MYSTIFYING MOVIES

NOËL CARROLL

MYSTIFYING
MOVIES
FADS & FALLACIES IN
CONTEMPORARY
FILM THEORY

COLUMBIA UNIVERSITY PRESS

New York 1988
 Columbia University Press
New York Oxford

Copyright © 1988 Noël Carroll

All rights reserved

LIBRARY O F CONGRESS
Library of Congress Cataloging-in-Publication Data
Carroll, Noel (Noël E.)
Mystifying movies : fads & fallacies in contemporary film theory /
Noël Carroll,
p. cm.
Bibliography: p.
Includes index.
ISBN 0-231-05955-8
1. Philosophy. 2. Film criticism. 3. Motion pictures. I. Title.
PN1995.C356 1988
791.43'.01—dcl9 87-36448
CIP

The Andrew W. Mellon Foundation, through a special grant,

has assisted the Press in publishing this volume.

Hardback editions of Columbia University Press books are

Smyth-sewn and arc printed on permanent and durable acid-free paper.

©
Printed in the United States of America

p 10 9 8 7 6 5 4 3 2 1
c 10 9 8 7 6 5 4 3 2 1
Dedicated to Annette Michelson
 CONTENTS

Acknowledgments ix

INTRODUCTION 1

I. PSYCHOANALYSIS: M E T Z A N D BAUDRY 9
Jean-Louis Baudry and " T h e Apparatus" 13
Metz's Imaginary Signifier 32
Metz's Daydreams 44
Conclusion 48

II. MARXISM A N D PSYCHOANALYSIS:

T H E ALTHUSSERIAN-LACANIAN PARADIGM 53
Althusser and Ideology: Marxism and the
Invocation of Psychoanalysis 56
Lacan and the Construction of the Subject 62
An Initial Questioning of the Presuppositions of
the Althusserian-Lacanian Paradigm 73
The Explanatory Power of Ideology 84

III. T H E C I N E M A T I C IMAGE 89
Illusionism 90
Photography and the Cinematic Image 106
Perspective 127
An Alternative View of the Cinematic Image 138
Vili CONTENTS

IV. NARRATION 147

Enunciation 150
T h e Internal S t r u c t u r e and F u n c t i o n of Narrative 160
An Alternative Account of Movie N a r r a t i o n 170

V. CINEMATIC NARRATION 182

Suture 183
An Alternative A c c o u n t of Cinematic N a r r a t i o n 199
T h e Power of Movies 208
A C o n t r i b u t i o n to the T h e o r y of Movie Music 213

C O N C L U S I O N . Problems and Prospects of

C o n t e m p o r a r y Film T h e o r y 214

Notes 235
Index 259
 ACKNOWLEDGMENTS

I H A V E dedicated this book to Annette Michelson who has been my

teacher—both formally and informally—for nearly two decades. In fact,
this book grew out of an opportunity generously provided by Annette
Michelson when she invited me to review Stephen Heath's Questions of
Cinema for the journal October. She spent a long, hot summer in Vermont
listening to me work out my ideas about contemporary film theory. And
though she often strongly disagreed with me, she graciously shared her
wide knowledge of film and the French intellectual milieu, enabling me to
clarify my arguments. She has also made helpful comments about my
treatment of Metz, my responses to Heath, and my thesis concerning the
power of movies. She has always managed to be critical and encouraging,
even about views of mine of which she is extremely skeptical; a better
mentor would be hard to find.
P. Adams Sitney, Tony Pipolo, Stuart Liebman, Leger Grindon, Amy
Taubin, Gerald O'Grady, Bruce Jenkins, Tony Conrad, Douglas Gomery,
Ted Perry, Peter Lehman, Bill Luhr, Berenice Reynaud, Kristin Thomp-
son, Johnny Buchsbaum, Jerry Rabkin, Raymond Carney, Michael Ryan,
Robert Sklar, Vance Kepley, Janet Staiger, David Rodowick, John Belton,
Carl Plantinga, and Daryl Davis have heard or read parts of this text and
have made careful comments and supported the project in various ways.
Perhaps the person in cinema studies who has been most influential in
the progress of this book has been David Bordwell. For nearly a decade, we
have sustained an intense dialogue ranging, I conjecture, over every topic of
film theory. He has forced me to reconsider my interpretations of and
arguments against contemporary film theory on numerous occasions. He
has shared his vast research in film history with me, saving me more than
X ACKNOWLEDGMENTS

once from a historical blunder. He and I share a conviction that cognitive

models of film theory have advantages that the reigning psychoanalytic
models lack. He has pursued this topic in greater depth than I have; I wish
him good speed.
Through Wesleyan's splendid Center for the Humanities, I have had the
opportunity to discuss my research with scholars outside my own field,
including Richard Vann, Hazel Carby, Michael Denning, Hubert O'Gor-
man, Christina Crosby, Richard Slotkin, Susan Foster, Richard Ohmann,
Richard Stamelman, Leo Lensing, Gayatri Chakravorty Spivak, Rena
Grant, Bruce Greenwald, Marylin Arthur, Neil Lazarus, and Ellen Rooney.
Some of these people will be stunned at finding their names in this acknowl-
edgment; but their criticisms, I feel, were both genuine and productive.
Most important, in this regard, are Elizabeth Traube and Khachig Tolol-
yan, who in my first semester at Wesleyan started a popular culture study
group that facilitated a great deal of lively and serious debate in a most
congenial atmosphere. Both Betsy and Khachig have also offered extensive
responses to my writings, from which I have benefited greatly. David
Konstan, late of Wesleyan, now at Brown, read large parts of the manu-
script and made pointed suggestions and comic asides.
I have many philosophers to thank for comments they have made after
hearing or reading parts of this manuscript. They include Marx Wartofsky,
Joe Margolis, Ted Cohen, Bruce Vermazen, Roger Shiner, Alex Sesonske,
Flo Leibowitz, Dale Jamieson, Richard Eldridge, Mary Devereaux, Allan
Casebier, George Wilson, Allan Weiss, Kevin Sweeney, Don Crawford,
Jack Glickman, Laurent Stern, Evan Cameron, Brian Fay, Joel Snyder, Jay
Bachrach, Chris Gauker, Annette Barnes, Sue Cunningham, Gary Shapiro,
Sam Kerstein, and Arthur Danto. George Dickie, Paul Guyer, Peter Kivy,
and the late Irving Thalberg, Jr., read the earliest stages of this project and
provided more than ample feedback. Stanley Cavell discussed my debate
with Heath with me insightfully, and Ian Jarvie, when I see him, has let me
bend his ear mercilessly.
My original editor at Columbia, Bill Germano, was immensely enthusi-
astic and expeditious. Jennifer Crewe and Joan McQuary have continued in
the same spirit.
Most of all, I have Sally Banes to thank for sharing her ideas, her life, and
her computer with me during the grumpy times in which such books are
composed. She not only tolerated the superstitious rituals I instituted while
writing, but was willing to examine critically the results as if a sane person
had produced them. Neither she nor anyone above is responsible for the
flaws in this text. I suppose I am.
 INTRODUCTION

T-L HIS is a book concerned with certain central topics in contemporary

film theory, a term of art by which I mean to designate film theory that
proceeds within a semiological framework, indeed, most particularly,
within a semiological framework amplified by Marxism and psychoanalysis.
"Contemporary film theory," in this book, stands in contrast to what might
be thought of as "classical film theory," that is, film theory prior to the
advent of semiology.1 Though this sounds like a strict chronological dis-
tinction, it is not. For works in the classical tradition, by theorists like
Stanley Cavell and V. F. Perkins, have been written since semiology arrived
on the scene. Rather, for our purposes, the mark of contemporary film
theory is a positive commitment to the dialogue opened by the rise of film
semiology, an event which, for convenience, we might date as vaguely in
the vicinity of 1966-1967 when Christian Metz wrote the essays that
compose his Essais sur la signification au cinéma.2
Contemporary film theory, conceived of as semiologically derived film
theory, moreover, has two stages: a first stage in which Saussurean linguis-
tics provided the dominant model, followed by what is sometimes called
the second semiology, in which Marxism and psychoanalysis became the
preferred conceptual tools. Contemporary film theorists do not regard this
transition as a sharp break because the brands of Marxism and psycho-
analysis they bought were themselves influenced by Saussurean linguistics.
From the vantage point of the reception of contemporary film theory in
America, the heyday of the first semiology was brief. Shortly after Metz's
earlier, Saussurean-derived essays on film theory were translated into En-
2 INTRODUCTION

glish, his more psychoanalytically oriented work began to appear along

with essays in the influential British film journal Screen which towed a
strident Marxist-psychoanalytic line. Since the early mid-seventies, at least,
contemporary film theory has primarily developed along the lines of the
second semiology. And it is the second semiology—an amalgamation of
psychoanalysis, Marxism, and semiology—with which this book is almost
exclusively preoccupied.
In the United States, Marxist-psychoanalytic-semiology is the dominant
form of film theory and it has enjoyed this position for at least a decade.
This is not to say that everyone in film studies, stateside, practices this
variety of film theory, but only that it represents the most vocal and most
influential school of thought in the field. It is the second semiology that has
attracted many of the younger and most ambitious scholars in film studies.
It provides much of the jargon for major academic film journals and film
conferences; and most books published in film theory by academic presses
are in its idiom; there are several anthologies devoted to its teachings, such
as Philip Rosen's Narrative, Apparatus, Ideology.3 In the United States,
Marxist-psychoanalytic-semiology is the film theory of the academic estab-
lishment. This book is designed to explore several of the central concepts,
presuppositions, and doctrines of the dominant film theory for the express
purpose of challenging them. Specifically, I shall argue that in their at-
tempts to show how movies purportedly mystify spectators, contemporary
film theorists, in fact, mystify our understanding of cinema.
The reception of contemporary film theory in America is an epicycle
within a larger historical process which involves the rapid expansion of film
studies as an academic field of study in the seventies. This expansion, in
turn, seems initially to have been a response to a resurgence of interest in
film in the sixties, marked, in part, by the rediscovery and reaffirmation of
the masterpieces of the Hollywood cinema. One might explain this by
noting that the movie boom of the sixties roughly corresponds with the
coming of age of the baby boom generation, a generation brought up on
old American films on TV and predisposed to take those beloved objects
seriously. The baby boomers supplied the enrollment for increasingly
larger film appreciation classes and, then, programs, as well as, in time,
supplying the faculty to those programs.
To a great extent, the French provided the most important intellectual
role model for the development of film appreciation in America in the
sixties. They had already extolled the glories of Hollywood cinema in such
journals as Cahiers du Cinema and that sensibility, as well, gained persuasive
force insofar as it was incorporated into the vital film movement known as
INTRODUCTION 3

the French New Wave, which comprised such artists as Godard, Truffaut,
Chabrol, Rohmer, et al. From the perspective of theory, the Americans
benefitted from the adoption of two conceptual tools developed in France:
auteurism and the theory of realism propounded- by André Bazin.4
Auteurism was the doctrine that maintained that films were commend-
able insofar as they manifested the distinctive personality of their maker
(author/fli«ic«r) who, in general, was identified with their director. A
routine Hollywood genre film, thus, could be a great film if it bore the
individual stamp of its director—e.g., of a John Ford, a Howard Hawks, an
Alfred Hitchcock. Indeed, for some, Hollywood films had the best chance
at greatness for it was presumed, mistakenly I think, that the Hollywood
studio system was a great obstacle to the personal expression of the master
filmmakers, and, in consequence, that their overcoming this obstacle made
their achievements all the more meritorious.
Auteurism was not really much of a theory; it was actually a form of
connoisseurship. However, it did have the virtue of supplying one with a
way to organize an unwieldy field of study in terms of a canon of great
directors, and with a research program—view all the films of the great
directors, identify their recurring features, and stay on the lookout for
undiscovered great directors. This may not sound like much; but it was a
way to at least start film studies during its period o f rapid expansion.
With Bazin, Americans encountered a sophisticated view of the ontology
of film which was also connected to a powerful account of the evolution of
film history. Bazin held that the film image was an objective re-presentation
of the past, a veritable slice of reality. Critically, this disposed him to
applaud realist cinema, notably that associated with the use of deep-focus
cinematography and with filmmakers such as Renoir, Welles, Wyler, and
the Italian Neo-Realists. Bazin was useful to the emerging field of film
studies in numerous respects. First, he showed people a way to think of film
stylistically rather than simply thematically. Moreover, he proposed a way
of dividing up film history stylistically, and, in the course of this, he
introduced the notion that there was a classical style of filmmaking, an idea
that still fuels much film theory. And, as well, Bazin offered a narrow
category of film realism which was fruitful for contrasting film styles and for
rediscovering forgotten directors (namely, those who used deep focus
cinematography).
Though the contributions of auteurism and Bazin were enabling, they
also had various shortcomings. Conceptually, auteurism reversed the crit-
ical order of things. Typically, we identify great artists by their master-
pieces; but auteurism seemed to imply that a work was a masterpiece if it
4 INTRODUCTION

was by a great artist, i.e., that a film by Preston Sturges, no matter how bad
it seems, is a masterpiece because it is by Preston Sturges. Similarly, BazirTs
account of the ontology of film may very well be incoherent5 and his
advocacy of deep-focus realism was parochial. Thus, by the early seventies,
American film studies was ripe for a new approach to theory.
Moreover, along with the conceptual liabilities of auteurism and the
Bazinian approach there was also the feeling that these projects were
politically insensitive. The expansion of American film studies in its early
stages, of course, corresponds to the political awakening that occasioned
the Vietnam War. Film scholars, like scholars in other fields, wanted to
make their research relevant to political life, and this project, furthermore,
was eminently plausible since films do seem somehow connected with
ideology. Auteurism and the Bazinian approach, however, were narrowly
aesthetic in their focus; whereas film scholars wanted theoretical tools that
would facilitate political engagement.
Again film scholars turned to the French from whose highly developed
film culture they had acquired the auteurist and Bazinian models. This time
around the reigning film theory was a form of semiology inspired by
Ferdinand de Saussure's Course in General Linguistics. As was already re-
marked, Christian Metz was the leading proponent of this mode of theoriz-
ing in film. Undoubtedly, part of the attractiveness of this style of theory
was its pretensions to being scientific. Metz was adopted by Americans at a
point when linguistics enjoyed special notoriety—when the name Chom-
sky was a household word (i.e., in academic households). Thus, it is no
surprise that a nascent field like cinema studies, with anxieties about its
legitimacy, should hitch its fortunes to a star. However, it seems to me that
an even more important factor in the adoption of semiology by young
American film scholars was their sense that it could be politically useful.
It may sound mysterious to claim that a theoretical model based on
linguistics could be thought to be relevant to political analysis. However,
the reasoning seems to have been that whereas someone like Bazin treated
the cinematic image as if it were a virtually natural phenomenon, a lin-
guistically oriented semiology at least regarded cinema as a symbol system
that is socially constructed. That is, semiology studied cinema as a social
means of communication and a cultural artifact. Thus, semiology, though
not explicitly political, brought together the study of film and the study of
society. And, of course, in the early seventies, many were prone to uphold
that anything social was at least implicitly political.
Again, the first stage of semiology was not long lived in this countrv.
Metz's theory had a number of problems: it seemed descriptive rather than
INTRODUCTION 5

explanatory; its crowning accomplishment, the grand syntagmatique, a ty-

pology of choices of the narrative structures of the classical cinema set out
like a flow chart, was difficult to apply with uniform results; and it was not
clear that those results were significantly informative (i.e., after you seg-
mented a film, then what and so what?).6 Moreover, though film scholars
thought that linguistically oriented semiology afforded more political po-
tential than rival theoretical models, they did not feel that it was a tool as
suited to the task of ideological analysis as what they desired. They believed
that the first semiology was keyed to the analysis of denotation rather than
connotation, and that the most important dimension of study for ideologi-
cal analysis was the connotative. Also, they thought that the first semiology
paid insufficient attention to the spectator whereas in order to understand
the ideological effects offilmone would need a psychology of the audience.
As a result, film scholars turned to a politicized version of psychoanalysis
for a framework with which to study the ways in which cinema affected its
spectators ideologically.
The second semiology, which in the second half of the seventies became
the regnant film theory in the United States, is fundamentally a combina-
tion of elements of Althusserian Marxism, and Lacanian psychoanalysis,
along with contributions from various literary theorists, most notably
Roland Barthes.7 Its animating idea, derived from Louis Althusser, is that
the primary function of ideology is to reproduce subjects who will carry out
the roles and uphold the values that are required for the continued existence
of a social order such as capitalism. This function is realized through a
process called subject positioning which is engaged in all instances of
communication, including cinema. The film scholar, who follows Al-
thusser, then, proceeds by attempting to discern the ways in which film
positions subjects. Film studies is aided in this by exploiting the work of
Jacques Lacan, for Lacanian psychoanalysis is the form of psychology
tailored for the study of subject positioning. Often when reading contem-
porary film theory, one is taken with the thought that it is really more a
form of psychoanalytic theory rather than of Marxist theory. However, it is
important to keep in mind that the putative justification for its psychoana-
lytic flights is the radical political conviction that ideology cannot be
understood without a theory of subject positioning.
This hasty sketch of the second semiology will be worked out in detail in
the rest of the book. The first chapter takes a close look at some of the
inaugural articles of the second semiology by Jean-Louis Baudry and Chris-
tian Metz. These articles differ in at least one important respect from the
material to be discussed in the rest of the book, viz., these articles are
6 INTRODUCTION

primarily psychoanalytic and not explicitly political. With the second chap-
ter, we begin to deal with the politicized psychoanalytic approach that
represents the core of contemporary film theory. There, specifically, I
outline and criticize the overarching frameworks and presuppositions that
contemporary film theory assembles from Althusser, Lacan, and others.
The next three chapters, in turn, explore the application and extension of
that framework to various topics in film theory. The third chapter deals
with the treatment of the cinematic image by contemporary film theorists,
reviewing their opinions about the status of pictorial representation, per-
spective, and the function of photography with respect to cinema. The
fourth chapter takes up recent views of the function of narrative in the
process of subject positioning. The fifth chapter discusses the ways in which
narratives are rendered cinematically by means of features like editing and
music, and what contemporary film theorists call "suture." The conclusion
will examine some of what I believe are the fundamental methodological
shortcomings of contemporary film theory and will make some recommen-
dations about the directions post-contemporary film theory should take.
Although I have attempted to set the stage for what follows by offering a
thumbnail sketch of the evolution of film studies, most of the rest of the
book is not organized historically but topically. Also, though this stage
setting has been done from the perspective of changes in American film
studies, the ensuing discussion will not be primarily about American film
theorists, but mainly about the French and British theorists who have
influenced them. Undoubtedly, one might write a different introduction
than mine from the vantage point of British or French film studies. I have
chosen the American perspective, however, not only because it is the one I
know best, but also because it seems the most effective way to situate and to
motivate the issues for an American audience. Furthermore, since this book
is not essentially a historical account of the rise of contemporary film
theory, it seems to me that my neglect of the significance of such things as
the student uprisings in Paris in May '68 and the politics of the British Film
Institute for cinema studies is permissable.
This book is extremely critical of contemporary film theory. It contests,
often in great detail, what I take to be the central tenets of contemporary
film theory. However, it would be a mistake to think that the book is purely
negative. For generally, after I have rejected the hypotheses offered by
contemporary film theorists to account for this or that filmic phenomenon,
I attempt to construct a positive hypothesis about the way in which the
phenomenon in question might be theorized. So, for example, after criticiz-
ing contemporary accounts of movie narration, I present an alternative
INTRODUCTION 7

account. These rival accounts are an integral part of my argument against

contemporary film theory; for, following the traditional method of scien-
tific debate, I am attempting to show the shortcomings of contemporary
film theory by demonstrating that there are more attractive, competing
theories that offer better explanations of the data at hand. That is, the case
against contemporary film theory does not rest with simply noting its
logical and empirical failings, but goes on to show there are superior,
alternative ways of explaining the phenomena.
This book is a selection of topics in contemporary film theory; I do not
pretend to canvass every topic in the field, though I think I manage to cover
some pretty foundational ones. I can imagine, and, in fact, did imagine a
book twice this length which would have delved into further topics such as
genre theory, the avant-garde film, documentary film, the relation of theory
to criticism, and so on. However, I finally settled upon the present, more
manageable project.
Nevertheless, there is at least one topic of contemporary film theory
which I have not broached directly and whose omission deserves comment.
The topic is feminism. Anyone familiar with film study knows that femi-
nism is one of its most active areas. So, one wonders, why I have not
devoted at least a chapter to it. In order to answer this question, I must first
note that feminism, including feminism in film studies, is not a monolithic
enterprise. Many feminists in film work in the area that is sometimes called
the study of the image of women in film. I have learned a great deal from
them and am grateful for what they have taught me about our film culture
and our society. They are not included in this book, however, because the
purpose of this book is to oppose that which I take to be wrong in the area
of contemporary film theory. And I do not take the feminists engaged in the
study of women's images to be wrong.
Of course, there are other feminists in film studies who work explicitly
within an Althusserian-Lacanian theoretical framework and I have not
dealt with them either. My reason for this is simply that the criticisms I level
at the framework of contemporary film theory are very foundational. So, if
the foundation goes, then everything that rests on it goes as well. One
cannot, for example, build a feminist theory of suture, if suture theory itself
is flawed in ways that have nothing to do with the perhaps patriarchal biases
of its inventors.
I predict that many feminist theorists will say that I mistakenly assume
that they accept the presuppositions of contemporary film theory whole
cloth, when, in fact, they reject its patriarchal orientation and are at-
tempting to alter it in fundamental ways by, for example, developing a
8 INTRODUCTION

theory of the female subject. Now, I think that they are probably right
about the patriarchal orientation of Lacanianism, and it is certainly true
that they arc attempting to offer a theory of the female subject. However, in
doing this they are essentially rebuilding the framework of contemporary
film theory from the inside, whereas if I am right one ought not to try to
rebuild it. One ought to scrap it entirely. One cannot develop a theory of
female subject positioning if the very notion of subject positioning is
insupportable.
Where there is opposition to contemporary film theory nowadays, it is
often voiced by people who are opposed to theory building in general. But
it is important to stress that my rejection of contemporary film theory does
not spring from an antitheoretical bias. I consider myself to be a film
theoretician. However, I am suspicious of the kind of theorizing wc find
practiced in film studies today. Broadly speaking, this theorizing is what I
call "totalizing" and "top down." That is, it attempts to answer all our
questions concerning filmic phenomena in terms of a unified theoretical
vocabulary with a set of limited laws (primarily concerned with subject
positioning) that are applied virtually like axioms. In contrast, I favor
theorizing that is "piecemeal" and "bottom up." That is, where contempo-
raryfilmtheory presents itself as The Theory of Film (among other things),
I prefer to propose film theortfi—e.g., a theory of suspense, a theory of
camera movement, a theory of the Art Cinema, etc.—with no presumption
that these small-scale theories will add up to one big picture. The contem-
porary film theorist can, of course, rebuke the piecemeal theorist by noting
that his or her system has more theoretical elegance than a series of disjunct,
small-scale theories. But that putative elegance is bought at the cost of
extravagant ambiguity and vacuous abstraction. Or so I shall try to con-
vince you.
 1 .

