, and the various characters in
legal dramas. Most people meet many more fictional surgeons, judges, and detectives
than they ever do real-life ones, and when they meet real-life ones their expectations are
influenced by their exposure to media. Hence the phrase, “You’ve been watching too
many movies!” The interest most media scholars have in stereotypes is aimed at
examining and counteracting this process whereby people mistake the types encountered
in representations for information about real-world people and groups.
In any given representation, it may not be feasible to discriminate among social
and genre typing. The distinction between these categories is theoretical, not practical.
In some cases, there are clear film-specific types (e.g., the maniacal genius villain of
Hollywood action films is a type rarely encountered in real life) and types that clearly
originate in society (e..g., most age, gender, ethnic, national, and racial types). Yet these
social types are informed by media representations, so that adolescents’ conceptions of
adolescence comes not only from experience in the world but also experience in the
world of representation. It works the other way, too. The maniacal genius villain is
typically male, is almost always an adult and a criminal, and is often a foreigner. These
are all social types that together make up aspects of the genre type.
Typing in Process: Passion Fish
Characterization by type is significant for understanding character psychology for a
number of reasons. Most basically, the range of associations quickly and automatically
generated by typing—which we might think of as a typically fast and dirty kind of
cognition—both opens up and constrains the range of inferences to be generated. These
may be folk psychology inferences about intentional states. Some of the associations that
come with a given type often refer to goals, desires and plans. We assume that lawyers
try to win cases, that doctors try to heal the sick, that car salesmen try to sell cars; we
attribute the causes of behavior to these desires based on our typing. But since typing is
part of what defines and characterizes narrative situations, emotion inferences are also
dependent on typing. To take but the most obvious example of typing and emotion,
women are often assumed to be “more emotional” than men, an idea that suggests that
many emotions are both more likely to occur in women and that they are more likely to
feel them strongly. Although this may seem empirically dubious, such types are socially
powerful and filmmakers utilize them in creating characterizations.
Typing often functions to establish a situation, and the other aspects of social
cognition generally often proceed with some type attribution already in mind. This is
because of the automaticity of characterization by type; it is virtually impossible to
represent a person in cinema absent some typing information. The only way I could think
to do so would be by having a character who is only referred to by name, never seen, and
whose name has neither denotative nor connotative meanings suggesting a gender,
nationality, ethnicity, age, etc. The character would have to be characterized so
minimally as to generate too few associations or causal relations to be typed into any
group. Such a character would be of minimal interest to spectators, but as soon as any
more of him or her was represented, the character would automatically become an
instance of one or more types.
This automaticity ensures that typing generally (though not always) precedes all
of the other aspects of characterization temporally in the comprehension of narrative.
Direct characterization and inferential, indirect characterization alike occur on a moment-
by-moment basis. While there are more global inferences covering large-scale goals (like
the overarching goals that lead from the opening scenes to the ultimate denouement in
many narratives), and while there are emotions (or more likely, moods) that seem to
pervade the whole representation, most folk-psychology inferences and most displays of
emotion occur in circumscribed instances, in reaction to specific events.31 These interior
states last for short periods of time. You are angry when someone insults you, but the
emotion typically lasts for only a few seconds or minutes.32 But typing is more of an
ongoing process than emotion or folk-psychology inference. Once typed a certain way,
the categorization informs comprehension of everything to do with the character and his
situation unless some contradictory information comes along. Furthermore, while typing
informs the other aspects of the process of characterization, the other aspects are not
likely to inform typing very often, though some film do strategically invite us to adopt
one type attribution only later to demand that we revise or reject our judgment on the
basis of developments in the narrative situation. This kind of development is a function
of the “primacy effect” which I discuss in Chapter 6. This is the principle that the first
information that a film gives us about a character biases our interpretation of all of the
subsequent information.33 We shall see that by exploiting the primacy effect, John
Sayles’s Passion Fish is able to present distinct character types during its initial
31 The notion of moods pervading a whole narrative is discussed in Greg M. Smith, Film Structure and the Emotion System (Cambridge: Cambridge UP, 2003), 42-51.
32 Ibid.
exposition only to complicate our understanding of characters when later on the film
introduces other ways of typing them that complicate our initial assessments.