PSYCHOANALYSIS
METZ AND BAUDRY

I n my introduction, I noted that the first semiology was beset by a

number of vexing problems, notably: it is descriptive rather than explana-
tory, the grand syntagmatique seemed difficult to employ with uniform
results, and even if uniform results could be obtained, it is difficult to see the
way in which those results were significandy informative. However, diffi-
culties of this sort were not historically crucial in the transition to what is
called the second semiology, which might also be called psychosemiology,
due to its heavy reliance on psychoanalysis. That is, contemporary film
theorists cite limitations, other than those just mentioned, as their primary
motives for turning to psychoanalysis.
Among the difficulties that contemporary film theorists say led them to
abandon their quasi-Saussurean model in favor of psychosemiology, is the
notion that the latter, especially in its Althusserian-Lacanian formulation, is
more useful for the analysis of ideology than the first semiology. That is, if
the first semiology was more sensitive to ideology than previous film
theory—insofar as the semiologist at least regarded cinema as a socially
constructed, coded artifact—it was still not adequate enough for the pur-
poses of ideological analysis. Related problems raised against the earlier
semiology include the charge that it was too formalist; it was concerned
with the cinematic sign to the exclusion of investigation of the effects of
cinema on spectators, something felt to be key to the development of an
account of the ideological power of movies.
Also, it is often repeated that the first semiology is best geared to the
examination of the denotative structure of film whereas investigation of the
10 PSYCHOANALYSIS: METZ AND BAUDRY

connotative structures is of greater importance, and, furthermore, it is

asserted that psychoanalysis is the suitable tool for such a project. What is at
issue in this contrast between "denotation" and "connotation" can be
confusing since film theorists do not use these terms univocally, nor do thev
use them in their accepted philosophical sense (where "denotation" refers
to the extension of a term and "connotation" is the meaning of a term). The
denotative structures that the first semiology is said to study so well arc the
categories of articulating the spatio-temporal relations propounded in
Metz's "¿¡rand syntagmatique." But what comprises the connotative dimen-
sion that supposedly still needs to be studied? For some theorists, this
amounts to studying the affective component of film which projects, then,
connects up with the study of the effects of film upon spectators. For other
theorists, influenced by Roland Barthes, however, "connotation" appears
to correlate with "ideology."1 Thus, for these theorists the need to studv
connotation reduces to the call to investigate ideology.
Marxism, supplemented by psychoanalysis, was proclaimed to be the
means by which these alleged shortcomings of the first semiology could be
overcome. Presently, this fusion is broadly informed by the linkage of
Althusserian Marxism and Lacanian psychoanalysis. The desire to studv
ideology is the generally favored rationalization for the supplementation of
Marxism with psychoanalysis. Thus, politics is invoked in order to ground
the embrace of psychoanalysis. But, as we shall see in later chapters, the
psychoanalytic component often appears to supersede the political compo-
nent in contemporary film theory.
The tendency of psychoanalysis to be the dominating concern of contem-
porary film theory is quite evident in several of the seminal essays that
attended the transition to the second semiology. Three of those essays are
the focus of this chapter: Jean-Louis-Baudrys "The Apparatus"; Christian
Metz's "The Imaginary Signifier"; and Metz's "The Fiction Film and Its
Spectator: A Metapsychological Study."21 have chosen to investigate these
essays prior to discussing the formation of the presiding Lacanian-Al-
thusserian paradigm in contemporary film theory for two reasons. First,
though these founding essays of the second semiology have been integrated
into the political polemic of the ruling paradigm, they are not themselves
stridently political. They are overwhelmingly psychoanalytic in their preoc-
cupations. Thus, though they can be adapted for ideological purposes, thev
are not essentially political, and should not just be lumped, without qualifi-
cation, with ostensibly politicized semiology. This is not to say that either
Metz or Baudry are averse to ideological analysis. Baudry, in another
essay,3 offers an ideological characterization of the apparatus, while Metz
PSYCHOANALYSIS: M E T Z A N D BAUDRY 11

makes asides in the essays examined here which suggest his sympathy with
ideological analysis. However, in these essays, Baudry and Metz are clearly
more concerned with the psychoanalytic effects of cinema on spectators
than in the explicit, detailed analysis of the ideological effects of the
medium.
A second difference between these three essays and the reigning Marxist-
psychoanalytic paradigm is that these essays, though they make use of
Lacanian psychoanalysis, are not as pervasively wedded to that model as is
much contemporary film theory. The psychoanalyst Bertram Lewin plays a
more important role than Lacan in Baudry's essay while Melanie Klein
seems at least as important to Metz as is Lacan. Thus, since none of these
essays is as Marxist or as Lacanian as is fashionable nowadays, I will treat
them as apart from the dominant framework of contemporary film theory,
while reemphasizing that these essays can be and have been easily dra-
gooned into the service of the established film theory by the addition of a
few crucial premises.
Baudry and Metz both employ psychoanalysis to describe and to explain
the ways in which cinema affects us. Their method is to draw analogies
between cinema and various forms of psychic phenomenon: daydreams,
night dreams, syndromes, such as voyeurism and fetishism, and early psy-
chosexual conditions and fantasies, such as primitive oral narcissism. The
thinking behind this analogizing appears to be that if cinema is sufficiently
like certain psychic phenomena, then we can begin to explain the power of
cinema over us by extrapolating psychoanalytic explanations of the power
that analogous psychic phenomena exert. That is, if cinema appreciably
resembles certain forms of psychic phenomena, then we can hypothesize
that the power of cinema is akin to the power of its pscyhic analogs:
whatever desires those analogs satisfy, cinema satisfies; whatever compel-
ling force those analogs mobilize, cinema mobilizes. Baudry, for example,
analogizes film and night dream and infers that film satisfies the oral
regressive penchant, also satisfied by night dream, of feeling undifferenti-
ated from one's environment. Metz, in one essay,4 introduces a corrective
analogy between film and daydream, but reaches, nevertheless, some con-
clusions that correspond to Baudry's.
Analogies between cinema and dream have been with us almost as long
as film. Since the turn of the century, with movies such as Edwin S. Porter's
Dream ofa Rarebit Fiend and Griffith's The Avenging Conscience, filmmakers
have attempted to mime or to represent mental processes, notably dream-
ing. In the art cinema, Bunuel's The Andalusian Dog and his The Age ofGold
are famous examples of this, while Bergman's Wild Strawberries provides
12 PSYCHOANALYSIS: M E T Z A N D BAUDRY

another well-known, if prosaic, case. Hitchcock's Spellbound represents an

attempt at dream illustration of a sort that is routine in popular cinema
while Pabst's Secrets ofa Soul is an essay in state-of-the-art Freudian theoriz-
ing about psychoanalytic interpretation which employs dream sequences
created through consultation with practicing analysts Karl Abraham and
Hans Sachs. In short, there is no question that a film or film sequence can
imitate dream processes. Superimposition can be made to function like the
primary process of condensation; disjunctive editing can ape the associative
"logic" of dream. However, the claim that Metz and Baudry advance is not
that some films are to be understood through analogies with psychic pro-
cesses such as dreaming, but rather they conceptualize all of cinema by
reference to psychic analogies. Their claim is theoretical and general, rather
than critical. That is, since some filmmakers model some films and film
sequences on psychic processes, it behooves a critic to explicate such films
with an eye to the ways in which the films construe, or more often,
misconstrue, the psychic processes they strive to represent. But the fact that
some films can be unproblematically designated as analogs or attempted
analogs to dreaming does not show that all films or that cinema as such is
analogous to dreaming. And it is the latter theoretical point that Baudry
and Metz are after.
The theoretical notion that film as such is analogous to dream is also
longstanding. An early proposal of this oudook occurs in Hugo von Hof-
mannsthal's 'The Surrogate for Dreams," published in 1921. 5 Like Freud-
ians of the period, Hofmannsthal conceives of dreams in terms of wish-
fulfillment. Examining film technique, Hofmannsthal notes two wishes
that it vicariously grants. Poetically, he refers to these wishes as the cloak of
invisibility and the magic carpet ride. That is, through editing, the spectator
travels through space; one minute she is in Europe and the next she is
circling the minarets of Baghdad. By editing, film realizes the myth of the
magic carpet. Also, the spectator is invisible to the spectacle; he watches
Antony and Cleopatra embrace with impunity. Cinema, like Albrecht the
Dwarf, supplies us with a cloak of invisibility, again counterfeiting a psy-
chologically significant myth. In effect, part of what Hofmannsthal is
arguing is that through its technique, film satisfies wishes for omnipo-
tence—the ability to travel anywhere—and for omniscience—the ability to
see anything while unseen. Interestingly, these notions reappear in Metz's
theory which emphasizes the psychic importance of cinema by referring to
the fantasy of perceptual mastery it evokes and to the possibility for voyeur-
ism that it purportedly affords. 6 And like Hofmannsthall, Metz's account of
the power of cinema presumes that spectators identify with the camera.
PSYCHOANALYSIS: M E T Z AND BAUDRY 13

American theorists, of course, are very familiar with the film/dream

hypothesis through the very influential and popular treatise by Susanne K.
Langer entided Feeling and Form.7 There Langer notes what she takes to be
three significant analogies between film and dream. In both phenomena,
she asserts that the subject feels at the center of the visual array while the
array itself promotes a compelling aura of givenness. She also notes that
continuity in film, i.e., editing, abides by an affective or associative logic as
does dream imagery. Langer's overall theory of art holds that the generic
function of art is to clarify and to objectify—that is, to manifest exter-
nally—forms of feeling, aspects of our felt emotive life. Each art form, in
turn, specializes, so to speak, in objectifying distinct domains of our emo-
tive life. Film objectifies dream. Correspondingly, both Metz and Baudry
identify part of the power of cinema as its manifestation of inner processes
of the unconscious to the spectator, though, unquestionably, both in their
psychoanalytic idiom and in their detailed conclusions about the signifi-
cance of film, they are philosophically at odds with Langer.
Despite its long heritage, however, the theoretical correlation of film
with such psychic phenomenon as dreaming is far from established. Baudry
and Metz represent two of the most sustained efforts to carry off the
analogy in a theoretically significant way. In what follows, I will examine
their arguments very closely. By way of a preview, let me say that I think
their project a dismal failure. After examining Metz's and Baudry's position
in detail, the chapter concludes with general observations about the in-
advisability of pursuing film theory through analogies with psychic phe-
nomena and about the theoretical inappropriateness of psychoanalyzing
the cinematic apparatus.

JEAN-LOUIS BAUDRY AND " T H E APPARATUS"

A major reason, given by contemporaryfilmtheorists, for their shift from
a semiological framework of study to a psychoanalytic one, is that the
semiological model is too narrow. It concerns itself with the structure of the
cinematic sign but does not, according to many contemporary film theo-
rists, pay sufficient attention to the effects of cinema upon the spectator.
The semiological model, at least in the ways it was employed in the late
sixties and early seventies, was felt to be myopically object-oriented. In
order to remedy this putative shortcoming, film theorists resorted to psy-
choanalysis.
Jean-Louis Baudry's essay "The Apparatus" was a seminal essay in the
turn to psychoanalysis. In this essay, he attempts to account for the impres-
14 PSYCHOANALYSIS: METZ AND BAUDRY

sion of reality that cinema is said to impart to spectators. He intends to use

psychoanalysis, that is, to analyze what he takes to be a paramount effect of
cinema on audiences. However, though the phrase "impression of reality"
recurs frequendy in "The Apparatus," one must be careful in that which one
identifies as the phenomenon with which Baudry is concerned. For Baudrv
does not contend that the impression of reality caused by cinema is equiv-
alent to our everyday encounters with the world; cinema is not a replication
of our ordinary impressions of reality. Rather, cinema is said to deliver an
impression of reality that is more-than-real. That is, less paradoxically
stated, Baudry wishes to deploy psychoanalysis to explain cinema's intense
effect on spectators; he wants to analyze the peculiarly charged relationship
we have with the screen when we attend movies.
Moreover, Baudry does not search for this effect by scrutinizing the
content of the images or the stories of particular films or even of particular
kinds of films. Instead he sees this effect as the product of what he calls
"the apparatus," a network which includes the screen, the spectator, and
the projector. That is, Baudry seeks the origin of the impression-of-reality
effect in the projection situation itself, irrespective of what is being
screened.
Baudr/s basic procedure for discovering the origin of the impression-of-
reality effect is to draw a series of analogies between dreams and the
projection context, or, as he prefers to call the latter, the apparatus. He is
motivated in this by a belief that dreams likefilmengender an impression of
reality that is highly charged, that is, an impression of what Baudry thinks
of as the more-than-real effect. Thus, Baudry hopes to extrapolate the
psychoanalytic explanation of the charged impression of reality in dreams
into an explanation of the impression of reality in cinema.
Though Baudry does not set out his case in a logically rigorous fashion,
his analysis implicitly takes the form of an inductive argument by logical
analogy. For example, he notes that the film viewer and the dreamer share
the property of having their movements inhibited, that both inhabit
darkened rooms, and that both film and dream impart an impression of
reality. Dream, in turn, is said to have this consequence insofar as it induccs
regression to an earlier psychosexual stage, that of primitive narcissism
where the self is supposedly not differentiated from the other nor is percep-
tion differentiated from representation. On the basis of the similar condi-
tions and effects, respectively, of film and dream, Baudry infers that the
impression of reality in film is brought about by a regressive mechanism
similar to that operative in dream.
Though arguments by analogy are not absolutely conclusive—they arc,
PSYCHOANALYSIS: M E T Z A N D BAUDRY 15

after all, inductive rather than deductive-—and though they are often
abused, they are a respectable form of reasoning. We use them all the time.
For example, if you have a 1964 Saab and I have a 1964 Saab, and both cars
are the same model, both engines are in exactly the same condition and state
of repair, both carry the same weight, use the same fuel, and have been
serviced in the same way by the same mechanic, and my Saab can go 55
mph, then we infer that (probably) your car can go 55 mph. Stated
formally, this type of argument takes the following pattern: If we have
items A and B, and they are similar in a number of relevant respects, say in
terms of properties pi through px-l, and item B also has property px, then
we infer that (probably) A has px. That is,
1. Item A has properties pi . . . px-l (e.g., A is a 1964 Saab).
2. Item B has properties pi . . . px-l.
3. Item B also has property px (e.g., B can go 55 mph).
4. Therefore, (probably) Item A also has property px.
Premises 1 and 2 set out the analogy; if more items are being analogized
more premises will be added here. Once the analogy is set out, these
premises can be combined with premise 3, which states a property known
to be possessed by B (but not observed to be a property of A), in order to
license the probable conclusion stated in 4. Obviously such arguments gain
strength when the number of items and/or the number of relevant proper-
ties cited are multiplied. Inversely, the argument loses force as either data
base for the analogy is diminished. This can be done by:
A. showing that the analogies cited fail (for example, your car is
really a 1921 Ford, not a 1964 Saab).
B. demonstrating that the analogies cited are irrelevant to what is at
issue (for example, that both cars are green).
C. noting relevant disanalogies between items A, B . . . in order to
challenge the purported similarity of the cases under comparison (for
example, that your Saab has no wheels).
With this sketch of the logic of Baudry's approach, we can go on to fill in
the details and evaluate the persuasiveness of his account of the psychic
mechanism that he believes causes cinema's characteristic impression of
reality.
As we have already noted Baudry holds that the conditions of reception
of film and dream are analogous; both involve a darkened room and the
inhibition of movement. The film viewer sits in his seat; the dreamer lies
abed. Inhibited motoricity is also a feature of the infantile state to which the
16 PSYCHOANALYSIS: METZ AND BAUDRY

dreamer is said to regress. Connected to this inhibition o f motoricity is

another feature: the lack of the means to test reality. Baudry writes:
In order to understand the particular status of cinema, it is necessary to
underline the partial elimination of the reality test. Undoubtedly, the means
of cinematographic projection would keep the reality test intact when com-
pared to dreams and hallucination. The subject always has the choice to close
his eyes, to withdraw from the spectacle or to leave but no more than dreams
does he have means to act in any way upon the object of his percepdon,
change his viewpoint as he would like. . . . His relative motor inhibition
which brings him closer to the state of the dreamer, in the same way that the
status of the reality he perceives (a reality made up of images) would seem to
favor the stimulation of a regressive state, and would play a determining role
in the subject-effect of the impression of reality, this more-than-real of the
impression of reality, which we have seen is characteristic not of a subject to
reality, but precisely of dreams and hallucinations.8

Here, Baudry notes that there is only a partial analogy in regard to reality
testing between film and dream. But he thinks this similarity is important
because along with inhibited motoricity, and perhaps because o f it, the lack
o f reality testing reproduces the conditions o f the infantile state o f primitive
narcissism that explains the impression-of-reality effect that the cinematic
apparatus is said to induce. Also, in this passage, Baudry alludes to two
other analogies between film and dreams: both traffic in the medium of
images and both deliver a more-than-real impression of reality.
So far five analogies between film and dream have been noted: inhibited
motoricity in the subject; lack o f reality testing; darkened rooms; the
medium of images; the more-than-real impression of reality. Are there
other analogies?
For Baudry, following the psychoanalyst Bertram Lewin, dreams, like
the cinematic apparatus, have screens. That is, Baudry argues that dreams
are projections onto dream screens in a way that is analogous to film
projection. Baudry writes:

That dream is a projection reminiscent of the cinematographic apparatus is

indeed what seems to come out of Lewin's discovery of the dream-scene, the
hypothesis for which was suggested to him by his patients' enigmatic dreams.
One young woman's dream, for example: "I had my dream all ready for you,
but while I was lying here looking at it, it began to move in circles far from
me, wrapped up on itself, again and again like two acrobats." This dream
shows that the screen, which can appear by itself, like a white surface, is not
exclusively a representation, a content—in which case it would not be neces-
sary to privilege it among other elements of the dream content; but rather it
would present itself in all dreams as the indispensable support for the projec-
PSYCHOANALYSIS: M E T Z AND B A U D R Y 17

tion of images. It would seem to pertain to the dream apparatus. "The Dream
screen is a surface on which a dream seems to be projected. It is the 'blank
background' (empty basic surface) which is present in the dream although it
is not necessarily seen; the manifest content of dream ordinarily perceived
takes place over it or in front of it." 9

Baudry also points out that people often describe their dreams as being like
movies.
The dream screen/film screen, dream apparatus/film apparatus analogy is
particularly crucial for Baudry. For Lewin has a psychoanalytic account of
the dream screen that is pertinent to the effect dreams have upon us. And
Baudry intends to extend that account to cinema.
According to Lewin's hyp»thesis, the dream screen is the dream's hallucina-
tory representation of the mother's breast on which the child used to fall
asleep after nursing. In this way, it expresses a state of complete satisfaction
while repeating the original condition of the oral phase in which the body did
not have limits of its own, but extended undifferentiated from the breast. 10