In Passion Fish, the two main characters are May-Alice (Mary McDonnell), who
has recently been left paralyzed by a car accident, and the nurse she hires to look after
her, Chantelle (Alfre Woodard). We infer based on these roles that May-Alice is angry
and sad, and that she hopes Chantelle will be able to help her. As for Chantelle, she is
introduced after several other nurses are made so exasperated by May-Alice’s
recalcitrance that they quit. We see her not only as a nurse but as the last nurse on the
agency’s list to send to May-Alice, the lowest on the pecking order. We infer that she
needs the job, that she wants to please her employer. As a nurse, we infer that she is
compassionate and that she sympathizes with the sick and infirm.
May-Alice’s and Chantelle’s mental states are not represented directly, yet both
characters are vivid and richly detailed. May-Alice’s anger and fear, her resistance to
therapy, her strong individualism, and her propensity to alcoholism are all explained by
her typing as an independent, well-off white woman and as the victim of a catastrophic
accident. Her quick temper and substance abuse may also by regarded as typical of May-
Alice’s show-business profession: she is an actress on daytime television. Chantelle’s
determination to do her job right, her preference to keep private her personal life, her
frankness and general skepticism are all explained by her typing as the nurse, the
employee, the working black woman. It doesn’t work the other way around, though.
33 Meir Sternberg, Expositional Modes and Temporal Ordering in Fiction (Bloomington: Indiana UP, 1978), 38; Bordwell, Narration in the Fiction Film, 38.
Determinations of type are not initially products of inferences about intentional states and
emotions. That is to say, you don’t first recognize that May-Alice is angry or that she
wants some wine and move from those data to a type-inference referring to her social
status. This is not to say that no emotion or intentionality data are informative in relation
to typing. A frequently angry character may be typed as easy to anger, and a character
who often believes falsely that people are conspiring against her would be typed as
paranoid. But note that these would be types that refer to mental states themselves.
Types that refer to social realities or genre conventions are harder to conceive as products
of represented or inferred intentional and emotional states. That is, you don’t ordinarily
type someone as a gangster, an Indian, a police officer, a child, a romantic partner, etc.,
on the basis of character psychology. Ordinarily, it works the other way around: you get
character psychology on the basis of the typing.
Typing is also significant because of the expectations it generates about narrative
design, specifically about character trajectories (a.k.a. arcs, see chapter 6). In Passion
Fish, the relationship between the two main characters is established early on as
unbalanced and a bit contentious. At the same time, the roles of caregiver and care
recipient inform all of their interactions. The audience expects that Chantelle will help
May-Alice rehabilitate and learn to get along in the world without the use of her legs. At
the same time, May-Alice is the employer and Chantelle is the employee. We expect
Chantelle to try to satisfy May-Alice and to heed her wishes. And while May-Alice is a
white woman of independent means, Chantelle is African American and has to work for a
living. We also expect their racial and class differences to come into play.
In any good story, some expectations are met and others are not. The characters
in independent cinema are often more surprising than those in Hollywood films. They
may have more contradictory traits or more complex combinations of traits. This
frustrates the process of typing but also adds interest to it. Passion Fish is a fine example
of this tendency. The central trajectory of the main characters’ relationship is a narrative
of reversal: at first Chantelle takes care of May-Alice, then it becomes apparent that May-
Alice is also taking care of Chantelle, and that she is strongly committed to being
Chantelle’s friend and helper. This reversal is the product of the narrative’s most
important surprise: we learn that Chantelle is recovering from drug addiction and that she
has a daughter who has been taken from her custody Keeping her job and doing well at it
are requirements for Chantelle keeping her on the path to full recovery and to getting her
little girl back. . Thus a parallel is established among the women’s traits that transcends
their typing and brings them closer: both are addicts, both have suffered through life-
changing traumas, and neither can afford to fail at recovery.
Critics have commended Passion Fish for taking the characters and situations of
contemporary melodramas (such as soap operas and disease-of-the-week movies) but
representing them with far more subtlety, nuance, and complexity than is conventional.34
This referencing is accomplished simply by the choice of subject matter, but also by
interpolating texts that reference these highly emotional, conventionally feminine genres,
such as soap operas and the melodrama What Ever Happened to Baby Jane?; it is also
significant that May-Alice is a soap actress. To some extent this compliment is
backhanded and sexist, as it gives Sayles credit for making the best of a bad genre while
implicitly it questions his choice of material and assumes that a film about female
friendship is automatically of less interest than one about baseball players or coal miners,
to name two other Sayles subjects (in Eight Men Out and Matewan, respectively). At the
same time, it is an implicit recognition that by subtle manipulations of trait-type
expectations, Sayles is able to surprise us with characters whose defiance of conventional
expectations lends the narrative credibility and emotional resonance that outdoes many of
his other, more didactic films.