Thus, via the dream screen, the dreamer regresses to and relives a stage in
our psychosexual development marked by primitive narcissism, a stage
where self and environment are said to merge and where perception and
representation are believed to be undifferentiated. Moreover, this regres-
sion satisfies a desire, a desire to return to that sense of undifferentiated
wholeness. It is this desire in turn which gives the dream imagery its special
charge and which accounts for the intensity with which we regard it.
Insofar as the cinematic apparatus mirrors relevant aspects of the dream
apparatus—inhibited motoricity, lack of reality testing, visual imagery, the
more-than-real impression of reality, projection, and a screen support—
Baudry feels warranted in adopting Lewin's hypothesis about the causation
of dream and the dream effect as an explanation of our animating desire for
and our experience of the cinematic apparatus. That is, a regressive mecha-
nism seeking to revive the experience of primitive narcissism is what draws
us to movies while satisfaction of that desire is what renders that experience
more-than-real. So, by postulating the return to primitive narcissism as the
operative agency in film spectators, Baudry thinks he has isolated the cause
of the cinematic effect while also, in the process, supplying an account of
why the movie experience is desirable to us. Later, we shall see, Metz poses
a question that the latter point might be thought to answer; that is, Metz
asks psychoanalysis to explain why we go to movies. From Baudrys per-
spective, the answer seems to be to relive that stage of primitive narcissism
where all-is-one, including a conflation of perception and representation.
Indeed, for Baudry, the cinematic apparatus incarnates a wish for a simula-
18 PSYCHOANALYSIS: M E T Z A N D BAUDRY

tion machine "capable to offer the subject perceptions which are reallv
representations mistaken for perceptions," 11 thereby recalling the dream
state which itself derives from a regression to an archaic state where percep-
tion and representation are not differentiated.
Summarizing Baudry's case so far, his argument looks like this:
1. The dream apparatus has the folllowing features: inhibition of
movement; lack of reality testing; an imagistic medium; a dark room;
projection; a screen; a more-than-real impression of reality; a tendency to
efface the distinction between perception and representation.
2. The cinematic apparatus has exacriy the same features noted in
premise 1.
3. A significant animating force behind the dream apparatus is the
desire for and regression to primitive narcissism which enactment causes
the charged experience of dreams.
4. Therefore it is probable that a significant animating force behind the
cinematic apparatus is the desire for and regression to primitive narcissism
which enactment causes the charged experience of cinema.
Baudry, of course, does not hold that films are mistaken for dreams. He
rather construes them as simulations of dreams. For this reason he might
want to say that the regression encountered in cinema is less intense than
that of dreams. Nevertheless, he would appear to hold that to whatever
degree the more-than-real impression of reality of film approximates the
analogous effect of dreams, it is a function of the process of some measure
of regression to the all-is-one state of primitive oral narcissism.
Because of the emphasis Baudry places on the cinematic apparatus as a
simulation of unconscious phenomena, specifically of the hallucinatory
aura of dreams, one might offer a slightly different interpretation of Bau-
dry's argument than the one just presented. That is, one might take Baudry
to be saying that since film replicates the most significant conditions of
dreaming—e.g., motor inhibition, lack of reality testing, and so on—it
triggers the same effect—regression to primitive narcissism. On this inter-
pretation of Baudry's strategy, the argument rides on the principle: same
conditions, same effects. However, whether one chooses this interpretation
or the interpretation of the analysis as an inductive argument by analogy is
logically indifferent for the purposes of evaluating Baudry's central claims.
For in either case Baudry's central assertions stand or fall on the basis of the
adequacy of the correlations he draws between cinema and dreams.
Baudry concludes his essay by asserting that the unconscious has an
instinctual desire to manifest itself to consciousness. This suggests that
PSYCHOANALYSIS: M E T Z A N D BAUDRY 19

cinema is one means for fulfilling this instinctual desire. For in simulating
dream, cinema satisfies the desire of the unconscious for acknowledgment.
This somewhat resembles a claim that Metz makes to the effect that by
combining elements of night dream and daydream film causes pleasure by
externalizing what is usually experienced as internal. But this claim of
Baudry's can only be accepted if cinema is a suitable simulation of dreams,
that is, if the analogies Baudry draws between film and dream are fitting and
if they are not outweighed by significant disanalogies.
However, before turning to an assessment of Baudry's central thesis, it is
important to take notice of an elaborate complication in the text which has
not been remarked upon so far. Baudry not only analogizes film and dream
but he also compares both with the description of the circumstances of the
prisoners in Plato's myth of the cave.12 Those prisoners are chained in a
darkened vault. Behind and above them, fires burn. As passersby walk
between the prisoners and the flames, the strollers' ambulating shadows are
cast upon the wall of the cave. The prisoners see these moving shadows and
take them for reality. Through this allegory, Plato sets forth his disparaging
estimation of the ordinary person's "knowledge" of the world. It is based
on illusion; it is nought but shadowy deception.
Baudry calls attention to the ways in which Plato's cave resembles the
cinematic apparatus. The cave is analogous to the apparatus in obvious
respects: the motoricity of the prisoners is inhibited as is their capacity to
test reality. The cave, like the movie theater, is a dim space. The shadows in
Plato's cave might be thought of as projections and the wall of the cave is a
screen of sorts. Moreover, the projecting device is above and behind the
prisoners and is, so to speak, hidden from them. The imagery in both
Plato's cave and the cinema is a matter of shadows or reflections caused by
passing something before a light. And in both cases, one can designate the
play of "two scenes": first, the scene in the world that gives rise to the
"shadows" and second, the scene comprised of the shadows themselves.
These analogies lead Baudry to conclude that Plato's myth of the cave
"doesn't merely evoke, but quite precisely describes in its mode of opera-
tion the cinematographic apparatus and the spectator's place in it." 13
Now the question immediately arises as to what logical purpose the
analogies between Plato's cave and the cinematic apparatus serve in the
context of Baudry's overall analysis of film in terms of regression to primi-
tive narcissism. For these added cave analogies do not function logically to
enhance Baudry's argument concerning film and dream. The conclusion of
that argument is that regression is the motor of the cinematic apparatus just
as it is the motor of the dream apparatus. For the analogies with Plato's cave
20 PSYCHOANALYSIS: METZ AND BAUDRY

to film and dream to bolster this conclusion—viz., that regression to

primitive narcissism is the motor of the cinematic apparatus—we would
have to have reason to claim antecedent knowledge to the effect that Plato's
myth was generated by the type of regressive mechanism that the argument
wants to attribute to the cinematic apparatus. But we do not have such
knowledge. Whether Plato's myth of the cave derives from a regressive
desire remains to be proved to the same degree that the cinematic appara-
tus' origin in such a desire does. That is, the psychoanalytic cause of Plato's
myth is not known prior to Baudr/s argument and in that sense is logically
in the same boat as whatever the animating force behind the cinematic
apparatus turns out to be. Thus, adding the analogies of Plato's cave, since
the cause of that myth cannot be supplied as a premise of the argument
(but, at best, as a corollary conclusion) does not strengthen the argument
by analogy between the cinematic apparatus and the dream apparatus. So
the question remains as to the point of Baudr/s ornate rendition of Plato's
myth.
Baudry, of course, has no wish to endorse Plato's epistemological posi-
tion which he, Baudry, misidentifies as idealism, a label more apt for a post-
Cartesian such as Berkeley (that is, Plato does not believe that all that exists
is mental and, indeed, the mental/physical distinction relevant to the for-
mation of an idealist philosophy does not appear to have been historically
available to Plato). Rather, Baudry tears the myth of the cave out of the
context in which it functions as an allegory, and he treats it as a fantasy ripe
for psychoanalysis. Among other things, Baudry claims that it is a proto-
cinematic wish, that is, a deep-seated wish for something very much like
cinema before the invention of cinema. Baudry sees evidence for similar
proto-cinematic wishes throughout history: the camera obscura, the magic
lantern, the praxinoscope. What Baudry seems to conclude from the exis-
tence of a proto-cinema stretching from Plato's cave to the praxinoscope is
that it supplies evidence for a transhistorical, psychical, or instinctual source
of desire behind the invention of the cinematic apparatus and its préfigura-
tions.14 And, of course, if the compelling force behind cinema is in-
stinctual, that may supply prima facie grounds for approaching it psychoan-
alytically. Thus, I take it that the point of Baudry^ use of Plato's cave in
'The Apparatus" is not to enhance the central argument about the causal
relevance of regression to film, but rather to mount a coordinated but
independent argument to persuade us that the recent invention of cinema is
really a manifestation of a long-standing, transhistorical or instinctual de-
sire of the sort that psychoanalysis is fitted to examine. That is, the discus-
sion of Plato's cave is meant to convince us of the appropriateness of
PSYCHOANALYSIS: M E T Z A N D BAUDRY 21

psychoanalyzing cinema and it is not properly speaking part of the argu-

ment by analogy that concludes with the assertion of regression as key to
the cinematic apparatus.
Of course, in speaking this way I am offering an interpretation of Bau-
dry's essay, one guided by the logical requirements of the type of argument
Baudry appears to advance. I admit that at points Baudry himself writes as
though the analysis of the allegory of the cave were an essential part of the
film/dream argument. Not only does this fly in the face of the logical point
made earlier, but it also promotes many extravagant and confusing, free-
associative leaps as Baudry attempts to forge connections simultaneously
between Plato's cave, film, and dream. For example, Plato's cave correlates
with the darkened room of the movie theater. What connection does this
have with dreams? Baudry says that the dreamwork often represents the
unconscious by means of underground places.15 Now even if this has some
connection with Plato's cave, what is its relevance to the cinematic appara-
tus? Films are not characteristically viewed in caves or underground places.
The effects of such whimsical flights of fancy can be minimized if we restrict
our attention to the central argument that analogizes film and dream which,
anyway, is the logical fulcrum of Baudry's case. Thus, a sense of interpretive
charity leads me to regard the film/dream argument and the analysis of
Plato's cave as making separate though coordinated points.
Undoubtedly, Baudry's essay also attempts to show that the "apparatus"
of Plato's cave has the same regressive mechanism behind it as does the
dream "apparatus." And Baudry's way of showing this is ostensibly an
argument by analogy like that concerning the cinematic apparatus. But this
is, logically, a parallel argument to the film/dream argument, one that
neither supports nor derives support from the speculations on film and
dream. That is, the cave/dream argument concluding with regression as the
motor behind Plato's myth is an induction to be pursued independendy of
the film/dream argument. Indeed, Baudry's discussion of Plato is only
relevant to film theorists—as opposed to historians of philosophy—insofar
as the myth of the cave can be demonstrated to be proto-cinematic. Thus, I
will restrict comment upon Baudry's cave/dream analogies to those points
that are relevant to establishing the existence of a proto-cinematic wish.
Baudry's "The Apparatus," then, contains at least two major arguments
for film theorists: that the apparatus of cinema importandy involves regres-
sion to primitive narcissism and that the archaic wish underlying cinema is
atavistic, reaching as far back in history as Plato's myth of the cave. Of these
two arguments the former seems to me of greater moment because, if it is
true, it is what gives the claim about a proto-cinematic wish precise sub-
22 P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y

stance, and also because it would supply an interesting and substantial

insight even if the claims about proto-cinematic wishes were false, i.e., if the
invention of cinema responded to a historically recent wish rather than to
an ancient longing of the human race.
The argument that the cinematic apparatus involves regression to a
period o f primitive narcissism where self is not differentiated from the
environment is an inductive argument by analogy and, therefore, its con-
clusions are only probable. This is not problematic for most of what we
value as knowledge is at best probable. The degree of warrantibility in such
an argument, however, depends on the strength o f the analogies cited in the
premises and on the presumption that there are not significant disanalogies,
in this case between film and dream, which would neutralize or outweigh
the persuasiveness of the analogies advanced. Thus, to assess Baudry's
thesis we must consider whether his analogies are apt and compelling, and
whether or not there are profound disanalogies between film and dream
which render Baudry's analogies fledgling.
The two analogies that Baudry repeatedly stresses involve the inhibition
of movement and the absence of reality testing, features purportedly shared
by the cinematic apparatus and dream. Supposing that these are features of
dreams, are they also features of film viewing? The dreamer is asleep;
insofar as he is not a somnambulist, his literal movement is restricted to
tossing and turning. O f course, his movement capacities as a character in his
own dream can be quite expansive. But insofar as he is asleep, the move-
ment of his physical body is involuntary. But what of the cinema viewer?
Conventionally we sit in our seats, moving our heads, arms, and so on
within a small perimeter of activity. But is our movement inhibited in a way
that is significantly analogous to the inertness of sleep? First of all, a kev
reason for speaking of motor inhibition, both in terms of sleep and in terms
of the infantile state of primitive narcissism, is that in those cases the lack of
mobility, for different reasons, is involuntary. However, no matter how
sedentary our film viewing is, we are not involuntary prisoners in our seats.
Of course, Baudry speaks of this lack of motoricity not only in respect to
film viewing and dreaming, but also with reference to Plato's prisoners
whose constraint is involuntary. So the dream state and that of the pris-
oners correlate along the dimension of involuntary motor inhibition. And
movie viewers are supposed to resemble the prisoners. But from this one
cannot surmise that it is correct to claim corresponding motor inhibition
for movie spectators since movie spectators resemble the prisoners only in
such respects as being viewers of reflection and not in terms of involuntari-
ness. Unlike Plato's prisoners, the film viewer can move her head volun-
PSYCHOANALYSIS: M E T Z A N D B A U D R Y 23

tarilv, attending to this part of the screen and then the next. What she sees
comes under her control, unlike the dreamer or the prisoner, in large
measure because of her capacity to move her head and her eyes. And, the
film viewer, as Baudry admits, can leave the theater, change her seat, or go
into the lobby for a smoke.
Though Baudry does not make this move, a proponent of the inhibition
analogy might claim that when a film spectator adopts the convention of
sitting before the movie screen, she adopts the pretense of having her motor
capacities inhibited. But there is no evidence that such a game of make-
believe is occurring. A more likely description of what the spectator does
when adopting the convention of taking a seat is that she opts for the easiest
method of attending to the film. Literally, her motoricity is not inhibited,
nor does she feel it or pretend it to be. If we are willing to describe film
viewing as involving motor inhibition, we should be equally willing to
describe witnessing baseball games and listening to political speeches as
involving motor inhibition. And to the extent that such descriptions of
baseball and speeches is inaccurate, so is a description of movie viewing as
movment inhibited inaccurate. Moreover, even if there is a sense in which
we might say that movement in all these cases is "inhibited," it is certainly
not a matter of motor inhibition, but a voluntary inhibition promoted by
respect for conventional decorum.
Of course, the point that sitting at movies is a social convention is central.
Movies can be watched with no loss ofeffect while standing; people frequently
walk to the rear of the theater and watch, stand in the aisles while they grab
a smoke or relax their bottoms. Nor are such standing filmgoers necessarily
stationary; if one watches the film while pacing across a side aisle, the
impression the film imparts need not be lost. Baudry connects the putative
impression of reality imparted by film to inhibited motoricity. Given this,
one would predict that that impression would not occur if the spectator
watched while also moving voluntarily. If there is such a phenomenon as
the impression of reality, then it should be an empirical matter to establish
whether it disappears when the spectator is in movement. In my own case, I
have found that I can back out of a movie theater while watching the screen
or return to my seat from the beverage bar with no discernible difference in
the impressions I derive from the screen when I am seated. I know I'm
walking in one case and sitting in the other, but these are proprioceptive
impressions and not screen impressions.
Perhaps Baudry would admit that the film viewer's movement is not
literally inhibited, but would attempt to save his analogy by saying that the
film viewer feels inhibited. Phenomenologically, I have never had such an
24 PSYCHOANALYSIS: METZ AND BAUDRY

experience. But even if others do have such experiences, this will not help
the analogy that Baudry wishes to draw. Why? Well, if we shift to a
phenomenological register, then the dreamer often frels in motion when
dreaming, for instance, when one dreams one is falling or being pursued by
a three-headed ogre. That is, the dreamer often feels in motion when he is
not, whereas the film viewer does not ordinarily take himself to be literally
in motion when he is not. 16 Of course, Baudry may say that the inhibited
movement which is attributed to the movie spectator is really metaphorical.
But why should a correlation between his metaphorical description of the
film viewer and the literal motor inhibition of the sleeper count as anything
more than an entertaining but fanciful piece of equivocation?
Baudry's second key analogy between the cinematic apparatus and dream
hinges on the claim that both involve an absence of reality testing. Of
course, in one sense, the film viewer is fully capable of indulging in reality
testing. He can go up to the screen and touch it; he can shift his view of the
screen, noting that the contours around the objects do not alter, and,
thereby, he can surmise that the projected array is two dimensional. Also,
things like coke botdes and cabbages can be and have been thrown at
movies, a dramatic measure for revealing the nature of the screen. Baudry is
aware of this; when he speaks of an absence of reality testing in film
viewing, his reference is not to an incapacity the viewer has in relation to
objects, such as screens, in the actual world; rather Baudry has in mind that
the viewer lacks the ability to test reality within the world of the film. That
is, the movie viewer cannot enter the visual array onscreen in order to
ascertain whether the buildings in Siegfried are concrete or merely card-
board.
Baudry also connects lack of reality testing with the inhibition of move-
ment. Plato's prisoners, and the preambulatory infants at the stage of
primitive narcissism cannot test reality at a distance because they are immo-
bile. But the same correlation between motor inhibition and absence of
reality testing cannot, as Baudry suggests, be extrapolated to film viewing
and dreaming. For if there is an absence of reality testing in both these
cases, then that is a function of the fact that, loosely speaking, there is no
reality to be tested in the world of the film and the world of the dream
(where "reality" is understood as the foil of "representation"). So if the
correlation based on absence of reality due to inhibited movement is key to
aligning film and dream with the infantile state of primitive narcissism, the
analogy is inaccurate.
We may also wish to know whether it is really appropriate to hold that
the film viewer has no means for testing reality inside the world of the film.
PSYCHOANALYSIS: M E T Z AND BAUDRY 25

Certainly it is true that we cannot walk into the world of Casablanca in

order to determine whether the characters are really drinking whiskey. But
at the same time, I think that it pays to recall a really overwhelming
disanalogy between film viewing and dreaming. Namely, films are publicly
accessible; they can be viewed by more than one person. Moreover, they
can be repeated; we can see the same film again and again, and we can fall
back on all sorts of evidence—production and distribution records, the
testimony of other viewers and of the filmmakers, the existence of similar
prints, and so on—to warrant the claim that the film we just saw, say
Captain Blood, is the same film we saw in the past. This is a radical disanal-
ogy with dreaming. Neither the analyst nor anyone but the dreamer has
access to the dream. And no one, including the dreamer, can be sure that his
report of a dream is accurate; there is no interpersonel validation available.
Even with a "recurring" dream, we have litde reason to be confident that
the dreamer experienced exacdy the same dream from night to night. What
does this epistemic disanalogy have to do with reality testing? Simply that
withfilmsthere is a way in which we can "test reality," i.e., corroborate our
experience of a movie. We can ask someone else if she saw what we saw.
Nor is this something we do only after a film is over. During Lifefbrce I
leaned over to my neighbor and asked "Did I really just see a vampire-nun?"
to which she replied "I saw her too."
There are, in short, means to test the veracity of our experience of films.
We cannot plunge into the image, but we can corroborate what we see
there, which is the sort of reality testing that is appropriate to visual fictions
(as opposed to what might be called ordinary visual "realities"). Moreover,
if we are worried about testing the fidelity of documentaries to their
subjects, that is also possible. My point here is simply that it seems to me
inappropriate to describe the film viewer as lacking the means for testing
reality. And if I am right in this matter, this short-circuits Baudry's second
key analogy between film and dream. However, if I am wrong and the
analogy is acceptable, the considerations I have just raised present another
problem for Baudry. For even if in some sense his analogy works, I have
also pointed to a major disanalogy between film and dream: viz., that film
experiences are open to interpersonal verification. This appears to me to be
important enough to outweigh analogies betweenfilmand quasi-solipsistic
phenomena, since it establishes that film viewing has an objective dimen-
sion and is not purely subjective.
Films are a visual medium and so are dreams. Is this a significant analogy
between the external and internal phenomena under comparison? Not
really. For memories are also often visual. Why not analogize film to
26 PSYCHOANALYSIS: M E T Z AND BAUDRY

memories as certain film realists might propose? Here, it might be argued

that dream is the appropriate analog—that we know to eliminate memory
as a viable candidate—because of the earlier analogies that correlate film
with dream rather than memory. But as I hope I have shown, those earlier
analogies are not so sturdy, nor, I might add, is the film theorist under any
imperative to identify any mental correlate for film.
However, there is also an important disanalogy between film imagery
and dream imagery which indicates that the two are not congruent. The
single film image is ordinarily complete, by which I mean that, typically, it
is visually articulated throughout. Dream images, on the other hand, tend
to be incomplete, foregrounds without backgrounds or figures in a void.
This can be simulated in film as can the "noise" of the dreamwork; note
Brakhage'sfilms.However, the ordinaryfilmimage does not "look like" the
ordinary dream image. That is, even iffilmand dream are imagistic, they are
radically dissimilar imagistic media. They are too unlike to be treated as
cognate phenomena. Later, as we shall see, Metz notices further, crucial
dissimilarities between film and dream,17 which added to the disanalogies I
adduce render Baudry's argument even more unlikely.
Baudry's analogy between film and dream in respect of darkened rooms
is also problematic. One can, of course, fall asleep mid-day on the beach.
And movies can be viewed in well-lit circumstances. I expect that Baudry is
probably right in asserting that most of the time we dream and viewfilmsin
the dark. But there is still something strange about this correlation. The
film viewer is not only objectively in a darkened room; she is experientiallv
aware of being in a darkened room. But even if the dreamer is objectively in
a darkened room, she is unaware of it. Indeed, she may believe that she is on
a blistering, sun-baked desert. That is, though there is a possible objective
analogy between the film viewer and the dreamer, their experiences are
disanalogous. Now if a film is supposed to simulate dreams or trigger the
same kind of response or mechanism in the subject, wouldn't it seem more
likely that what the viewer experiences be key to the dream analogy rather
than the objective, physical conditions of reception? That is, the dreamer
does not have the same experiential awareness of a darkened room that the
film viewer does. So why would the film viewer's awareness of a darkened
room remind one of dreaming or simulate dreaming for the unconscious?
In this case, Baudry seems to overvalue the significance of correlations
between the objective, physical conditions of film viewing and dreaming
while forgetting the crucial, phenomenological disanalogies between the
film experience and dream experience. We saw that there was a similar
problem with his treatment of inhibited motoricity, where he ignored the
P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y 27

fact that objectively, physically leaden sleepers often feel in vigorous

motion.
Another analogy that Baudry proposes between film and dream is that
both impart what he refers to as a more-than-real impression of reality. Part
of the problem with evaluating this claim is the vagueness o f the notion o f a
more-than-real impression of reality. That is, even if film and dream impart
such broadly describable impressions, are their respective impressions the
same in analytically revealing respects? Both quartz and lemurs can be
described as matter but they are matter o f such different sorts that the
observation does not tell us much that is useful. Are the respective more-
than-real impressions proffered by film and dream very alike or very un-
alike?
Answering this question is difficult since Baudry tells us next to nothing
about the phenomena to which he wants to refer. However, if the impres-
sions he has in mind are a matter of imagery charged with affect, we can
remark that dream imagery is, if not just more vividly charged, it is at least
more invariantly charged than movie imagery. The reason for this, which
Metz notes, 18 is that the affect that attaches to dream imagery originates in
the dreamer and her personal associations whereas the affect derivable from
film imagery comes from an external source—such as the imaginations o f
screen writers and directors—which may or may not correspond to the film
viewer's emotive life. Thus, if the more-than-real impression o f reality o f
films and dreams is identified with a constant correlation of imagery and
affect, then the film apparatus and dream are very different. Moreover, if
the more-than-real impression of reality is not a matter of a constant
coincidence of affect and imagery, what is it? Merely occasionally exciting
imagery? But isn't that enough to correlate a televised chess game to
dreams?
But there is another way to probe the problems with Baudry's use of the
notion of the more-than-real impression of reality. In dream, this impres-
sion appears to refer to imagery charged with affect. In film, we are told that
this impression is one that diverges from our ordinary encounters with
mundane life. And, admittedly, the events we witness on film are most
often more exciting, more expressively characterized, and more emotion-
ally arresting than those of quotidian existence. However, it is important to
note that films of this sort, though common, are also very special. They are,
in the majority of cases, fiction films or they are films otherwise designed
explicitly to promote intense affective responses. A film like Greed, or
Sunrise, or Potemkin, or The Passion of Joan of Arc, may leave an impression
that is, as they say, "larger than life." But this sort o f impression is not a
28 PSYCHOANALYSIS: M E T Z AND BAUDRY

function of simply throwing an image on the screen. It is the internal

structure of these films that accounts for their effect, not the fact that they
are projected. Not all films bestow comparable affective results. Home
movies, or bank surveillance footage, especially of persons unknown to us,
may appear affectless, flat, and lackadaisical. That is, many films are pro-
jected, but few are chosen. Now this is an important point against Baudry.
For he claims that the more-than-real impression of which he writes is a
consequence of the cinematic apparatus, a claim tantamount to predicting
that whatever is projected onscreen will be swathed in affect. But this is
downright false; just recall Warhol's Empire.
Another major analogy between film and dream that Baudry produces
asserts that both have screens, or projection supports. It seems reasonable,
barring the complications of TV, to agree that films are normally projected
onto screens. But frankly the claim seems shaky in regard to dreams. The
evidence appears to be that some of Lewin's patients reported dream
screens in their nightly reveries. I have no reason to question their reports.
However, does this amount to evidence that something like a dream screen
is an essential or normal element in all dreams? Perhaps Lewin's patients
had personal associations with movie screens and this accounts for the
appearance of screens in their dreams. Why suppose that a screen element is
a characteristic feature of all dreams? Were there visions of screens in
dreams before there were screened entertainments? Were there visions of
screens in cultures without screened entertainments? And, furthermore,
what general criteria are there for establishing that phenomena, like the
appearance of a screen in a dream, are organic ingredients of dreaming
rather than the associative imagery of given dreams? Until these questions
can be satisfactorily answered the dream screen/film screen correspon-
dence—and with it the dream/film apparatus analogy—appear extremely
dubious. Of course, as I have already admitted, dream imagery is often
incomplete; but where the dream "picture" is unarticulated it is not neces-
sarily the case that the dreamer apprehends a screen, white, silver, or
otherwise. There is rather just a void.
So far Baudry's argument by analogy has been attacked by showing that
his analogies are hardly compelling and by remarking upon salient dis-
analogies between film and dream. The accumulated force of these objec-
tions shows that Baudry's argument by analogy is without substantial
warrant. I think the previously cited disanalogies are enough to swamp his
case. But also the premises that set out the analogies between film and
dream are virtually without support. 19 Thus, Baudry's argument fails to go
through. 20
P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y 29

Of course, questions, as well, might be raised not only about Baudry's

analogies but also about the crucial premise that asserts that the underlying
mechanism in dream is regression to primitive, oral narcissism. This is a
psychoanalytic claim, not a film theoretical one, and, as such, we should
probably not pursue this issue in depth here. Yet, it pays to remember that
the Lewin-Baudry hypothesis about regression is extremely controversial.
The regression hypothesis appears to ride upon the postulation of a
dream screen which, in turn, can be associated with a mother's breast. The
evidence for a dream screen as an organic, essential, or merely characteristic
element of dreaming has already been challenged. Insofar as the dream
screen serves as a linchpin for the inference of regression, and insofar as the
dream screen phenomenon is not generic, then there is no evidence for a
generic regressive mechanism, o f the specificity Baudry claims, in dream-
ing.
Also one must at least question the purported screen/breast association.
What is its basis? And how extensive is it? Maybe some white people
envision breasts as white and then go on to associate the latter with white
screens. But not everyone is white. And I even wonder if many whites
associate breasts and screens. Certainly it is not an intuitively straightfor-
ward association like that between guns and penises. For example, screens
are flat; and lactating breasts are not. A screen is, ideally, uniform in color
and texture; but a breast has a nipple. Nor will the association work if it is
put forward by saying that breasts are, for the infant in the state of oral
regression, targets of projections as are screens. For according to the theory
of primitive narcissism, the mother's breasts are part of an undifferentiated,
all-is-one experience, and, therefore, could not have been recognized way
back then by the primitive narcissist, and, thus, cannot be recalled now to
be targets of projection. For primitive narcissism admits no distinctions
between targets of projections, projections, and projectionists.
I do not deny that there may be some people who associate screens and
breasts, thereby at least suggesting the hypothesis of oral regression in
those cases. After all, it is probably psychologically possible to associate
anything with anything else. But even if some people associate breasts and
screens, that does not provide enough evidence to claim a general pattern of
association between breasts and dreams such as might support a theory
about all dreaming. And if oral regression is not the general causal force
energizing the dream, then it cannot be extrapolated by analogy as the
causal force behind the cinema apparatus.
Baudry's central argument in "The Apparatus" is beset by problems at
every turn. Do his subsidiary arguments fare any better? By analogical
30 PSYCHOANALYSIS: M E T Z A N D BAUDRY

reasoning, he links Plato's myth of the cave with dreaming, and also with
filmgoing, inferring that all three can be explicated by reference to regres-
sion. Plato's cave, furthermore, is identified as proto-cinematic, indicating,
to Baudry, that the instinctual desire that propels filmgoing is ancient.
The analogies between Plato's prisoners and dreamers are rather weak,
often in ways reminiscent of the problems with Baudry's film/dream analo-
gies. Both prisoners and dreamers are said to be immobile, in a darkened
place, bombarded with visual imagery, and unable to test reality. Contra
Baudry, we must note again, with reference to darkened places, that people
can sleep and dream in broad daylight, while the "world of the dream" need
not be dark. And, as was pointed out earlier, the dreamer as a character in a
dream need not be immobile, while the sleeper, unlike Plato's prisoners, is
unaware of being immobile. Also, what Plato's prisoners see differs radi-
cally from the dreamer's imagery—i.e., Plato's prisoners see uniformly
black figures whose only features are shadowy contours rather than inter-
nally articulated figures with eyes and moustaches. And, of course, Plato's
prisoners are awake while dreamers are not, which reminds us that Plato's
prisoners do literally see something—even if they misinterpret it—and this
indicates that they can objectively correct each other about the look of the
shadows before them, a type of reality testing not available to the dreamer.
One could go on at length discounting Baudry's dream-analysis of Plato,
but this appears to be more of an issue for historians of philosophy, if it is an
issue for anyone, than for film theorists. So let the preceding, hurried
refutation of Baudry's version of Plato's cave suffice.
Baudry also claims that Plato's myth of the cave is a préfiguration of
cinema. Needless to say this ignores the philosophical purposes Plato
designed the myth to serve. But Baudry believes that the myth evinces a
myth deeper than Plato was aware of. Baudry holds that Plato's cave is
proto-cinematic, which leads him to claim that cinema answers a desire of
ancient, instinctual origins. Many, more historically minded, film theorists
might wish to question the existence of transhistorical, transcultural desires
of the sort Baudry postulates. Nor is the existence of such a transcultural
desire absolutely integral to the project of psychoanalyzing the cinematic
apparatus. For cinema might be the answer to culturally specific desires of
the nineteenth and twentieth centuries.
Baudry's analogies between the cinematic apparatus and Plato's cave are
underwhelming. Both, purportedly, involve projection from behind the
spectator. But films are often rear-projected and early Japanese cinemas
positioned their audiences at right angles to the projection apparatus (not
to mention the possibility of projecting films via large video screens). Are
PSYCHOANALYSIS: M E T Z A N D BAUDRY 31

either of these practices uncinematic or do audiences at, for example, rear-

projected movies have different film experiences than those at cinemas with
projectors behind the audience. Baudry speaks of the projection in the cave
and in cinema as hidden. But film projection is not always hidden. One can
set up a projector in one's living room or in a classroom or boardroom, for
all to see, and still have a typical film experience. We have already noted the
film viewer is not necessarily immobile. Plato's prisoners are shackled at the
neck; even the seated film viewer can move her head. Furthermore, the
images of film are immensely different than Plato's shadows. The prisoners
see solid black blotches on the wall while film viewers see internally articu-
lated pictures. How, for example, could Plato's prisoner see the eyes of one
of the people who cast the shadows on their wall? And isn't this disanalogy
far more striking than the analogies that Baudry defends?
Plato's prisoners cannot enter the world of the shadows for purposes of
reality testing and neither can viewers of Casablanca belly up to Rick's bar
for a fast Scotch. But, on the other hand, film viewers can touch the screen,
and what is more important, they are aware that there is a screen, and that
they are watching a movie. Thus, their status, epistemologically, is exactly
opposite that of Plato's prisoners. Also, I think we would agree that one
could see a film that portrayed the world accurately whereas Plato's shadow
representations are putatively always deceptive. There are surely some
surface resemblances between Plato's prisoners and film viewers—they
both see projections. But this is hardly sufficient for supporting the claim
that both film and the myth of the cave address the same psychic need and
mobilize the same psychic mechanism.
Baudry believes that Plato's cave is part of the prehistory of cinema.
Apart from the inductive weakness of the analogies he draws, there is also
something strained in Baudr/s use of the notion of the prehistory of
cinema. Is every instance and/or report of shadow projection, prior to
1895, to be considered part of the prehistory of cinema? What criteria
determine that which we are to count as legitimately and four-squaredly
part of the prehistory of cinema? The camera obscura, Marev's repeating
camera, the praxinoscope, and Muybridge's batten,' of cameras are clear-cut
examples of what people include in the prehistory of cinema. And it is easy
to see that what these devices have in common is that they can figure in
causal accounts of the invention of cinema. But there is no historical
argument in sight to show that Plato's myth of the cave literally played a
role in the invention of cinema. Thus, even if Baudrys analogies were more
convincing, it is not clear that it would be appropriate to consider Plato's
cave as part of the prehistory of cinema.
32 PSYCHOANALYSIS: M E T Z AND B A U D R Y

The central argument, as well as the subsidiary ones, in "The Apparatus"

appear, upon scrutiny, to be utterly groundless. Undoubtedly, some
readers may complain that my methods o f examining Baudry's hypotheses
are not suitable and perhaps even boorish. For in my account, Baudrv is
making a series o f literal, logical, scientific claims about a causal process.
Thus, Baudry's conclusions are assessed with the rigor one would apply to
any scientific hypothesis. However, many contemporary film theorists
might object that Baudry is doing something more "interesting" or "impor-
tant" than pedestrian science.
Personally, I find it difficult to see why a claim about the isolation o f a
causal mechanism (in this case, regression as a motor force behind cinema)
should not be treated as a scientific hypothesis. And so treated, Baudry's
arguments by analogy are woefully inept; he fails to consider significant
disanalogies, while the analogies he presents are loose and superficial. But
many contemporary film theorists, especially those with backgrounds in
literature, may counter that that which I call "loose" and "superficial" is
really graceful, imaginative, and ingenious. They admire Baudry, as they
admire Barthes, for his supposed expertise in belle lettres. But however
seductive to some literary sensibilities "The Apparatus" may be, contempo-
rary belle lettres does not afford the means to defend causal claims about the
processes underlying cinema. Film theorists with backgrounds in literature
may bewail this fact; but it is unavoidable. Moreover, as we shall see, the
confusion o f belle lettres, on the one hand, with scientific and philosophical
reasoning, on the other, is one o f the most egregious problems in contem-
porary film theory. Indeed, the extremely detailed, literal-minded, and
argumentive style o f this chapter and of the rest of this book is mandated bv
my conviction that contemporary film theorists, with their penchant for
belletristic expression, including slippery analogies and metaphors, must be
shown that they are using the wrong tools for the tasks at hand.

METZ'S IMAGINARY SIGNIFIER

Christian Metz was the leading theorist o f the first semiology. At the
same time, his essay "The Imaginary Signifier" was one o f the seminal
contributions in the transition to the second semiology. In this essay, Metz
attempts to deploy psychoanalysis in order to explain both why people go
to the cinema, and how people are able to assimilate the rules and conven-
tions o f cinema. Thus, like other practitioners o f the second semiology,
Metz is motivated toward psychoanalysis as a means o f investigating the
conditions o f cinematic reception. Metz's purview has expanded from his
earlier, quasi-Saussurean concentration on the structure o f the cinematic
PSYCHOANALYSIS: M E T Z AND BAUDRY 33

sign to a concern with cinematic signification as it interacts with spectators.

Yet Metz does not regard the adoption of psychoanalysis as making a
totally radical departure from his earlier more classical, semiotic stance. For
psychoanalysis, he believes, like semiology, studies signification. Psycho-
analysis supplements semiology narrowly construed. Psychoanalysis can be
employed to continue the study of cinema in new areas.
Similarly, other practitioners of the second semiology agree that the
mobilization of psychoanalysis is not a research program that is discon-
tinuous with the Saussurean one. For the variant of psychoanalysis most
frequendy endorsed by psychosemiologists is Lacanian. And since Laca-
nian psychoanalysis has pretensions to linguistics a la Saussure, the transi-
tion to the second semiology is not conceived by its proponents to be a
rejection of the earlier phase but an expansion thereof.
"The Imaginary Cinema" proposes to tell us why we attend films and
how we are able to understand them by maintaining that cinema has roots
in certain unconscious phenomena, notably: imaginary identification, voy-
eurism, fetishism, and disavowal. The way that Metz identifies a species of
unconscious phenomenon as a root for cinema is by analogy. That is, an
unconscious phenomenon is a root or prototype of cinematic experience
where the unconscious phenomenon in question evinces an analogous
structure in what is called the play of presence and absence, as does a correla-
tive species of cinematic experience. Once Metz proposes a type of uncon-
scious phenomenon as a root or prototype of cinema, he can go on to use it
to answer his central questions. That is, in some cases, Metz can account for
our desire for cinema by suggesting that the medium can afford, in some
measure, whatever presumable pleasures, satisfaction of desire, and com-
pelling force that the psychic phenomena he has isolated provide the
unconscious; while, in other cases, Metz can argue that we arc able to
comprehend various cinematic practices in virtue of their having the perti-
nent and analogous psychic processes, structures, and syndromes as psychic
prototypes. In other words, we understand various cinematic structures
and phenomena because we have already encountered them in the course of
our psychosexual development.
Though the point of'The Imaginary Signifier" is really quite simple, the
essay is very difficult to read. One reason for this is that the essay does not
set out its central argument until it is nearly half finished. The piece opens
with meandering ruminations about cinephilia, emphasizing its dangers for
film theorists, and with elaborate statements of Metz's specific semiological
and psychoanalytic allegiances. One plows through over thirty pages before
we know where the essay is going.
Metz is at pains to differentiate his deployment of psychoanalysis from
34 P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y

other lines of cinematic research that might also invoke psychoanalysis. For
example, a psvchoanalvticallv inclined film scholar might analyze the per-
sonality of a filmmaker through his works, relating Hitchcock's attitude to
authority, say, to his relation to his father. Since Hitchcock's father was a
grocer, such a study might be initiated by examining the violence done to
and by means of food in Hitchcock's films. Or a scholar might attempt to
derive a typology of directorial syndromes, outlining the structure of Lan-
gian paranoia or the Wellesian obsession with power. But Metz rejects
these two approaches as theoretical options because they are concerned
with authors rather than the cinematic signifier. Yet another option might
be to employ psychoanalysis in the study of individual films conceived of as
textual systems. However, this approach, as well, does not come to grips
with the cinematic signifier as such, i.e., with the nature of cinema apart
from the unique characteristics of individual textual systems. Alternatively,
Metz writes "Another approach—the one I have had in view and have now
reached—consists of a direct examination, outside any particular film, of
the psychoanalytic implications of the cinematic."21
That is, Metz intends to isolate the film specific features and effects of
cinematic representation, and, then, to assess what psychoanalysis can
reveal about cinema by applying psychoanalysis to film's distinctive charac-
teristics. Specifically, Metz wants to unearth the psychic significance of the
essential feature of cinema in such a way that our questions about why we
go to films and how we understand films will be answered. Here it is
extremely important to note that though at times Metz has eschewed
essentialism, his methodology in "The Imaginary Signifier" is essentially
essentialist.
Metz's candidate for the essential feature of cinema is derived primarily
through a contrast with theater. Here, the essential, differentiating feature
between the two media, which captures Metz's attention, is that in cincma
what he calls the signifier involves a unique play of presence and absence.
That is, the cinematic signifier, or representation, is present to the spectator
(e.g., an image of a locomotive), but what it signifies or refers to (viz., the
locomotive itself) is absent, i.e., it is not in the screening room. This is
thought to contrast with theater where a character or a prop is represented
by something that is actually present to the audience, e.g., a living actor or a
chair. Metz contends
. . . what is true o f Sarah Bernhardt is just as true o f an object, a prop, a chair,
for example. O n the theater stage, this chair may, as in Chekhov, pretend to
be a chair in which the melancholy Russian nobleman sits every evening; on
the contrary (in Ionesco), it can explain to me that it is a theater chair. But
P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y 35

when all is said and done it is a chair. In the cinema, it will similarly have to
choose between two attitudes (and many other intermediate or more tricky
ones), but it will not be there when spectators see it, when they have to
recognize the choice; it will have delegated its reflection to them. 2 2

Having identified the play of presence and absence as the unique feature
of cinematic representation, Metz looks to psychoanalysis in order to
review what it has to tell us about the play of presence and absence,
especially with regard to our relations or potential relations to visual arrays
where those are marked by a play of presence and absence analogous to that
purportedly found in cinema. Metz finds analogous themes of presence and
absence in four types of psychic phenomena: imaginary identification,
voyeurism, disavowal, and fetishism. Metz proceeds to attempt to illumi-
nate the cinematic experience by reference to these psychic phenomena
conceived o f as prototypic analogs to the cinematic experience.
The first analog that Metzfindsfor the cinematic experience of the play of
presence and absence is imaginary identification. This phenomenon is a
crucial element in Lacanian psychoanalysis, a topic we will consider in
greater detail in the next chapter. In the present context, it is important to
recall that, according to Lacanian psychoanalysis, one of the momentous
points in our psychosexual development is what is called the mirror stage.
Purportedly, at around eighteen months of age, the child acquires a power-
ful sense of identity during an encounter with its own mirror image. The
child derives its sense of identity by identifying with its own reflection, an
experience occasioned by jubilance and a sense of perceptual mastery.
Cinematic representation is somewhat analogous to this seminal experience
since what is reflected onscreen and in the mirror is not literally the source,
i.e., the very referent itself, o f the reflection. That is, the image is present
but the referent is absent from the reflection itself.
What can this analogy be used to explain about cinema? First, presum-
ably the encounter at the cinema with a dynamical relationship of presence
and absence similar to that of the mirror stage triggers a corresponding
sense of perceptual mastery, thereby accounting for some of the pleasure we
derive from cinema.
Second, the invocation of the mirror stage explains why, according to
Metz, film viewers identify with the camera. This, in turn explains, why we
so readily accept certain cinematic conventions, such as camera pans.23 But
actually the engendering of identification with the camera explains much
more for Metz. For he believes that all communication requires identifica-
tion of some sort in order to be intelligible. Thus by triggering identifica-
tion with the camera, through the play of presence and absence (which
36 PSYCHOANALYSIS: METZ AND BAUDRY

harkens back to the mirror stage), a film secures a necessary condition for
appearing intelligible. Metz says

But with what, then, does the spectator identify during the projection of the
film? For certainly he has to identify': identification in its primal form has
ceased to be a current necessity for him, but he continues, in the cinema—if
he did not the film would become more incomprehensible than the most
incomprehensible films—to depend on that permanent play of identification
without which there would be no social life (thus, the simplest conversation
presupposes the alternation of the I and the jo», hence the aptitude of the two
interlocutors for a mutual and reversible identification). What form does this
continued identification, whose essential role Lacan has demonstrated even in
the most abstract reasoning and which constituted the "social sentiment" for
Freud . . . take in the special case of one social practice among others, the
cinematic projection?
Obviously the spectator has the opportunity to identify with the character
of the fiction. But there still has to be one. This is thus only valid for the
narrative representational film, and not for the psychoanalytic constitution of
the cinematic signifiers as such.24