In part, Sayles is able to achieve this because, unlike most of his films, Passion
Fish is centered on a dyad rather than an ensemble, and so achieving nuanced, complex
characters is a different kind of challenge than in films like Return of the Secaucus 7,
Matewan, City of Hope, Eight Men Out, and Sunshine State. Each character is on screen
for far longer in Passion Fish than in any of these other films (I return to discussing
ensemble films in Chapter 6). Each of the characters in Passion Fish belongs to more
types; in turn each one confirms or defies type expectations in various ways. In Passion
Fish, Sayles accomplishes with two characters what in other films he does with a half
34 For example, see Roger Ebert, “Passion Fish,” Chicago Sun-Times, 29 January 1993, URL: http://www.suntimes.com/ebert/ebert_reviews/1993/01/839268.html. In an interview, Sayles confirmed this reading of Passion Fish as taking a soap-opera situation but treating it a way that would defy the audience’s genre expectations. Trevor Johnson, “Sayles Talk” in Jim Hillier (ed.), American Independent Cinema: A Sight and Sound Reader (London: BFI, 2001), 215-219.
dozen or more: presents characters of multiple dimensions. This is a matter, to a large
extent, of spinning each character into a variety of types, some of which fit better than
others into the total characterization.
While the main characters in Passion Fish are genre types familiar from women’s
pictures, and while the initial narrative situation may sound clichéd, the film itself was
never taken to be a soap-opera or disease-of-the-week movie. This was not the category
that most critics and spectators used to identify it when it was released in 1993. More
likely, it was understood as an independent film or a John Sayles film, or more simply as
a drama and a film about two women. There are elements of the characters in Passion
Fish that are typical of all of these categories. The patient-nurse dyad is typical of
disease movies. The strong-willed individual is typical of Sayles and of indie films. So
is the character coming to terms with her identity. Passion Fish has none of the truly
iconic Hollywood genre types seen in some indie films, such as prohibition-era gangsters
in Miller’s Crossing. Its characters still instantiate genre types, though they are not as
rich and vivid as genre types as those in films by genre-obsessives like the Coen brothers,
for whom genre-typicality is a central preoccupation of characterization.
In general, independent films have a distinctive approach to typing, based on an
implicit suspicion of the notion of fixed categories and on an interest in exploring issues
of social identity, as I discussed in Chapter 1. This interest is especially marked in a
writer/director like Sayles, whose progressive politics informs all of his storytelling. In
independent cinema, the process of typing is complicated by the incompatibility of
multiple types and by the introduction of surprising types that occasion reevaluation of
the character. We have seen that in Reservoir Dogs this is accomplished when Mr.
Orange is revealed as an undercover cop. Of course, mainstream narratives are also full
of surprises, and the surprises in Reservoir Dogs are not all that radical (or for that matter
surprising—we probably suspected Mr. Orange from pretty early on). The distinction in
indie films is that a wider and more diverse complement of social and genre types is
introduced for main characters, and with more types comes greater potential for friction
among them. It is partly from this effect that characterization is made prominent as a
feature of the text, and that character is made to seem more interesting.
But as indicated in Chapter 1, there is also an tendency in independent cinema of
making characters clear emblems of a particular social identity, such as a racial, ethnic, or
gender identity. Much independent cinema, and especially independent cinema that is
politically engaged, is driven by multiculturalism, by a desire to explore the specificity of
particular cultural and sub-cultural experiences. In this vein, independent films are often
just as concerned with exploring the dimensions of a single identity and the
contradictions inherent within it as with multiplying types as in Reservoir Dogs and the
Coen brothers’ genre-bending films. In Passion Fish, both of the main characters’
cultural identities (Cajun and African-American) are made prominent, and both are
explored through an effort at recognizing their complexity. In May-Alice’s case, there is
also a concern to understand a character’s ambivalence with her culture.
These approaches to typing, both the multiplication of types and the exploration
of specific cultural identities, are partly achieved through the film’s narration, in
particular its delayed and distributed exposition, whereby events preceding the beginning
of the plot are referenced later on. It is also partly a function of changes undergone by
the characters. That is, some of the type-multiplication and type-exploration is a feature
of information being introduced only selectively, and some is a feature of new types
emerging either to complement or to supplant old ones. Essentially, the more we learn
about the characters, the more we see them as multidimensional and contradictory.
Characterization in Passion Fish is not an additive process of accreting traits and types,
but a transformational process wherein each new type introduced can effect a reshaping
of the character, a new conception and understanding of her.
May-Alice is the one introduced first, and the film begins as her story. She is
introduced first as a white, American woman (about mid-thirties, certainly neither young
nor old) and a spinal injury patient. We subsequently learn that she is a soap-opera
actress, that she has no immediate family, that she is from the Louisiana Cajun country,
though she has no trace of an accent, and that she is fairly well off. She then takes on the
role of employer to a series of live-in nurses, none of whom can stand her unpleasant
demeanor and hostile behavior. As the first half of the film continues, she takes on more
types as we observe more behavior and meet more secondary characters: alcoholic,
amateur photographer, old friend, colleague, potential romantic partner. Some of these
new type categories stand in a contradictory relationship to other, earlier established ones.
Her romantic relationship with an old married friend, Rennie (David Strathairn), is
unexpected given her inability to experience sexual pleasure and her earlier, angry
statements about this inability. Her photography is represented as a creative outlet that
replaces her soap acting, which in contrast seems less personal. She begins to make
peace with her Cajun roots, which we learn she had earlier tried hard to repress and
forget. Instead of having no family, Rennie and Chantelle become a surrogate family.
May-Alice’s recovery arc is a reversal, substituting types given a positive valence for
negative ones.
Chantelle’s characteristics are based on a different constellation of type
categories. First, she is a black woman, about May-Alice’s age, working for May-Alice
as her nurse. Quickly we discover that she is not like May-Alice’s other nurses in one
important respect: she is more committed than they are to doing her job well no matter
what obstacles she faces. We assume that she has no significant other and no children
because none are referenced in any dialogue and none appear at the house. When she
begins a romantic relationship with Sugar LeDoux (Vondie Curtis-Hall), a black cowboy
with several children, it seems clear that although Chantelle is single and available, she is
very wary of getting involved with anyone. But then her ex-husband appears, and then
her father and daughter. Delayed exposition again surprises us. We learn that she is a
recovering addict, a daughter, a mother. We learn that she grew up well off, not poor, in
Chicago. Chantelle’s recovery arc is characterized by admissions and revelations, a
coming to terms with who she is.
Characterization in Passion Fish relies on a strategic order of exposition for its
effects. The primacy effect emphasizes certain type assignations, but later on a “recency
effect” balances the primacy effect by demanding a revision of our assessment of the
characters.35 Sayles saves May-Alice’s creative and romantic characteristics for later in
the film, instead beginning with her as an angry accident victim who has to give up a
career and a busy life in New York for an isolated one in Louisiana. Her only
interpersonal relations are with members of the health care professions. With Chantelle,
it’s a different approach. Sayles saves her most personal material, her own trauma, for
later in the film. The effect is to put Chantelle on equal footing with May-Alice in terms
both of suffering and of recovery, but to save this symmetry for the latter part of the film.
For each of the main characters, a relationship could derail or spur her recovery,
and so each approaches her male counterpart with a combination of desire and
trepidation. This reinforces the parallel between the main characters. As well, it is
significant that the film ends with the women together, but without closure in the
romance sub-plots. The film has many secondary characters, most of whom appear in
only one or two sequences, and two of whom, Rennie and Sugar, are more recurring and
significant. In general, secondary characters have fewer characteristics than primary
ones, and as such often seem simpler and more straightforward.
35 Recency effects are not nearly as strong as the primacy effect, but it is of course possible for later-encountered information to bias earlier-encountered information in certain cases of “strong contrast effects.” Richard Nisbett and Lee Ross, Human Inference: Strategies and Shortcomings of Social Judgment (Englewood Cliffs, NJ: Prentice-Hall, 1980), 172.