What is the object o f identification o f the cinematic sign as such? By a

process o f elimination, Metz concludes it is the camera. This is an identifica-
tion facilitated by the replication o f the conditions o f identification—i.e.,
the play o f presence and absence—of that primal moment o f identification,
the mirror stage. This process o f identification with the camera is called (by
Metz) "imaginary," named after a faculty for identification, the Imaginary,
which Lacan alleges is acquired at the mirror stage. Insofar as cinema
reactivates this faculty, it is an imaginary signifier.
Along with the pleasures o f perceptual mastery, the cinematic play o f
presence and absence also supposedly affords voyeuristic pleasure. Accord-
ing to Metz, cinema recapitulates the theme o f presence and absence in-
volved in voyeurism insofar as the film spectator is absent from the fictional
world o f the movie. She wears, as Hofmannsthal would have it, a cloak o f
invisibility. Metz claims that in contradistinction to the theater, where the
actor is aware o f the presence o f the audience, the film spectator is not
present to the film actor. Metz adds to this that film viewing, like voyeur-
ism, is experienced in an essentially solitary way whereas theater is a far
more communal experience. 25
The psychic phenomena o f disavowal is also correlated to the cinematic
play o f presence and absence. "Disavowal" refers to the supposed infantile
coming to terms with castration. A male both believes but does not believe
that females have penises—or alternatively, that they both are and are not
castrated. This capacity for believing something is present while knowing it
PSYCHOANALYSIS: M E T Z AND B A U D R Y 37

is absent really underwrites the cinematic experience for Metz. This is how
we can react as though Sylvester Stallone is before us when we know that he
is not. Disavowal is the mechanism that makes this possible, thereby en-
abling us to comprehend cinema's most basic convention. 26
Connected to the psychic phenomena o f disavowal is that o f fetishism. In
place of the fantasized absent penis o f the female, the fetishist finds an
object—a foot, for instance—that stands for the missing phallus. In film,
Metz believes, the absence o f the image, the referent, is compensated for,
emotionally, by technique. This displacement is not only mobilized by
cinéphiles but by all filmgoers insofar as they esteem the well-madeness o f a
movie. Metz asserts:
The cinema fetishist is the person who is enchanted at what the machine is
capable of, at the theater of shadows as such. For the establishment of his full
potency for cinematic enjoyment he must think at every moment (and above
all simultaneously) of the force of presence thefilmhas and of the absence on
which this force is constructed. He must constandy compare the result with
the means deployed (and hence pay attention to the technique), for his
pleasure lodges in the gap between the two. Of course, this attitude appears
most clearly in the "connoisseur," the cinéphile, but it also occurs, as a partial
component of cinemadc pleasure, in those who just go to the cinema: if they
do go it is partly in order to be carried away by thefilm(or thefiction,if there
is one), but also in order to appreciate as such the machinery that is carrying
them away: they will say when they have been carried away that thefilmwas a
"good" one, that it was "well made" (the same thing is said in French of a
harmonious body). 27

"The Imaginary Signifier" is constandy cited by practitioners o f the

second semiology. For though it is not overly concerned with ideology, its
emphasis on cinema's effects on the spectator is compatible with the direc-
tion of analysis that ideological analysis is currendy supposed to take. But
though "The Imaginary Signifier" is accepted as authoritative, it is an
extremely problematic work.
Metz's method is hardly clear. Having identified film's essential feature,
Metz searches for the occurrence o f the analogous theme—that o f the play
of presence and absence—among psychic phenomena. But what principles
of selection lead Metz to exactly the four (rather heterogeneous) phe-
nomena he arrives at? Are there only four correlations? In answer to this,
Metz may say that he never claimed exhaustiveness. But, apart from this
trifling issue, it must be observed that the manner o f the way in which Metz
moves from the essence o f cinema to his analogous psychic prototypes is
murky.
There are no uniform, consistent principles stated that enable us to
38 PSYCHOANALYSIS: M E T Z A N D BAUDRY

reason smoothly from the essence of film to Metz's favored psychic phe-
nomena. The notion of the play of presence and absence is not applied
univocallv throughout the essay. In the treatment of voyeurism, for exam-
ple, the relevant absent element is the spectator who is not party to the
world of the film and who watches unobserved as Gregory Peck blows up
the guns of Navarone. But in Metz's other analyses what is significantly
absent is the source of the film's imagery—the guns of Navarone, for
example. Indeed, using concepts as vague as "presence" and "absence" with
little or no attention to what is present or absent in each case, allows for
correlations that seem no more than equivocating sleights-of-hand. And,
furthermore, why is it that when we come to fetishism, we are no longer
speaking of presence, but of the force ofpresence?
However, the lack of an explicitly stated, consistently applied method is
the least of Metz's problems. Greater difficulties arise with Metz's essential-
ism. First, one wonders whether he is correct in isolating the essence of
cinema primarily by a contrast with theater. Aren't the characters of novels,
such as Emily in Mysteries ofUdolpho, as absent to the reader in Metz's sense
as E.T., Mighty Joe Young, and Johnny Ringo are to the filmgoer? Metz
will undoubtedly say that he is working with the species of multi-channeled
visual arts and that he is seeking the differentia of cinema in that class. But
we shall see that it is his forgetfulness of the art of fiction that makes his
proposed differentiation between theater and cinema problematic.
Metz's account of the differential play of presence and absence in film as
contrasted to theater is suspect. For if we are speaking of fiction—i.e.,
fiction films and fictional plays—then, ontologically, Shylock is no more
present to the theater spectator than Fred C. Dobbs is present to the film
viewer. Neither Shylock nor Fred C. Dobbs can be hit by a disapproving
spectator with a dissenting tomato. I may stop a performance of Hamlet by
running onstage just as I might stop a showing of The Cabinet of Dr.
Caligari by plunging into the screen. But I cannot prevent Gertrude's death
nor cure Francis. I can't reach out and touch one of the yellow bricks on the
road to Oz; but then I can't wear Lear's crown either. Once we are
considering the realm of fiction, it makes no sense to speak of the differ-
ences between cinema and theater in terms of what is absent to the specta-
tor. In both fictional film and theatrical fiction, the characters are absent
from the continuum of our world in the same way.
There is no reason to think that the theoretical distinction between
theater and film can be drawn in the way that Metz desires, especially when
we recall that what Metz has in mind are fictions. The issue of presence and
absence which Metz raises has no relevance where what is being communi-
P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y 39

cated is first and foremost fictional. Admittedly, barring such cases as

cartoons in film and shadow play in theater, Metz's comments about
presence and absence have some connection to the most typical ways of
producing images in film and theater. But this contingent fact about the
way these fictional images are caused to be is of no ontological moment.
Nor is it clear why these features o f the production of fictional images
should have any psychological repercussions, particularly when what the
audiences in both film and theater are primarily concerned with are fictional
worlds.28
Further difficulties beleaguer Metz's analyses of the interrelations and
analogies between film and his supposed psychic prototypes. Even if, for
the moment, we ignore the question of whether the operative Lacanian
account o f imaginary identification has a secure scientific foundation, we
still have more than enough to trouble us with Metz's characterization of it
in regard to film. Why, one wonders, if there is a faculty of the Imaginary
would it be triggered by film? An encounter with a film is so different,
physically and phenomenologically, from an encounter with a mirror. We
do not see ourselves reflected from the silver screen. And even if we are
Richard Burton or Elizabeth Taylor, we do not see ourselves behaving or
looking as we are in the screening room as our celluloid simulacrum
unfolds. Metz appears to be aware of these disanalogies yet he pushes on
with the account o f imaginary identification. I don't understand why.
Perhaps it will be argued that in our culture the mirror is a metaphor for
visual representations such as painting, drama, and cinema. Hamlet advises
the players to hold a mirror up to nature. But why should this conceit cause
imaginary identification? Is our unconscious moved by literary images,
unspoken in the film? And what of spectators unfamiliar with this cultural
association? In any case, it is strange to think that such a cultural association
could stand in the place of the putative causal conditions of imaginary
identification.
Of course, it is undoubtedly the metaphoric association of mirrors with
certain types of art that attracted film scholars to Lacanian psychoanalysis in
the first place, since talk of mirror identification could be segued, in an
equivocating fashion, with the conceit that somehow film is a visual process
exactly like mirroring. But this, lamentably, only shows that much contem-
porary film theory is precariously based on metaphors. It does not show
that something as radically different from my mirror image of myself as a
film o f somebody else can serve as the efficient cause that mobilizes my
supposed faculty of the Imaginary in a way analogous to the alleged, primal
mirror experience.
40 P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y

Nor docs Metz appear on target when he describes our relation to the
camera in terms of identification. If I truly identified with the camera, I
suppose that I would experience the entire visual array of the projection as
coextensive with my visual field. Yet, when I look at a film image, I only
focus on part of it, usually upon what is represented in the foreground or
upon that quadrant of the screen where the primary action of the narrative
transpires. That is, often the camera's field of view is broader than mine; my
field o f vision is not coextensive with its field of vision. Moreover, this
disparity is frequendy used by film directors in a way that brings it to our
attention forcefully. In a horror film, the monster may be brought into the
soft-focused background of the shots just moments before the audience
sees it. The creature is visible onscreen before we are aware of it. But how
can it be that we identify with the camera when it "sees" more than we do?
Furthermore, it seems to me that we are often made aware of the fart that
the camera sees more than we do. In the deep-focus, realist style of cinema
theorized by Bazin, the spectator discovers important dramatic details in
the background of shots. This is a standard source of aesthetic pleasure in
film. But that sense of discovery hinges on finding something that was
already in the camera's field of vision. We could not savor this sense of
discovery if we took the camera's field of vision to be coextensive with our
own. For there would be nothing for us to discover visually.
Another way to make this point is to consider offscreen sound. Or-
dinarily, when I hear a noise close by I turn my head to identify its source.
However, it is quite common in film to register an adjacent offscreen
sound—e.g., the unearthly growls of some fiend or madman—without
cutting to a view of the source of the sound. Insofar as this is at variance
with our customary perceptual practices, we are aware that the camera's
"perception" is not ours, and identification with the camera would appear
to be blocked. The audience viewing Fritz Lang's M is obviously aware that
the camera is defying their normal pathway of vision by keeping Peter
Lorre just off camera for roughly the first half o f the film. How can they be
said to identify visually with that which palpably frustrates their characteris-
tic perceptual lvflexes. Nor is this use of offscreen sound esoteric; it occurs
throughout popular films as well as art films.
O f course, one might drop the idea that one identifies with the camera.
But Metz can't. For we have seen that Metz holds that all communication
requires a subtending process of identification. And he argues that camera
identification is the most plausible candidate for this role when it comes to
film. It is this commitment to the necessity for identification that drives
Metz to explain film reception in terms of imaginary identification with the
PSYCHOANALYSIS: METZ AND BAUDRY 41

camera. But I think that it is outlandish to accept the general supposition

that every communication, in order to be intelligible, requires some sub-
tending process of identification. I overhear a department store sales atten-
dent tell a pregnant woman that maternity clothes are on the second floor. I
understand these remarks, I find them intelligible, without in any meaning-
ful sense of the word identifying with either the attendant or the woman.
And even if identification were necessary it would be hard to come up with
compelling reasons why I would have to identify with the attendant rather
than the woman, or vice versa. Perhaps it will be argued that we must
identify with both; I find it more theoretically efficacious to say we identify
with neither.
Metz's account of voyeurism in cinema is as confused as his account of
imaginary identification. Undoubtedly voyeurism can be made a pressing
issue in the context of specific films. One thinks of Rear Window, The
Conversation, Blow-up, and Blow-out. But Metz believes that voyeurism is
relevant to the operation of allfilms.Why? Because the film spectator is not
in the presence of the film performer, which suggests that the film spectator
is viewing unobserved. This, as has already been stressed, inexplicably
reverses the way in which Metz usually discusses the play of presence and
absence. But the analogy between the film spectator and the voyeur also
seems wrong. Metz asks us to think of the film actor as if he were unaware
that he is the object of an ontologically absent audience—this in contradis-
tinction to the stage actor who is conscious of the presence of a breathing
audience in front of him and who is complicit in their act of watching. The
authentic conditions of voyeurism for Metz appear to require a victim who
does not know he is being watched. The film actor purportedly approxi-
mates this state but the stage actor does not. But does this make any sense?
Surely every film actor in typical films is playing for an audience—quite
knowingly I might add. Film actors are in no way like unwary apartment
dwellers who accidendy leave their curtains open and who are victimized by
prying eyes. Film actors are just as complicit as stage actors in their exhibi-
tion of themselves for popular consumption. Nor would it help matters to
say that film viewing is a situation in which the spectator pretends to be
voyeuristic, for how, in principle, could we be stopped from mounting the
same argument about theater viewing? And in any case, I, at least, doubt
that we usually pretend to be voyeurs at either movies or theatrical specta-
cles. That would require a mental act that I am sure we would all remember.
Metz claims that, like voyeurism, cinema is experienced as solitary
whereas theater viewing is communal.29 This observation is arrestingly
parochial. Perhaps film viewing in first-run cinemas in Paris is privatized.
42 PSYCHOANALYSIS: M E T Z AND BAUDRY

But, for example, ghetto audiences and teenage audiences (in the United
States at least) treat movies as opportunities for communal participation.
They crack jokes loudly at the expense of the film, currying the approval of
fellow viewers; they try to scare each other with premature screams; they
shout out criticisms and offer advice to the characters. The audience with
whom I watched Halloween kept warning Jamie Lee Curtis to stop daw-
dling and to get out of that house pronto. Also, middle-class adult au-
diences, as encountered in upper east side theaters in New York City, are
quite communal, talking to each other very audibly throughout the film.
Indeed, in my experience communal interaction is far more common, in
America, during movie screenings than during theatrical performances
(though this does not prompt me to say theatrical spectatorship is solitary).
Metz does say that his observations in "The Imaginary Signifier" are not
cross-culturally valid, and that they apply only to Western moviegoing. But
that caveat cannot be used to evade the preceding counter-examples. For
these examples are Western and they indicate that, among certain groups of
Western audiences, movies present an occasion for communal participation
or at least for a kind of raucousness which aims at mutual recognition.
Nor is Metz's characterization of film fetishism particularly persuasive.
There may in fact be some film fetishists. But I question whether all film
viewing involves fetishism. Putatively a fetishist fastens upon one object in
order to deny the absence of another object in whose presence the fetishist
has a stake. But what is the relevant absent object in the case of cinematic
representation? According to Metz, what is absent is whatever had been
filmmed—the desert, for example, in Lawrence ofArabia. But how many
viewers of the latterfilmhave a stake in the presence of the desert, and what
might that stake be?
Indeed, it is just implausible to attribute to normal spectators a desire for
the actual presence, in the screening room, of the objects, persons, and
events that they see represented on film. Consider what the desire to be in
the presence of the absent objects of certain films would amount to: being
amidst bullets whizzing by in Scatface; amidst cascading glass in The Posei-
don Adventure; amidst an attack of driver ants in The Naked Jungle. What
normal viewer literally has a stake in inhabiting such prospects?
Moreover, the supposed desire, that Metz postulates, of the viewer for
the actual presence of the referents of cinematic representations, is at
variance with one of the most crucial features of representation. That is, as
Aristode and Arthur Danto have stressed, we are interested in the represen-
tations of things in ways that we would not be interested in the things that
serve as the pretexts for the representations. We are, for example, interested
PSYCHOANALYSIS: M E T Z AND BAUDRY 43

in still lifcs of dead fish and not in the dead fish that modeled for them.
Representations facilitate cognitive and emotive responses that their refer-
ents "in nature" do not. Given this easily documented function of represen-
tation, is it likely that what consumers of representation really have a stake
in is the actual presence of the referent of the cinematic representation in
the screening room? Metz's account of the audience's desire for the pres-
ence of the absent referent conflicts with the cognitive and emotive pur-
poses for which we seek out representations rather than the "real things"
they portray. And surely we have more grounds for accepting the latter
account of representation than we do for accepting Metz's postulation of
arcane fetishistic desires.
Furthermore, what is the basis of the psychoanalytic association that
links film techniques as the substitute for the absent cinematic object?
When a foot fetishist substitutes a female's foot as a replacement for the
allegedly missing female penis two factors motivate the choice of symbol:
first, a sort of rough resemblance of the foot to the male organ and second a
metonymic relation, i.e., the foot is on the itinerary of the male gaze as it
travels up the female's body toward the genitals. But what motivating
factors promote the replacement of the absent planet Jupiter by the camera
movements in the "Star Gate" sequence of 2001 ?
In connection withfilmfetishism, Metz says it is the process of disavowal
that enables the film viewer to believe that the objects that cause the image
are present despite the fact that the spectator knows they are not. Some
such process as disavowal might be operative if it were appropriate to say
that in some sense film viewers characteristically believed, while simulta-
neously disbelieving, that Robert DeNiro is in the screening room with
them. But Metz has supplied no reason to postulate the existence of this
contradictory state in spectators. Specifically, he has not shown that specta-
tors in any way believe that the objects and persons shown in film are really
before them, "in the flesh," so to speak.
We shall see that there is a general tendency in contemporary film theory
to maintain that film spectators are rapt in the illusion that what is repre-
sented—the cinematic referents—are really present. This presumption will
be attacked throughout this book. In contrast, my own position is that the
spectator is aware that she is watching a film and does not mistake the
images for their referents. With movies we generally focus our attention
upon what is represented, though we are aware in a subsidiary way that it is
a representation to which we are attending. Rather than saying that we
believe and simultaneously disbelieve in the presence of the cinematic
referent, I would say, following Polanyi, that our mode of attention is
44 PSYCHOANALYSIS: M E T Z AND B A U D R Y

better characterized by way of two, simultaneous, ntmconfiicting modes of

awareness: a focal mode, directed at what is being represented, and a
subsidiary mode, through which we remain constantly aware that what is
before us is a representation. And if this characterization captures our
relation to cinematic representation, then we have no need to hypothesize
the operation of disavowal in film viewing for these states of conscious-
ness—focal and subsidiary—are not contradictory states of belief such that
one of them must be disavowed in order for the organism to maintain
equilibrium. That is, our relationship to the screen in terms of awareness is
not analogous to the fetishist and does not, therefore, call forth a corre-
sponding process of disavowal.

METZ'S DAYDREAMS
Metz's "The Fiction Film and Its Spectator" might be read as a corrective
to Baudry's 'The Apparatus." For Metz acknowledges certain crucial lim-
itations of the analogy between night dreams and films. However, he
believes that certain fundamental insights of the analogical approach can be
sustained by reorienting the analogy by linking film and daydream. This
allows him to conclude that part of our desire for cinema resides in its
capacity to externalize inner processes. He writes: "This is the specific joy of
receiving from the external world images that are usually internal images,
images that are familiar or not very far from familiar, of seeing them
inscribed in a physical location (the screen)." 30
This recalls Baudry's conclusion that through the cinematic apparatus
the unconscious manifests itself insofar as film simulates the dream. This is
gratificatory, signaling a return or manifestation of the repressed. That is,
one consequence that both Metz and Baudry draw from their analogies
between film and mental phenomena is that in simulating the unconscious,
film satisfies an unconscious desire that its internal processes be exter-
nalized.
In "The Fiction Film and Its Spectator," Metz embarks on a lengthv
examination of the analogies and disanalogies between film, daydream, and
night dream. As already noted, the analogy between film and night dream
has been with us at least since the twenties. And, of course, it is echoed in
popular idioms such as the "Hollywood Dream Factory." Some of the
grounds for the film/night dream analogy which Metz cites include that
both phenomena involve a darkened room, diminished mobility, involun-
tary reception, and, purportedly, lowered wakefulness. However, Metz
also notes a number of strong disanalogies between film and night dream:
P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y 45

the film spectator almost always knows that he is at the movies, whereas the
dreamer does not usually know he is dreaming; film perception is real—
there actually is an external reflection on the screen—whereas dream per-
ception is not real in this sense; film narratives are more rational than dream
narratives; and, connected to the preceding disanalogy, film is less viable as
a source of hallucinatory wish-fulfillment than one's own dreams.
On the other hand, the brunt of these disanalogies can be tempered
somewhat if we analogize film to daydreaming rather than to night dream-
ing. The daydreamer is aware that she is daydreaming just as the filmgoer is
aware that she is watching a movie. The narrative of the daydream is more
subject to secondary revision than is that of a night dream, resulting in
something that is less gappy and discontinuous, and more rational than the
night dream. This would, o f course, suggest that the daydream, like film, is
less effective than the night dream in affording hallucinatory wish-ful-
fillment.
The film viewing process, though not identical with the process of
daydreaming, is closer to the daydream than to the night dream. Yet, the
film, the daydream, and the night dream all supposedly resemble each other
because each putatively involves diminished wakefulness. Film, then, is said
to blend elements of night dream and daydream. But our perception of a
film is, in the sense already mentioned, real. This leads Metz to conclude
that part of the unique power of film is that it mixes elements of real
perception, daydreams and night dreams, accounting for the specific object
of joy of cinema, the external manifestation of things usually internal. Thus,
Metz offers an answer to the question of why people attend cinema and of
what desire cinema satisfies.
The major consideration that Metz offers for the analogy between film
and night dream is that both involve lowered wakefulness. Undoubtedly,
night dreams involve lowered wakefulness. But I see no reason for believing
that film diminishes wakefulness. Metz writes:
In contrast to the ordinary activities o f life, the filmic state as induced by
traditional fiction films (and in this respect it is true that these films demobil-
ize their spectators) is marked by a general tendency to lower wakefulness, to
take a step in the direction o f sleep and dreaming. When one has not had
enough sleep, dozing off is usually more a danger during the projection o f a
film than before or afterwards. 31

But is this dozing off case really convincing? The problem of lowered
wakefulness here cannot be attributed to the film but to the fact that the
viewer is already fatigued. If he fell asleep eating his soup or reading a
newspaper would we say that soup and newspapers mix elements of night
46 PSYCHOANALYSIS: M E T Z AND BAUDRY

dream. Of course not. Metz indicates that a fatigued person is more likely to
fall asleep during a film than prior to or after the film. He gives no reason
for believing this. Moreover, I see no grounds for thinking that given two
subjects, equally fatigued, the one watching a movie and the other reading a
novel, that the moviegoer will be the one who falls asleep faster.
How does Metz know that films diminish wakefulness? He has not even
attempted such simple empirical corroboration as measuring heartbeats, or
breathing, or brain activity. Instead he notes that the narrative film does not
incite action and adds:
it encourages narcissistic withdrawal and the indulgence of phantasy, which,
when pushed further, enter into the definition of dreaming and sleep; with-
drawal of the libido into the ego, temporary suspension of concern for the
exterior world as well as the cathexis of objects at least in their real form. 3 2

Now I am not sure that film in fact encourages all these things. For if one
is attentively focused on what is going on in a film, one has little time for
narcissistic withdrawal or fantasy. If you are concerned with whether James
Bond will thwart the flooding of Silicon Valley, you are too preoccupied
with the action to drift off into narcissistic fantasy or any other kind of
fantasy for that matter. Of course, Metz may say that following a James
Bond exploit is itself a form of fantasizing. But such a maneuver would
seem to me to be a misapplication of the category of fantasy, and a ques-
tion-begging one at that.
What is even more peculiar in Metz's account here is that, granting him
the unlikely premise that films cause narcissistic withdrawal and fantasy,
one still wants to know what these have to do with lowered wakefulness.
Fantasizing, for example, often involves planning, imagining in detail, for
instance, elaborate measures of revenge. Surely, one can sit silently plan-
ning without having one's wakefulness lowered. Metz suggests that nar-
cissistic withdrawal and fantasizing are definitive of lowered wakefulness.
But that is a strange use of the notion. I can sit envisioning my acceptance
of the Academy Award for Best Actor and be in a very excited state
physiologically. And lowered wakefulness is, first and foremost, a phys-
iological state. To stipulate that fantasizing constitutes lowered wakeful-
ness would, of course, seem to attempt illegitimately to setde a question of
fact with a definition.
Returning to the issue of film, we note that Metz appears to think that
the facts—that films do not incite us to action, and that when we attend to
the screen, we tend to ignore other portions of the world around us—
suggest that the film-viewing state is one of lowered wakefulness. But, of
PSYCHOANALYSIS: M E T Z AND B A U D R Y 47

course, the film does promote a kind of response or action, one appropriate
to it, viz., film viewing. Does Metz think that film viewing correlates with
lowered wakefulness in any straightforward sense? Are surgeons nodding
off when they view medical instruction films? Metz appears to believe that
films correlate with lowered wakefulness because when we are at a film, we
usually ignore what is happening in our environmental surroundings. But
that is called paying attention, not lowered wakefulness.
Does Metz believe that his wakefulness is lowered when he, riveted to a
semiotics lecture, is unheeding of those who leave the room or shuffle
papers? If not, then why should he suppose ordinary film viewers, display-
ing the same sort of concerted attention, are in a state of lowered wakeful-
ness? Moreover, if Metz is not in a state of lowered wakefulness when he
attends a semiotics lecture, would he fall into such a state were he to view a
film of the self-same lecture? Not likely. Metz might object to the preceding
thought experiments on the grounds that he is speaking of fiction films, not
lecture films. But then, let us place the semiotics lecture in the context of a
fiction film—the semiotics professor is a part-time secret agent just as
Indiana Jones is a part-time adventurer, and a semiotics lecture is part of an
establishing scene. Will Metz's wakefulness be lowered as he attends to
Indiana Saussure's fascinating syntagmas? Metz's claims about the lowered
wakefulness of film viewing sound false and are completely unsubstanti-
ated. To the extent that the assertion that cinematic representation pos-
sesses elements of night dream depends on the hypothesis of lowered
wakefulness, Metz's thesis is altogether without proof.
Metz claims that film contains elements of daydream. Thus, when we see
a film we are gratified by encountering the externalization of something
generally internal. Insofar as thefilm/daydreamconnection depends on the
notion of lowered wakefulness, it must confront the preceding arguments.
But Metz, of course, has analogies other than lowered wakefulness to
support thefilm/daydreamanalogy. Some of these are not very compel-
ling—the filmgoer is aware he is filmgoing and the daydreamer is aware he
is daydreaming—but then, of course, the beer drinker is aware she is beer
drinking, and so on. Indeed, as the latter comment shows, Metz's analogies
between film and daydreaming are all rather weak because they are not
straightforward analogies but analogies selected relative to the weaknesses
of the night dream/film analogies. For example, film narrative is more like
daydream narrative than night dream narrative. But how directly analogous
are film narratives and daydream narratives? Enough to make the analogy
informative?33 That is, a systematic problem with Metz's reasoning in "The
Fiction Film and Its Spectator" is that most of the analogies he draws
48 P S Y C H O A N A L Y S I S : M E T Z AND B A U D R Y

between film and daydream are not really tight; they are just a little tighter
than certain film/dream analogies. But that does not amount to much
theoretically.
However, even if we accept Metz's analogies between film and dav-
dream, we must still ask whether these analogies will be adequate to
support Metz's major claim, that what is gratifying about film is that it
presents us with a type of visual experience that we recognize as primarily
internal.34 Daydreaming is a talent that is acquired. Not everyone does it.
Children apparendy learn to do it through play and through stories. Gradu-
ally, they learn to internalize their games with dolls and their stories, often
picture stories (which in our culture would include movies and TV).
Daydreaming appears to be the internalization of externalized forms of
representation such as play and narrative, including narrative movies and
T V 35 What is strange about contemporary film theorists who, extrapolat-
ing from Mecz, want to explain how we understand cinematic conventions
by means of invoking the putatively prototypical structures of daydream-
ing, is that, in all probability, the narrative devices of daydreaming, in many
cases, derive from pervasive cultural forms of narrative such as film. Thus,
when we encounter films we are not encountering something that is gener-
ally internal, but rather structures of representation found both internally
and externally and whose provenance is most likely external.

CONCLUSION
Apart from the questions of detail, which we have pursued at length, one
also may doubt the advisability of the overall theoretical strategy endorsed
by Metz and Baudry. Both Metz and Baudry propose to tell us about the
nature of film by analogizing it to the mind. This places them in a long
tradition offilmtheorists which probably begins with Hugo Munsterberg's
Film: A Psychological Study. Munsterberg analogized film to various cogni-
tive processes of rational thought whereas Metz and Baudry and others
analogize film to irrational processes. The conceptual foundation of these
approaches, however, are roughly the same; they differ primarily in the
aspects of the mind that they choose for analogy. Nevertheless, even if there
is a long tradition of mentalistic analogies in film theory, it is not clear that
this tradition has much to recommend it. For we must ask if the analogy
approach—whether employing rational or irrational analogs—is very prof-
itable, since, in fact, so litde about the mind and its processes is known.
For an analogy to be informative, we should know more about the item
that is meant to do the illuminating than we do about the item that is
PSYCHOANALYSIS: M E T Z AND B A U D R Y 49

supposed to be illuminated, e.g., we should know more about dreams than

we do about films. This is basic to the logic of analogy. But I am not sure
that this condition is met byfilm/dreamanalogies. Indeed, I suspect that we
probably know more about the workings of film than we do about the
workings of the mind. Undoubtedly, this will sound strange to film theo-
rists. But it may be an occupational hazard on their part to believe that
cinema is more mysterious than it is. We know quite a lot about how film
works, about why it works, about its conventions, and about its techniques.
Far less is certain about the mind in either its rational or its irrational
aspects. Again, I do not deny that specific films, like Last Tear at Marien-
bad, can try to mime the mind's operation in such a way that it is appropri-
ate for a critic to call attention to attempted analogies with thought. But as
a theoretical project, I think that film/mind analogies have little to tell us
given our present state of knowledge of the mind, both rational and
irrational. How much do we learn by being told that films are like such
things as daydreams or night dreams when we know so litde about dream-
ing? We do not even know why we sleep. Dreaming is far more mysterious
than cinema.
The reason, of course, thatfilmsare not mysterious is that we make them.
We make them to work a certain way and in the majority of cases they work
the way we designed them to work. In general, we understand our own
tools and inventions better than that which we have not created.36 That is
not to say that we understand our creations perfectly, but only that we
know a great deal about them in virtue of making them to perform those
services that they successfully perform. At present, the computer, a product
of our invention, is being used by cognitive scientists as a model or analog
to the mind. This is a profitable strategy because having designed com-
puters, we know a great deal about them, and we can extrapolate that
wealth of information to mental operations. In the past, theater and even
film provided useful analogs for the mind though ones not so powerful as
those of artificial intelligence. This is not to say that in the long run theories
of mind built on analogies to theater, film, or artificial intelligence will be
found to be true, but only to say that as research programs they have the
right logical structure whereas analyzing film through mind analogs does
not.
A second major methodological problem with the approach adopted by
Metz and Baudry is that it presupposes that cinema—the cinematic appara-
tus or cinematic representation—irrespective of specific films is an appro-
priate subject for psychoanalysis.37 Arguments to justify the use of psycho-
analysis to explicate cinema as such are remarkably absent in Metz and
50 PSYCHOANALYSIS: M E T Z AND BAUDRY

Baudrv. That is, it is never asked whether we should attempt to psycho-

analyze cinema as such.
Of course, if one believes that everything is ripe for psychoanalysis, then
one would not require an argument for its application to anything of
human origin, including the cinematic apparatus. But the pressing meth-
odological question is whether things like the cinematic apparatus or cine-
matic representation are licit topics for psychoanalysis. Now it seems to me
that not everything that is human is grist for the psychoanalytic mill. For
example, the explanation of the particular construction of a given bridge in
terms of psychoanalysis would be bizarre. We would seek, instead, explana-
tions of the construction of the bridge by reference to traffic, commerce,
patterns of exchange, and the stare of the art of bridge building. And if we
wanted to explain the existence of bridges as such, we would do it bv
reference to certain general facts about geography, human society, eco-
nomics, and so on. It would be peculiar to answer questions about bridges
by reference to psychosexual symbolism (though bridges may have psycho-
sexual portents in someone's dream).
But if some phenomena bear psychoanalyzing and others do not, what
criteria guide us these matters? And are things like the cinematic apparatus
and cinematic representation such that our criteria indicate they should be
analvsands or, rather, should they be in the same category as bridge con-
struction?
To specify the criteria for the appropriateness of psychoanalytic explana-
tions, we must recall that psychoanalytic theory is designed to explain the
irrational. The general paresis and epileptic fits, due to injury of Broca's area
in the brain, are wonrational and not a subject for psychoanalytic enquirv.
Similarly, when an agent does something that is rational, we have no prima
facie reason to investigate into the psychoanalytic causes of his behavior.
That is, a methodological constraint on psychoanalytic explanation is that it
not be mobilized until there is a recognizable breakdown in rationality
(which cannot be explained in terms of the intervention of some nonra-
tional process).
Given this requirement, which is mandated by the very concept of
psychoanalysis, it is clear that not all beliefs, not all emotional, social,
aesthetic, and cognitive responses are candidates for psychoanalytic inves-
tigation. Insofar as psychoanalysis is designed to conceptualize irrational
behavior, which is only identifiable as a deviation from rational behavior, there
is no work for psychoanalysis to do where the behavior is of an unmistak-
ably rational sort. That is, where adequate rationalistic explanations are
available, we do not require psychoanalysis.
PSYCHOANALYSIS: M E T Z A N D BAUDRY 51

So to mobilize psychoanalysis, we must show that the phenomena or

behavior in question are irrational rather than rational. At the very least,
one problem with Metz, Baudry, and their epigones is that they have not
bothered to show that such things as the cinematic apparatus or cinematic
representation are irrational or that our interactions with them as viewers
are irrational. More importantly, it is not clear that anyone could show that
these phenomena are irrational. The cinematic apparatus, as Baudry calls it,
was created through a process of invention that applied scientific reasoning
to the realization of a purpose, the projection of moving images. There was
nothing irrational in the technological solutions that resulted in the de-
velopment of the celluloid film strip or the projector gate. Nor was the
move from the early arcade movieolas to screened film prompted by an
irrational longing for the mother's breast; it is explained by the degree to
which it promised greater profitability.
Perhaps it will be urged that even if the development of the screening
apparatus was not an irrational process, but a matter of finding means to
an end, the end itself, the purpose which propelled that development, was
irrational. That is, the aim of having projected, moving images is irra-
tional. But this is rather implausible. The Lumières, for example, em-
ployed projected pictures to disseminate information about the world.
What is irrational about that? And given the way humans are built, the
communication of information by projected moving images seems emi-
nently sensible. Nor does the ensuing adaptation of cinema for entertain-
ment and artistic purposes seem irrational, unless one believes that enter-
tainment and art are necessarily irrational. Similar objections can be
leveled at Metz's unquestioning mobilization of psychoanalysis to investi-
gate cinematic representation. For it is not clear at all that the practice of
cinematic representation—the denotation of absent referents by means of
moving images—is inherently irrational. Given the way humans have
evolved it is a brilliant means of communication. The burden of proof here
is with Metz and Baudry. They must show that the cinematic medium is
irrational in origin or effect before they start psychoanalyzing it.
Now I do not intend to deny that individual films may be irrationally
constructed or that specific films or types of films mav elicit irrational
responses from spectators. And in such cases psychoanalysis may be an
appropriate tool for analysis. But this admission does not serve to warrant
the approaches Metz and Baudry advance. For they arc not speaking of
specific films or groups of films. They propose to analyze the apparatus as
such and cinematic representation as such, i.e., regardless of what films or
types of films are being shown. And this requires a demonstration that the
52 PSYCHOANALYSIS: METZ AND BAUDRY

structure of and interaction with the apparatus and cinematic representa-

tion is inevitably irrational. But neither Metz or Baudry has established
this, nor do I think either can. For the cinematic projection apparatus and
cinematic representation are calculated technological inventions, not fan-
tasies, which were developed to serve tangibly rational purposes. Indeed,
given the role in modern communication that these structures serve and the
importance of such communication for human social life today, it seems
perverse to think they are necessarily irrational. The notion that representa-
tion per se is irrational, given the way humans are built and the way they
communicate, is absurd. The practice of representation, like that of bridge
building, is a normal part of human life.
Without question, there is a tendency in psychoanalytic theorizing to
regard every aspect and institution of human existence as afieldfor psycho-
analytic speculation. And Metz and Baudry belong to this tendency. But
just because there is a tendency to ignore the obvious constraints on
psychoanalytic theorizing discussed earlier, that does not legitimatize such
speculation. Such theorists should restrict their field of study to the phe-
nomena which can be plausibly demonstrated to pertain to the irrational.38
And that, I submit, excludes cinematic representation as such and its
projection apparatus from the psychoanalytic ballpark. To think otherwise
is a bit of psychoanalytic imperialism and we shall see other variations of
this imperialism in the ensuing chapters.
 2.

MARXISM AND
PSYCHOANALYSIS
THE ALTH US SE RIAN - LACAN IAN
PARADIGM

T H E aspirations o f the Althusserian-Lacanian approach in contempo-

rary film theory must be understood against the broad background o f
Marxist aesthetic theorizing in the West. Roughly speaking, Western
Marxist art theory has been preoccupied with two large problems. The first
is whether art can perform an emancipatory role in the class struggle, and, if
it can, by what means can this role be effected. Some, like Lukacs, propose
realism as the means for emancipation, while others such as Brecht and
Benjamin advocate disjunctive, avant-gardist techniques. Adorno affirms
that works o f modernist art have a subversive capability; however, he is
exceedingly pessimistic about the effects this can bring about in societies
dominated by mass culture.
The second problem area theorized by Marxist philosophers o f art starts
with a question that is not specifically aesthetic, although the answer it
receives involves a theory o f the role and structure o f mass art and mass
media in modern, Western societies. The question is why haven't the
masses o f the industrial West revolted against capitalism or acted in other
ways in order to transform their societies in the direction o f communism?
Why, that is, have Marx's predictions failed? One way this question is
answered is by reference to ideology. Capitalism, through its mass popular
art industry—the movies, T V , radio, popular music, and so on—confuses,
mystifies, and manipulates our minds in such a way as to thwart the
development of emancipatory consciousness. Marcuse writes:
Marxian theory soon recognized that impoverishment does not necessarily
provide the soil for revolution, that a highly developed consciousness and
54 M A R X I S M AND PSYCHOANALYSIS

imagination may generate a vital need for radical change in advanced material
conditions. The power of corporate capitalism has stifled the emergence of
such a consciousness and its imagination: its mass media have adjusted the
rational and emotional faculties to its market and its policies and steered them
to a defence of its domination. 1

Given the view that ideology, as disseminated by capitalist mass culture,

is a major force in stultifying emancipatory consciousness, the Marxist
aesthetician, or cultural theorist, proceeds by isolating the features of mass
art and media through which our minds are manipulated. Adorno, for
example, argues that the standardization of art for the purposes of mass
distribution contributes to the "standardization" of the art consumer while,
at the same time, reinforcing the capitalist myth of individualism. Thus, the
consumer behaves in a regimented manner while believing that her be-
havior is a matter of her own personal choice. Speaking of the composition
of popular music, Adorno says:
He (the composer) must write something impressive enough to be remem-
bered and at the same time well-known enough to be banal. What helps here
is the old-fashioned individualistic moment which in the production process
is voluntarily or involuntarily spared. It corresponds as much to the need to
hide the all-governing standardization of form and feeling from a listener
who should always be treated as if the mass product was meant for him
alone.2

The question of why the Western working class fell short of Marx's
predictions has been with Marxists at least since the thirties, as has been the
hypothesis that the ideological operation of mass culture adequately an-
swers this question. Of course, different Marxist commentators have of-
fered different theories of the ways in which mass culture works in dis-
seminating ideology. Althusserian Marxism differs from Marcuse's variant
in many details; however, they can both be seen as attempts to answer the
same question in roughly the same ways, i.e., by invocation of the mind
manipulating capacities of ideology in the service of capitalism.
Historically, it is the case that the influence of Althusserian Marxism in
film theory has been primarily as a tool for analyzing the operation of
ideology. Althusserian Marxism also tries to give an answer to the question
of the emancipatory capability of art, but, as we shall see, it is unclear
whether this account makes sense in the context of the view of ideology that
Althusserians propound.
If the problem of the docility of the working class has been with Western
Marxists since the thirties, it is also true that it was felt to be particularly
pressing in the late sixties and early seventies when contemporary film
MARXISM A N D PSYCHOANALYSIS 55

theory began to consolidate. The failure of the high hopes of the New Left
forced Western Marxists to offer an account of their lack of success in
encouraging the working class to join with student protesters and life-style
revolutionists in the overthrow of the capitalist establishment. The hypoth-
esis of ideological, mind-manipulation, of course, presented a ready frame-
work with which to explain the failure of the New Left. And the Al-
thusserian version of this approach became the preferred method of
ideological analysis for contemporary film theorists.
Americans radicals were probably predisposed to ideological, mind-
manipulation hypotheses. Marcuse was a presiding figure for the American
New Left and he explicidy endorsed such a hypothesis. As well, the post-
World War II generation of Americans which was to comprise contempo-
rary film theory was brought up on the immensely popular exposes of adver-
tising of Vance Packard who, from an non-Marxist perspective, convinced
many that capitalism was surreptitiously molding our consciousnesses from
every direction. 3 At the same time, mind-manipulation hypotheses are also
in a way quite flattering to film theorists. For in identifying ideology, as
disseminated by such things as mass movies, as the key to the riddle of the
working class, one's analysis of film acquires vast political significance. This,
however, may reveal one of the shortcomings of the contemporary theory of
ideology, including especially the Althusserian variant. For the case for
supporting mind-manipulation theories may invest ideology with more
power than it actually has. Or to put the point differently, ideology may not
be what stops the working class from resisting capitalism. But these are
issues to which I shall return in the conclusion of this chapter when I assess
the overall advisability of the Althusserian explanation of ideology.
Contemporary film theory is a product of the sixties. Undoubtedly, it is a
beneficiary of the movie craze of that decade. Especially in America, a post-
World War II generation, nurtured on old films on TV, came of age,
extolling movies as a central cultural icon and encouraging the amateur and
professional study of film. Coincident with this boom of interest in cinema,
of course, were the major political and cultural upheavals of the sixties and
early seventies. Not surprisingly, many of those who participated in the
movie boom were also, at the same time, immersed in radical politics. As a
result, there was a predictable tendency to attempt to combine prevailing
interests in movies and politics. Thus, the desire was spawned to create a
Marxist study of film, one aimed at explicating the ways in which film
instilled ideology in its spectators, a task rendered more urgent, as already
noted, by the belief that ideology represented a central problem to be
solved by Marxists.
However, such a project, perforce, required some grasp of the psychol-
56 MARXISM AND PSYCHOANALYSIS

ogv o f spectators. And, though Marxism had a notion o f ideology, it lacked

a settled account o f the way in which that ideology was implemented
psychologically.4 It was in this context that the writings o f Louis Althusser
became increasingly important for film researchers. For Althusser sug-
gested an approach that incorporated psychoanalytic theory—notably that
o f Jacques Lacan—into the characterization o f ideology. Film researchers
seized upon, while sometimes also modifying, Althusser's framework, and,
as well, took Althusser as a precedent whose work legitimatized further
appropriations o f Lacanian psychoanalysis in the name o f Marx. The pur-
pose o f this chapter is to introduce an overview o f Althusser's approach to
ideology, emphasizing the putative linkage with psychoanalysis. Mention
will also be made selectively o f expansions on this approach by contempo-
rary film theorists. As well, I will examine certain Lacanian characteriza-
tions o f language and o f psychosexual development where those have had
significant impact on film studies. The chapter should not be mistaken as a
primer on Althusser and Lacan. Themes have been drawn from their work
on the basis o f their influence on contemporary film theory.