In Passion Fish as in many films, one function of secondary characters is to refine
and clarify the characterization of the leads. As I shall discuss in Chapter 6, Characters’
are understood in comparison to other characters’, so that a character who seems racially
insensitive (as we judge May-Alice’s old acquaintances) may make May-Alice seem
enlightened by comparison. This process that is especially significant in the ensemble
dramas in which Sayles generally specializes. When these friends come to visit, we also
get a glimpse of the milieu of her upbringing. She is much less “Southern” in her accent
and mannerisms, and she relates to Chantelle in a much less patronizing way (the old
friends assume her to be a servant), suggesting that racial prejudice was a problem among
local whites that May-Alice either avoided or overcame. A similar comparison occurs
when May-Alice’s former colleagues from the television show visit. We notice how she
is one of the actresses, sharing their camaraderie but a bit bitter over her part being taken
by someone else. This sequence also affords the opportunity to compare Chantelle with
Dawn, an African American woman of roughly her age, who also grew up in Chicago
(Angela Bassett). We learn here that Chantelle grew up well off (better off than Dawn),
which sparks our curiosity about how she wound up working as a nurse so far from
home.
The love-interest characters, Sugar and Rennie, are each portrayed as slightly
oddball. Sugar is a black Louisianan who speaks French, trains horses, wears a cowboy
hat, and comes off as a kind of Lothario, but also dotes on his adolescent daughter. He is
typical for indie films in his atypicality. He certainly seems to Chantelle to be an
incongruous combination of traits. Rennie is married to a bible-thumper who forbids
television and music in their home. He has many children with her, yet he spends long
periods away from them. He does not share their piety. But he does seem to give some
credence to local superstitions. When he takes Chantelle and May-Alice out in his boat,
he catches a fish for their lunch and as he is gutting it he tells of how the locals believe
you can see your future by looking in its belly. In it are two passion fish, and he has each
of the women make a wish, according to custom, of who they want for a lover. Rennie
seems to occupy a middle position, neither adopting all of the traditional local beliefs, nor
being a devout Christian. Like the other characters, he is caught in between differing,
contradictory social roles. He seems romantically available to May-Alice, but also seems
committed to his family. It is typical of the film that it does not make this contradiction
into a source of real dramatic conflict, instead allowing him to occupy both roles. The
effect of vivid, life-like characters in Passion Fish is achieved by just this sort of device
of the multiplication of types: Rennie is both the family man and the lover, the believer
and the skeptic, someone who feels a sense of belonging to place and someone restless to
escape.
As spectators, we approach the film armed with types, and the film is all too eager
to oblige us, offering characters who fit more of them than we expect. This is the effect
of freshness that the film’s admiring critics identify. Ultimately, by combining these
various types, by using the techniques of delayed exposition and surprise, and by
introducing the secondary characters who contrast with or complement them, the
characters are given greater specificity than one would expected in a mainstream
Hollywood film.
Conclusion
As with all of the techniques of character construction, typing depends on spectators to
seek coherence, to make things fit together as best they can. This means two things.
First, typing needs to cohere with the other techniques of characterization. It must inform
and be informed by what we know about characters’ mental states and observed
behaviors. Second, typing needs to be internally consistent—to have as little
contradiction and as much consistency as possible. The search for coherence is a global
desire of narrative comprehension, just as it is a primary sense-making strategy in social
cognition, 36 and categorization is a way of making things cohere, or appear to cohere.
Avant-garde art often seeks to frustrate our desire for this, but independent and
mainstream cinema, each in its own way, more often seek to exploit it. Categorization
can be seen as one strategy of making the world seem to hang together: rather than seeing
an infinite mess of disparate entities, we experience reality in meaningful chunks with
clear patterns of interrelation among them.
Just as films are sorted into genres, characters are sorted into types, and these two
modes of categorization are clearly related by correspondent categorizations. We simply
36 For an argument connecting social cognition with coherence-seeking, see Paul Thagard and Ziva Kunda, “Making Sense of People: Coherence Mechanisms” in S. J. Read & L. C. Miller (Eds.), Connectionist Models of Social Reasoning and Social Behavior (Hillsdale, NJ: Erlbaum, 1998), 3-26.
could not make sense of cinema without these categories. Some theorists of genre and
genre types criticize the categories used in cinema for appearing to be natural, immutable
or fixed, while beneath this surface of coherence is a welter of contradiction.37 While
there may be some truth to this notion on the level of appearances, on the level of the
function of types it fails to appreciate the cognitive utility and the aesthetic potential of
categorization. The significance of cinematic types is that that they are useful and
meaningful, and that it is natural that we would create them.
37 Rick Altman, Film/Genre (London: BFI, 1999).