ALTHUSSER AND IDEOLOGY: MARXISM AND T H E

INVOCATION OF PSYCHOANALYSIS

Althusser has made contributions in numerous areas o f Marxist philoso-

phy, including the philosophy o f science (broadly construed), historiogra-
phy, and the theory o f ideology. In several ways, these various concerns are
often interlocking. For example, Althusser's accounts o f science and ideol-
ogy would appear to be mutually interdependent. Yet, most contemporary
film theory, in terms o f what it uses from Althusser's work, relies primarily
upon his writings on ideology, effectively in isolation from his larger
cpistemological projects. This is not to say that contemporary film theorists
necessarily demur from Althusser's other commitments (though some do),
but rather that they do not materially and explicitly deploy Althusserian
epistemology and historiography in their film analyses. Instead, the ideo-
logical analyses are focal, especially those found in "Marxism and Human-
ism," "Freud and Lacan," and, most importantly, "Ideology and Ideologi-
cal State Apparatuses (Notes towards an Investigation)." 5 Consequently, I
shall concentrate on these essays, along with making some comments on
Althusser's essays on art. 6 Also, I shall try to indicate the way in which
Althusser's treatment o f ideology afforded an opportunity for film theorists
to apply Lacanian psychoanalysis to the study o f film ideology even though
Althusser himself did not develop such an analysis.
MARXISM AND PSYCHOANALYSIS 57

Althusscr's approach to ideology is essentially functionalist.7 In order to

continue its existence over time, any social formation must not only involve
itself in production—i.e., the production of goods and services to meet its
current needs; the social formation must also be engaged in a process of
reproduction. That is, it must reproduce itself in order to sustain itself. This
requires not only the replacement of depreciating machinery and dwindling
resources, but also the replacement of the labor force. Moreover, the labor
force—indeed the entire population of the social formation—must not
only be physically reproduced through the supply of vital necessities; the
population must also be reproduced in such a way that the roles, values, and
even the metaphysics requisite to the social formation remain relatively
stable from generation to generation.
In capitalism, the state is the nodal point for the reproduction of the
social formation. It secures the reproduction of the social formation by
means of two sorts of mechanisms which Althusser calls "apparatuses":
Repressive State Apparatuses, e.g., the army and the police, whose medium
primarily is force; and Ideological State Apparatuses, whose medium might
be said to be persuasion, or, less informatively, ideology. Ideological State
Apparatuses include the family, religion, political parties, the media, the
arts, and even trade unions.8 The legal system has aspects of both the
repressive and the ideological systems. Moreover, full reliance upon the
Repressive State Apparatuses (RSAs) is rare because the pervasive working
of the Ideological State Apparatuses (ISAs) guarantee that people will
behave in the ways that the underlying structure of the social system
demands.
All social formations require ideology. Althusser writes: "Human so-
cieties secrete ideology as the very element and atmosphere indispensable to
their historical respiration and life."9 That is, every social formation prof-
fers a range of roles and values to which the population accedes. Moreover,
the relation of these roles and values to the real conditions of the social
formation is imaginary.10 That is, the actual function of a given social role is
always misrecognized or misperceived by the individual due to the machi-
nations of ideology; the reformist politician, for example, thinks he is
benefiting the oppressed whereas he is really whitewashing capitalism.
Though all social formations deploy ideology, the organization of social
forces may change over time. Under feudalism, the church and the family
were the primary levers of ideology, whereas under contemporary capital-
ism, the educational system (rather than the church) and the family are
dominant. Significandy, Althusser includes under the umbrella of Ideologi-
cal State Apparatuses such often private enterprises as the media and such
58 MARXISM AND PSYCHOANALYSIS

private institutions as the family. Whether he is warranted in these classi-

fications is a matter to which we will have to return.
The ISAs reproduce the roles and values of the capitalist social forma-
tion. The educational system and the family reproduce a populace with the
values and attitudes that facilitate the acceptance of the roles the capitalist
system needs filled in order to sustain itself over time. That is, ideology
functions to reproduce capitalist subjects. The ambiguity of the notion of a
subject, here, is crucial. For Althusser, it refers both to one's belief in oneself
as a unity, an autonomous "I," the center o f one's own experience, and as
the source of free action; and it refers to one who is subservient to some
system of domination. That is, Althusser's account plays off of associations
of "subject" with autonomy, on the one hand (subject = the self as free
agent), and with associations with servitude (e.g., "I am subject to your
will"), on the other. The capitalist subject, who believes himself to be an
autonomous, unified, free agent, is really dominated. He misrecognizes his
acts as free acts when they are really implementations of the dictates of the
social formation. Althusser writes that through ideology the subject is
constituted or constructed by society as

a (free) subject in order that he shall submit freely to the commandments of the
Subject, i. e., in order that he shallfreelyaccept his subjection, in order that he shall
make the gestures and actions of subjection "all by himself." There are no
subjects except by and for their subjection. That is why they "work all by
themselves."11

Here we see that Althusser believes that central to the reproduction of

capitalism is the reproduction of subjects with mistaken views of them-
selves. This is not unlike the view of critical theorists, like Adorno and
Marcuse, who think that our view of ourselves as making free choices is the
ideological grease that lubricates our adoption of behaviors preordained by
capitalism. According to Althusser, we are subjects who believe we are
autonomous agents, but whose "free" choices have really been engineered
by ideology. So, for Althusser, it is a fundamental task of capitalist ideology
to promote our misapprehension or misrecognition that we are free, auton-
omous agents.
But how does ideology achieve this? How does ideology instill a sense in
me o f a certain type of agency? How does it construct me as a subject, i.e., as
a certain kind of subject, viz., as a subject who mistakes his subjection for
free action? It is at this point that psychoanalysis enters the picture. For
psychoanalysis, especially in the fashionable formulation sponsored by
Lacan, is reputedly a discipline that studies subject construction. Thus, at
MARXISM AND PSYCHOANALYSIS 59

least putatively, the analytical categories of Lacanian psychoanalysis can be

marshaled by Althusser and his followers to describe the different phases
and dynamical interrelations that result in the construction of subjects in
general as well as subjects possessed of certain misrecognitions concerning
the nature of their agency.
What does this have to do with film? Well, recall that media is an ISA.
Thus, film is a means of propagating ideology, or a means through which
ideology functions, which in the Althusserian dispensation means it is at
least a device through which subjects, in his special sense, are produced.
Moreover, insofar as the construction of subjects is a necessary condition
for the adoption of roles, this operation of ideology is thought to be rather
foundational. So the film researcher, taken by this approach, identifies her
task as the discovery of the ways in which film operates ideologically to
construct subjects. That is, the cinematic image, narrative construction, film
editing, sound, and so forth are to be analyzed in order to show how these
elements of cinema have been employed to facilitate subject construction,
or, more specifically, the construction of the kinds of subjects with at-
tributes congenial to the continuation of capitalism.
But how does something like film editing participate in the fabrication of
"agents" who take, or rather mistake, themselves to be free? Here we must
return to the details of Althusser's account of ideology. Althusser sees a
prime function of ideology to be the construction of subjects. Society has
various means to this end. These include not only government with its laws,
courts, and cops, but also the educational system, religion, the arts, and so
forth. These latter ISAs are not mere bodies or ideas. They are modes of
address. They have a discursive component and it is through this discourse
that subjects are significantly constructed.
How does this happen? To understand this, we must start with a distinc-
tion between the individual—the human organism, a biological, numer-
ically distinct entity—and a subject—a socially constructed identity which
is less than the sum of all the properties of the numerically distinct entity.
This entity acquires an identity—becomes a subject—by being addressed
by an institution or apparatus in a certain way. Consider this heuristic
example: I get a letter from the Internal Revenue Service that says: "Dear
Mr. Carroll, Please appear at our New York office to explain your deduc-
tion of $7,500 for moviegoing on your 1978 taxes." The IRS is implicitly
addressing me as a subject of a specific sort, viz., as a taxpayer, as opposed
to, say, a dutiful husband. The discourse, in this case the letter, determines
what I am within this situation, that is, what my relevant role and response
should be. Althusser calls this process interpellation—the letter addresses
60 MARXISM AND PSYCHOANALYSIS

me in a certain way; it requires that I react as a certain kind of subject in

terms of a specific role or position (hence, the process is also often called
positioning). 12 The subject is characterized by the address, frequently
implicidv. The subject is said thereby to be constituted by (or in) the
discourse, or to be positioned by (or in) the discourse. Moreover, this
process of address is held to be an aspect of all ideology since all ideology
has a discursive component. It is always addressed to concrete individuals13
who are then transformed into concrete subjects (subjects with a certain
position).
Lest my modest tax example be deceptive, it is instructive to pause to
spell out its extremely wide implications. It is being maintained that all
ideology has a discursive component, an address; every instance of ideology
functions to position a subject, thereby reproducing a necessary condition
for the reproduction of capitalism, or for the reproduction of whatever
other relevant social formation happens to prevail. Note, too, that here the
concept of ideology is much broader than that which is usually encoun-
tered. The constitution of a subject—of identity or agency—is something
we normally ascribe to the workings of a culture, and, ordinarily, we allow
that a culture may or may not be ideological in nature.
Althusser's remarks here are generally restricted to examples of address
which occur in the context of ISAs. However, some researchers, following
Althusser, have extended it to apply to virtually every unit of discourse.14
That is, virtually all discourse is thought to address—to position and to
construct—subjects in a way that is ideologically portentous. Furthermore,
under the sway of semiotics, these researchers have a rather expansive view
of discourse. Almost every aspect of civilized life—from sentences to
clothing—has an address or a discursive component. So, virtually every
element in the culture is participating in the construction of subjects in an
ideologically significant way. Althusser's own examples of ideological sub-
ject positioning are perhaps not so inclusive. He mentions conversational
rituals, and the discourse of priests and educators. But film theorists think
that the overall model of address can apply as well to television programs,
photographs, novels, and films.15 Indeed, one commonly reads that the
subject is constantly constructed in every intelligible act of speech by means
of imaginary projections of wholeness and unity. 16 Furthermore, since
most contemporary film theorists are persuaded of the semiotic view that all
cultural artifacts have a signifying dimension, it is difficult to see how, in
principle, one could avoid saying that every cultural artifact participates in
subject positioning. I am not sure Althusser would want to say this; but I
suspect that many contemporary film theorists would not be disturbed by
this point.
MARXISM AND PSYCHOANALYSIS 61

We might think of subject positioning by saying that just as a piece of

discourse performs some illocutionary act, it also has the perlocutionarv
effect of positioning its receiver as a subject. When thinking of examples,
like my IRS letter or Althusser's of a policeman shouting at someone, it is
easy to comprehend the political significance of this perlocutionary effect
because we can name the purported subject position by means of a well-
defined social role. But when a TV announcer proclaims "It's 2:37 p.m.,"
one wonders what the accompanying subject position might be. Or, if I'm
singing a popular tune in the shower and a passerby overhears me, and
understands me, how have I positioned her? That is, where the discourse
can be defined as addressing an individual in terms of a well-defined, social
role, one can perhaps grasp the Althusserian point, and even, provisionally,
concede that it might have some ideological or political relevance. But what
of cases of discourse that do not appear to stipulate a readily identifiable
social role for their prospective receivers? These counter-examples may not
be a problem for Althusser since, except for his examples of conversations,
most of his illustrations evoke well defined social roles. However, what will
those who have extended the notion of subject positioning say here?
To answer this, we must note that followers of Althusser, especially in
the humanities, stress that the subject may be positioned not only in terms
of a nameable role. But, as well, coherent discourse, intelligible discourse—
for example, "It's 2:37 p.m."—supposedly has the force of manipulating
the individuals, who understand it, into believing that the apparent unity
and coherence of the discourse is a property of themselves as subjects.
Discourse addresses the individual as a unified subject, and the individual
mistakes the seeming intelligibility, unity, and coherence of the discourse
and its address as its own unity as an autonomous "I." That is, coherent
discourse, as such, has the capacity to instill faith in the subject in the beliefs
that she is unified and free. However, Althusserians hold that such beliefs
are illusory, i.e., misrecognitions; the individual is neither unified, free, nor
autonomous. Moreover, the myths that the individual is a unified, self-
determining, autonomous, homogeneous subject/agent is ideological in
nature. That is, the perpetuation of capitalism, supposedly, requires the
belief in unified subject/agents, and discourse, including filmic discourse, is
a vital mechanism in securing this effect.
Lacanian psychoanalysis plays an essential role in this theory by provid-
ing an account of the psychic mechanisms that underlie the individual's
misrecognition of herself as a unified subject. Althusser himself does not
explicitly work out the way in which the psychic mechanisms isolated by
Lacan operate in subject construction in "Ideology and Ideological State
Apparatuses." Rather, he invokes Lacan by using the concept of an imagi-
62 MARXISM AND PSYCHOANALYSIS

nary relationship in his account o f ideological subject production. This

opens the door for contemporary film theorists to enter into Lacan's writ-
ings in order to work out the details of subject positioning while, at the
same time, using Althusser's notion that the production o f "autonomous"
subjects is the crux o f ideology as their political justification for turning to
psychoanalysis.

LACAN AND T H E C O N S T R U C T I O N OF T H E S U B J E C T
In the concluding remarks o f Althusser's "Freud and Lacan," an assess-
ment o f the significance o f these psychoanalysts for Marxism is suggested.
What can be learned from their researches is
that the human subject is de-centered, constituted by a structure which has no
"centre" either, except in the imaginary misrecognition of the "ego," i.e. in
the ideological formations in which it "recognizes" itself.
It must be clear that this has opened up one of the ways which may perhaps
lead us to a better understanding of this structure of misrecognition, which is of
particular concern for all investigations of ideology.17

Clearly, this passage all but predicts the writing of "Ideology and Ideo-
logical State Apparatuses." However that essay, though inaugurating a
research program—by making subject construction the key to ideology—
does not offer a detailed psychoanalytic account of the dynamics of the
psychical operations and mechanisms that underpin subject construction.
The essay may allude to them by use o f terms like "imaginary," but for a
fuller picture o f what is supposedly going on in subject construction, one
must turn not only to Althusser's writing on Lacan but to Lacan himself.
Like Freud, Lacan believes that the person, the human subject, is con-
structed over the course o f one's psychosexual development. According to
Lacan, our prenatal experience is one o f plenitude—of a kind of primordial
wholeness. Birth ends this, severing us from the mother. This is felt as a
loss; human life begins in alienation and separation which is often referred
to by the term "lack" in Lacanian jargon. The human subject is said to be
constructed on this loss o f plenitude or lack during the latter stages of
psychosexual development. 18
Initially, the child feels fragmented not only by its loss o f the plenitude of
the womb but because o f its lack of motor coordination. As well, the child
is completely dependent upon the ministrations o f others. These postulated
feelings o f insufficiency and the corresponding wish for wholeness provide
a background for the child's first step in the direction o f acquiring an
identity, that is, o f acquiring subjecthood. This first step involves the
MARXISM AND PSYCHOANALYSIS 63

development of the faculty called, by Lacanians, "the Imaginary." We have

already encountered this theoretical entity in our discussion of Metz and, as
well, it also is central to Althusser's account of ideological subject con-
struction.
The Imaginary is thought to arise around the child's sixth month and it
continues to develop for roughly the next twelve months.19 This period of
the child's life is called "the mirror stage" by Lacan in honor of the crucial
role mirrors, literally and metaphorically, are thought to play in our acquisi-
tion of an identity. The Imaginary emerges in our first experience of
ourselves as individuated entities. This occurs between six and eighteen
months of age when the child's immature body is relatively uncoordinated.
However, via the agency of the child's experience of its own image in the
mirror, its feelings of fragmentation and dependency give way to a sense of
ideal unity and illusory autonomy. That is, the mirror experience triggers
the operation of the Imaginary, a faculty which like the imagination, in
older theories of mind, has a unifying function (albeit in a different domain
of mental activity for Lacan). The fragmented, dependent child, beset by
feelings of lost wholeness, recognizes (or misrecognizes) its mirror image
in a wave of jubilation. The external reflection of its own image, which
shows the body, which it experiences as a fragmentation of functions, as a
total form or Gestalt, gives the child a sense of wholeness and unity as an
entity numerically distinct from other objects and persons. This ideal unity
is also invested with a sense of autonomy. Lacan writes:
This jubilant assumption of his specular image by the child at the infans
stage, still sunk in his motor incapacity and nursling dependence, would seem
to exhibit in an exemplary situation the symbolic matrix in which the I is
precipitated in a primordial form, before it is objectified in the dialectic of
identification with the other and before language restores to it, in the univer-
sal, its function as subject.
This form would have to be called the Ideal-I, if we wished to incorporate it
into our usual register, in the sense that it will also be the source of secondary
identifications, under which term I would place the functions of libidinal
normalization. But the important point is that this form situates the agency of
the ego, before its social determination, in a fictional direction, which will
alwavs remain irreducible for the individual alone, or rather which will rejoin
the coming-into-being (le devenir) of the subject asymptotically. 20

Two factors are important here: our sense of unity and autonomy as
subjects comes from the outside and it comes from what might be thought
of as a representation (the mirror). Moreover, the faculty of die Imaginary,
initially acquired during this mirror event, stays with each of us for the rest
of our lives, constandy functioning (in tandem with other psychic mecha-
64 M A R X I S M AND P S Y C H O A N A L Y S I S

nisms to be discussed below) to support our continuing faith in our unified

subjecthood. That is, throughout life, the Imaginary operates as a psychic
mechanism that instills illusions of subject unity through representations
or, speaking more expansively, as contemporary theorists are wont,
through discourse. Thus, it is through the operation of the Imaginary that
ideological addresses transform us into positioned subjects a la Althusser.
Similarly, contemporary film theorists, often inspired by Althusser, regard
cinematic representations as the sort of thing that can trigger the Imagi-
nary, thereby contributing to the construction of ideologically positioned
subjects.
The supposed "mirror stage" event, in the child's psychosexual develop-
ment, need not be taken literally. Rather, it is offered as a label that
summarizes our constitution as subjects, shaped, as it were from the outside
(by parents, society, and the desires they inculcate through discursive
interaction with the infant).21 We distinguish ourselves as selves versus
others. But, in fact, we are constituted by others. We have biological needs,
such as that for nourishment, but that is transformed into desires for the
kind of food our culture serves. Our desires, that is, are socially mediated.
Through and through, we are molded from the outside, by what theorists
call the other. Who we are and what we want arrives from the other, despite
the fact that we take ourselves to be autonomous. The mirror stage experi-
ence, which initiates this sense of wholeness and autonomy, ironically is
based upon the other. This sets forth what might be regarded as a continu-
ing contradiction. We believe that we are unified, autonomous subjects
defined in opposition to the other, but this is based upon an extrapolation
from the intuition of outside unity: the image in the mirror, specific
culturally signifying selections of nourishment, etc.—in short, the other.
Discourse, representations, continue throughout our existence to have the
power to instill this illusory sense of unity and autonomy.
But we are not autonomous. Rather we are products o f the other. And
we are not unified subjects, but composite organisms with roiling and
conflicting drives, instincts, and desires. Yet, through the Imaginary we
construct a mythic inner identity, unified and free, which is a metaphysical
illusion masking the facts that we are both heterogeneous complexes and
the products of the other, and assuaging our sense of lost plenitude.
It is difficult to gauge the evidential support for this account. Lacan
spends almost no methodological energy laying it out in the essay in which
he postulates the mirror stage. Nor does he explain how he knows with
such precision how children feel prenatally, in their first six months, and in
their first encounter with the mirror. One is taken aback when one realizes
MARXISM AND PSYCHOANALYSIS 65

that so much contemporary theory is based upon so little documentation.

Moreover, it is admitted that the mirror event might not even occur—
though it is not said at what frequency—and that the story about the
encounter with the mirror should be understood metaphorically as a way of
indicating the influence of the other during our first eighteen months. But
isn't it straining the metaphor of the mirror to apply it to parental com-
mands and caresses? And, anyway, if the mirror experience is construed
metaphorically, doesn't that undercut the analogy between the mirror and
film upon which people like Metz rely?
Contemporary theorists seem either unaware or untroubled by such
considerations. For them the Imaginary exists and the mirror stage oc-
curred. Moreover, the Imaginary continues to operate through later de-
velopmental stages, and it functions to transform apparent (misrecognized)
external unities into confirmations of the illusion of subject unity. But since
it is believed that these external unities themselves are only apparent, the
Imaginary is also the mechanism through which illusions of external unity
are apprehended (or rather misapprehended) as unities. The apparent
coherence of a movie ("apparent," perhaps because it is made up of frag-
ments through editing, or perhaps because it is full of "contradictions"), for
example, produces the illusion of subject unity in a spectator, which illu-
sion, in turn, supports the persuasiveness of the illusion that the movie is
unified. That is, the two illusions of unity are interdependent and mutually
reinforcing.
The Imaginary is the mechanism that responds to apparent external unity
by producing illusions of subject unity. So, when (apparently) intelligible
discourse, a movie say, addresses an individual, the Imaginary swings into
action, positioning the spectator as a unified subject. And, for a moment,
returning to the political realm, recall that such a discursive effect is said to
have ideological significance because, putatively, capitalism requires that
individuals believe that they are unified autonomous subjects.
At this stage in the exposition, it is useful to note the extraordinary
variety o f apparent unities that, according to contemporary theorists, the
Imaginary can use to project subject unity. These include, among other
things, human bodies and images thereof, parental imperatives and any
meaningful sentence as well as coherent films and novels.22 One wonders
how these very different stimuli can function in the same way as inputs in
the causal system of subject construction (also called subject production or
subject positioning). On the face of it, it would seem to be the vagueness
and lack o f specificity o f the operative concept of unity in this account that is
allowing its proponent to treat things as dissimilar as bodies and sentences
66 MARXISM A N D PSYCHOANALYSIS

as if they were functionally equivalent. But the only connection between

them seems to be a verbal one.
We call bodies unified in virtue of their continuity within the limits of
their contour; coherent utterances are unified in virtue of their meaningful-
ness. Despite the fact that we may use the same word to describe these
things, we are nevertheless discussing different kinds of things. Once we
undertake to describe them less amorphously, we must ask how such
disparate causes manage to evoke the same effect; how does the Imaginary
move from projecting subject unity on the basis of perceiving bodies or
images thereof to doing so on the basis of hearing a coherent sentence? In
short, we may need a further mechanism to explain the mechanism of the
Imaginary, unless the latter has merely been fabricated through the equiv-
ocation of the word unity.
There is no reason to assume that everything we are willing to call
"unified" shares the same causal potential for the Imaginary. Moreover, if
we say that the Imaginary is able to use all these different unities to project
subject unities because these different things all have the common property
of self-identity, then we have traveled in a circle, since it is the Imaginary
that is supposed to explain how things are grasped as self-identifiable
entities. Thus, it would be somewhat embarrassing to invoke the principle
of identity to explain how the Imaginary works. And, of course, since
contemporary theorists presume that the external stimuli that engage the
Imaginary are not truly homogeneous (in a way that putatively defies the
principle of identity) in any case, something like the principle of identity
would be of no use whatsoever.
A related problem that besets the use of terms like "unity" and "homoge-
neity" in contemporary film theory is that it is not clear that these terms are
always being used intelligibly. I know the criteria for applying the concept
of unity to a landscape painting, a TV news program, a novel, and a piece of
point-of-view editing in a film. But there does not seem to be a commonly
shared feature between the unity of a news program, point-of-view editing,
a landscape painting, and a novel. To use the concept of unity in a meaning-
ful way, we have to use it with reference to specific kinds of cases. However,
often in contemporary theory, one has the sense that words like "unity" and
"homogeneity" are being used as if there were a decontextualized sense to
them—that is, as if the unity of a picture and a sentence or a piece of point-
of-view editing and a body were all the same. This is pertinent to the
discussion of the Imaginary, since in contemporary theory anything that
can be denominated a "unity" is grist for the Imaginary^ processing. And,
as we shall see, similar questions arise with an opposing set of terms such as
"decentered" and "heterogeneous."
MARXISM AND PSYCHOANALYSIS 67

In Althusser's "Ideology and Ideological State Apparatuses," it seems as

though the operation of subject construction, which Althusser calls "inter-
pellation," rests on the Imaginary. However, Lacan also identifies another
faculty—the Symbolic—as critical for subject construction. As a result,
many contemporary film theorists have sought to expand upon Althusser's
account by including the operation of the Symbolic in their stories of
subject construction. In this, they remain essentially within the Althusser-
ian research program since they are motivated by what they regard as
Althusser's central discovery, viz., that ideology is fundamentally a matter
of subject construction. Where they seem to part company with "Ideology
and Ideological State Apparatuses" is in its apparent reduction of subject
positioning to the operation of the Imaginary which might also be stated as
denying that subject construction is solely a matter of interpellation.23
The Symbolic stage of psychosexual development corresponds roughly
to what Freudians call the oedipal stage. For Freudians, this is a par-
ticularly crucial step in the construction of the individual because it is at this
time that the organism is culturally defined in terms of its sexual identity.
For example, it is the period in which the male child, putatively fearing
castration by the father, leaves the quest for the mother and seeks to
emulate the father in a process called introjection. That is, the boy child
introjects the father, which means that he attempts to take on the values,
rules, and behaviors of the father. He strives to become manly in the way
that the father is manly. The child, at this point, is "sexed" as male; he
assumes a male identity. But this male identity is not simply a biological
matter; it is also, and more importantly, a cultural matter. For the child is
assuming the behaviors, roles, values, and beliefs of what it is to be a man in
the culture his father was raised in. For Freudians, the introjection process
of the oedipal stage is the means by which the culture reproduces itself,
assuring, from generation to generation, the cultural reproduction of the
kind of, for example, male identity which is required in order for society to
function—warriors in one case, perhaps workaholics in another.24 The
values, roles, attitudes, and ideals introjected at this point extend to every
aspect of the individuals' social life. And if Freudians are right in what they
assert of introjection at the oedipal stage, then this would obviously be the
nodal point in the production of subjects.
Lacanians agree with the main points of the Freudian scenario of the
oedipal stage, but they have a more elaborate version of it. They would not
say that it is a different account, for they believe that Lacan was involved in
discovering the real Freud.25 Be that as it may, for Lacanians, the investi-
ture of the child with its social roles, ideals, and values at the oedipal stage
is also the point at which the child is said to enter language. The oedipal
68 M A R X I S M AND P S Y C H O A N A L Y S I S

stage is rechristcncd the Symbolic stage, the psvchosexual stage at which

the faculty of the Symbolic is acquired, which is marked by the child's
passage into language. Lacan writes:

This is precisely where the Oedipus complex—in so far as we continue to

recognize it as covering the whole field of our experience with its significa-
tion—may be said, in this connexion, to mark the limits that our discipline
assigns to subjectivity; namely, what the subject can know of his unconscious
participating in the movement of the complex structures of marriage ties, by
verifying the symbolic effects in his individual existence of the tangential
movement towards incest that has manifested itself ever since the coming of a
universal community.
The primordial Law is therefore that which in regulating marriage ties
superimposes the kingdom of culture on that of a nature abandoned to the
law of mating. The prohibition of incest is merely its subjective pivot,
revealed by the modern tendency, to reduce to the mother and the sister the
objects forbidden to the subject's choice, although full licence outside of this
is not yet entirely open.
This law, then, is revealed enough as identical with an order of language.
For without kinship nominations, no power is capable of instituting the
order of preferences and taboos that bind and weave the yarn of lineage
through succeeding generations. And it is indeed the confusion of genera-
tions, which in the Bible, as in all traditional laws, is accused as being the
abomination of the Word (verbe) and the desolation of the sinner.26

Here we see the Oedipus complex linked with the anthropological phe-
nomena of the incest taboo, a correlation which Freud himself had made.
Moreover, like Freudians, the Lacanians regard the prohibition against
incest as crucial to social organization—"covering the whole field of our
experience with its signification." The taboo, as well, is linked with lan-
guage. Why? Because incest is determined by things such as clan, moiety,
and family. With whom one can and cannot legitimately mate is a matter of
one's position in these networks which is marked by how one is named. The
law of incest, which Lacan appears to regard as the fundamental social law,
is underpinned by a system of names in which the name of the father, so to
speak, serves as the anchor. Who one mates with is calculated in terms of
the relation of one's name with the name of the father in question. This
leads Lacanians to see social laws, "the Law," in their jargon, as a system
with the "name of the father," used as a technical term, as its fulcrum. "It is
in the name ofthe father that we must recognize the support of the symbolic
function which, from the dawn o f history, has identified his person with the
figure of the law." 27 (The anchorage provided by the name ofthe father, as a
MARXISM AND PSYCHOANALYSIS 69

technical term, can also be referred to as the phallus, also used as a technical
term, not referring to the literal penis but to the centrality of the patriarch
in the Law).
This suggests that social laws, the Law, if you will, is connected with
language, or at least with names. But the passage claims more: "This law,
then, is revealed clearly enough as identical with the order of language."
The child maturing into the jurisdiction of the Law, in turn, is said to enter
the order of language, the Symbolic Stage. But why, one still wonders, are
the order of the Law and the order of language identical?
Lacan seems to have reached this conclusion by the triple superimposi-
tion of a theory of language (Saussure's) onto an anthropological theory
(Levi-Strauss's) onto his own variant of psychoanalysis. This superimposi-
tion is facilitated by the fact that both Lacan's theory and Lévi-Strauss's find
their inspiration in Saussurean linguistics. Lévi-Strauss thought that the
laws of thought that underlie the type of phenomena anthropologists
study, including incest, have the same structure that the semiotic phe-
nomena that Saussure theorized have, viz., one in which the meaning of the
signs in a language is diacritical or differential, i.e., the meaning of the
terms is not defined in isolation but in relation to other terms in virtue of
their differences.
In entering the Law, the child finds its place in a system of names
diacritically structured around the name of the (his) father; it finds its place,
that is, in the most crucial pattern of definition in its culture, one from
which other cultural distinctions flow and which reflects the diacritical
structure of language. The Law is both the anchor of the system of cultural
distinctions embodied in language (though Lacanians might not like my
phrasing here) and it is a particularly powerful example of the diacritical
structure of semiotic signification. Obviously, the child has been a language
user before the oedipal stage. But it is at that time that it enters language in
the sense of entering what is thought to be the key relation of the system,
thereby becoming a fully incorporated participant in the language. By
positioning itself in relation to the name of the father, the child reaches a
new level of subject construction through becoming a full-fledged lan-
guage/culture user.
There is a hornet's nest of assumptions here, many of which involve
intricate questions in the philosophy of language too detailed to examine at
length here.28 One problem is whether the theory of language Saussure is
thought to offer is an adequate theory at all. That is, a constraint on an
adequate theory of language is that it be able to explain how language
learning and communication are possible. But Lacan appears to interpret
70 M A R X I S M AND P S Y C H O A N A L Y S I S

Saussurc's diacritical conception of meaning to entail that meaning is

disconnected from reference to things and hinges solely upon the interrela-
tions of difference between the signs in the language system. However, this
would make it difficult to understand how language ever got off the ground
as well as how it is taught to children. Each sign would just send us to
another sign regressively. That is, for language to begin and for it to be
taught requires some fixed relations between some words and their refer-
ents. 29 Thus, the Saussurean theory under Lacan's dispensation fails to
explain such basic facts about language as language learning.
One could avoid the aforesaid regress, of course, by maintaining that
some words are not defined in terms of other words but are taught osten-
sively. But Lacan does not. Rather, he takes his putative disconnection of
words from the world as yet another mark of humanity's alienated nature.
The Symbolic stage of development is marked by division and difference
(because it is articulated by the diacritical structure of language). Human
life is completely mediated by language, language anchored in the Law of
the name of the father. Lacan writes:

Symbols in fact envelop the life of man in a network so total that they join
together, before he comes into the world, those who are going to engender
him "by flesh and blood"; so total that they bring to his birth, along with the
gift of the stars, if not with the gifts of the fairies, the shape of his destiny; so
total that they give the words that will make him faithful or renegade, the law
of the acts that will follow him right to the very place where he is not yet and
even beyond his death; and so total that through them he is not yet and even
beyond his death; and so total that through them his end where the Word
absolves his being or condemns it. 3 0

At the Symbolic stage, the child becomes crucially integrated in what are
thought of as the codes of human role assignment and communication; its
very desires are constituted within the context of the Law. 31 The mecha-
nism of the Symbolic, then, is associated with language, and language
processing, which is to say that it operates with respect to difference, since
the structure of language narrowly and cultural coding widely is diacritical.
The Symbolic does not supersede the Imaginary, the faculty for the
projection unity and wholeness, even though the Symbolic is associated
with difference, division, and heterogeneity. Indeed, the Imaginary abets
the positioning of the Symbolic stage. For the child must perform an
imaginary identification with its position in the Law/language in relation to
the name of the father.32 For example, in order to assimilate the sentence "I
cannot tell a lie," it is thought that some imaginary identification with the
MARXISM AND PSYCHOANALYSIS 71

subject of that utterance must be effected. That, in turn, requires the

operation of the Symbolic in order to position that " I " within a system of
the differences o f the Law/language. So, it is said that the Imaginary and
the Symbolic, unity and difference, work together in order to culturally
position the subject as a unified, autonomous agent in relation to the
Law/language. Moreover, this is not an interaction that occurs only at the
Symbolic stage. Given the countervailing forces o f unity and difference
which underlie the subject position, it must be renewed again and again
throughout life as unity gives way to difference which calls forth the
operation of the Imaginary once again. Contemporary film theorist Colin
McCabe writes:
Language in the realm of the imaginary is understood in terms of some full
relation between word and thing: a mysterious unity of sign and referent. In
the symbolic, language is understood in terms of lack and absence—the sign
finds its definition diacritically through the absent syntagmatic and paradig-
matic chains it enters into. As speaking subjects we constantly oscillate
between the symbolic and the imaginary—constantly granting ourselves
some full meaning to the words we speak, and constandy being surprised to
find them determined by relations outside our control. 33

The subject's position in culture—which in the Althusserian-Lacanian

paradigm is equivalent to its ideological position—is constituted by the
work o f the Imaginary and the Symbolic which appear to have to renew the
subject position in every act of discourse (where discourse is broadly
construed). Furthermore, this process is strongly deterministic.
. . . the "theatrical machine" imposed by the Law of Culture on every
involuntary, conscripted candidate to humanity, a structure containing in
itself not only the possibility of, but the necessity for, the concrete variants in
which it exists, for every individual who reaches its threshold, lives through
it. 34

Moreover, this determinism is rooted in the operation o f language

(broadly characterized). Contemporary theorist Kaja Silverman holds that
the speaking subject is not really in control o f her own subjectivity. Why?
To begin with the subject's discourse is constrained by the rules of language;
it can only speak by means of a pre-existing linguistic system. Moreover,
"language" must here be understood in the broadest possible sense, as en-
compassing not only the operations of denotation, but those of connotation.
It is spoken not only bv the palpable voice of a concrete speaker, writer, or
cluster of mechanical apparatuses, but the anonymous voices of cultural codes
which invade it in the form of connotation. 35
72 MARXISM A N D PSYCHOANALYSIS

From Althusser, the contemporary film theorists derive the idea that the
primary operation of ideology is subject construction. From Lacan and
other psychosemioticians, and with the endorsement of Althusser, contem-
porary film theorists get the idea that subject construction occurs through
the interaction of the Imaginary and the Symbolic with the language of the
other. Subject construction, moreover, is a strongly determinist process.
Language, of course, in this context is to be thought of in very expansive
terms. Film, for example, can serve as an instance of the language of the
other. From these assumptions, a straightforward approach to the analysis
of the ideological operation of film can be drawn.
Since the construction of subjects possessed by the faith that they are free
and unified is the foundational operation of ideology, contemporary film
theorists will take as their central task the explanation of the ways in which
film does this. Since Lacanian theory purportedly is best suited for analyz-
ing subject construction, the film theorist will concentrate on the ways in
which film engages the psychic mechanisms that stimulate subject produc-
tion. In part, this involves triggering the psychic mechanism called the
Imaginary which again and again rehearses its mirror stage performance by
projecting the sense of subject unity on the basis of the apparent unities
issuing from the other. The task of film researchers, then, becomes the
isolation of the features of films that impart impressions of apparent unity.
For these will be ideological levers that trigger the psyche to endorse the
illusion of subject unity. Some of these features of film, as ensuing chapters
will elaborate, include the perspectival image, narrative structure, syn-
chronized sound, point-of-view editing, and a panoply of other cinematic
devices. The film researcher will also have to examine the way in which
cinema engages the Symbolic in its process of subject construction, and this
will involve, most especially, showing how, despite the intimations of
difference and heterogeneity that come with engaging the Symbolic, film,
particularly of the sort called movies, contains the impression of hetero-
geneity in favor of the illusion of wholeness and homogeneity which
promotes confidence in the supposed sine qua turn of ideology: the unified
autonomous subject.
Operating within the Althusserian-Lacanian framework, the film re-
searcher has her path clearly marked off. To study ideology is first and
foremost to study subject construction. Subject construction is crucially
brought off by the projection of apparent unity by the other. So one should
isolate the unity projecting features of film and explain the ways in which
their apparent unity is counterfeited, including explaining the ways in
which the recognition of heterogeneity by the Symbolic is contained. This
will account for the way in which subject unity is induced. As this might
MARXISM A N D PSYCHOANALYSIS 73

suggest, such a program will involve a great deal of abstract psychoanalytic

description of the interaction of postulated intrapsychic forces rather than
what we ordinarily identify as political criticism. That is, a great deal of time
will be lavished upon the formation of abstract subject constructions—
subjects necessary to bear social roles (subject supports), rather than on the
ways in which films abet specific, ideologically significant beliefs, such as
"The unemployed are just lazy." To the worry that the Althusserian-Laca-
nian paradigm may defer us from concrete social criticism in favor of
abstract psychoanalysis, the contemporary film theorist is likely to reply
that the psychoanalysis of subject construction is a central form of political
analysis. But whether such an assertion survives scrutiny is a topic with
which much of the rest of this chapter is much concerned.

AN I N I T I A L Q U E S T I O N I N G O F T H E P R E S U P P O S I T I O N S
OF THE ALTHUSSERIAN-LACANIAN PARADIGM
In the chapters that follow, I shall be primarily examining attempts by
contemporary film theorists to extend the Althusserian-Lacanian paradigm
to cinema along with the problems this project raises. However, before
exploring in detail the fortunes of the paradigm as applied to film, it is
worthwhile to spend some time interrogating some of the basic presup-
positions of the paradigm. In a book devoted to film theory, it is impossible
to expend the amount of space necessary to show that the foundations of
the paradigm are irretrievably ill-conceived. However, in what remains of
this chapter, it should be possible to show that many of the central assump-
tions of the reigning paradigm are questionable enough and their central
concepts so poorly defined that the burden of their proof can be shifted to
their proponents. In what follows in this section and the next, I will
primarily, though not exclusively, introduce problems that erupt in the
Althusserian wing of the paradigm, leaving many of the inadequacies of the
Lacanian wing to be taken up in subsequent chapters.
Perhaps the most glaring problem with the research program sketched so
far is its presiding concept of ideology. It is too broad. By identifying
ideology with subject construction, the concept has become roughly coex-
tensive with that of a culture, thereby losing its pejorative force. Ordinarily
we do not want our ideas and our thinking corrupted by ideology. For if a
belief is ideological, then that implies it is (1) false and (2) that it is a
rhetorical tenet that functions to uphold some practice of social domina-
tion. For example, "All black men desire to rape white women" expresses
such an ideological belief.36
Not all beliefs enunciated or presupposed by a culture are ideological in
74 MARXISM AND PSYCHOANALYSIS

this way. Moreover, it seems self-defeating for a Marxist to employ such an

overblown concept of ideology, since Marxism proposes to found a culture
which under this approach will be ideological. Every community will have
to have a culture. Althusser seems willing to accept this consequence.
However, the negative force of ideology will be lost if every conceivable
culture must be ideological. There will be no point in extricating ourselves
from ideology if we do so only at the price of entering a new ideology,
especially where that new ideology engages in the same, putatively suspect,
process of subject construction. Whereas when the idea of ideology is
connected with practices of demonstrably avoidable domination and false-
hoods, one clearly feels the negative pressure of the charge of "ideology"
and is thereby moved to dissociate oneself from what is ideological. The
hortatory force of ideology is lost when ideology becomes culture, since the
critical sense of the concept requires us to disavow ideology, while it is
probably literally impossible (and politically unnecessary) to disavow all
culture.
The preceding observations are related to the question of the usefulness
of the idea of the ideological construction of subject unity. It is held by
contemporary theorists that every bit of discourse enforces a subject posi-
tion. In the hands of contemporary film theorists, this focuses analysis on
what might be thought of as relatively small sets of phenomena—camera
movements, cutting patterns, the narrative structures of whole films, and so
on. The contemporary film theorist provides moment-by-moment recita-
tions of how, with each device, the projection of coherence results in the
construction of subject unities, which is said to involve interminable oscilla-
tions. However, the analysis of subject unities at the level of the way each
piece of discourse addresses the individual will prove very awkward for
examining the kind of ideological subjects that are relevant to social anal-
ysis.
Let us make a rough, pragmatic distinction between two kinds of sub-
jects: the occurrent subjects, i.e., unified subjects constructed on a moment-
to-moment basis by the way in which discourse engages the Imaginary and
the Symbolic; and stable subjects, i.e., subjects with, among other things,
personality profiles that remain relatively stable over long periods of time.
An obedient worker, a misogynist, a bigot, and a dedicated bureaucrat are
examples of what I call stable subjects. But it is hard to see how we are to get
from the construction of momentary subject unities to the stable sub-
ject/agents of the type just listed. It is an act of faith rather than of science to
believe that these momentary constructions amount to particularized, sta-
ble, ideologically instilled subject/agents. What are the patterned variables
MARXISM A N D PSYCHOANALYSIS 75

and regularities that transform momentary subject unities into stable ideo-
logical subjects?
The contemporary film theorist describes batteries of subject unities, but
it is impossible to see how these combine into ideologically mystified agents
with the inclination to uphold practices of social domination. If it is true
that discourse produces subject unities, these are extremely formal in na-
ture, almost like Kantian transcendental egos. But it hard to see the way in
which, logically, one could derive anything informative to say about the
formation of stable ideological agents with fixed dispositions on the basis of
such formal unities.
One way to appreciate the force of the preceding charge is to note that
though I may be addressed ideologically by some discourse, and though I
may take that discourse to be a coherent utterance—for example, "Unions
are by their very nature anti-democratic"—I may reject the utterance. Or,
to take an example from film, I may resist the heroic portrait of the KKK in
The Birth ofa Nation. These examples show that I am not the simple reflex
effect or imprint of ever)' coherent signifier. But how do these resistances
happen? And what are the differentia between being positioned as a be-
liever-subject-unity versus a disbeliever-subject-unity. An answer to this
might begin by establishing the relation between what I have called occur-
rent subject constructions and stable ideological subjects. But contempo-
rary film theorists, who use this model, have nothing to offer here. At the
very least, this suggests that the construction of subjects through discourse
simpliciter is not a very useful concept for supplying us with an analysis of
the formation of ideological agents. The animus of my objection here is not
essentially political. I am not saying the problem is that the concept of
subject unity is not helpful for revolutionary praxis (though probably it is
not). Rather, I am making the logical point that the notion of the con-
struction of subject unities through discourse falls short of what it ostensi-
bly promises to explain—the creation of ideological agents.
A major assumption of Althusserian-Lacanian film theorists seems to be
that the construction of subject unity, in and of itself, is politically negative.
On the face of it, this seems an unlikely premise. For if the problem of the
construction of subject unity is that it supplies a necessary condition for
further ideological operations—for example, the creation of ideological
agents—then it must be admitted that the construction of unified subjects
also appears to afford a necessary condition for individuals to adopt eman-
cipatory roles. Am I not positioned by "Workers of the world unite!" and
wouldn't the subject unity imparted by the coherence ofthat text become a
condition for revolutionary agency? If this is so, then how can subject unity
76 M A R X I S M AND PSYCHOANALYSIS

constitute a special problem apart from the problem of the specific kind of
role-agencv with which it is, in fact, correlated? That is, the theory's
negative attitude toward the construction of subject unities is unmotivated
unless a persuasive account is offered of the way revolutionary agents are
constituted without presupposing the construction of subject unities. It is
of no theoretical value to say that true revolutionaries are or will be
constituted differendy from those under capitalism: an adequate theory for
analyzing ideology in the capitalist epoch must portray the why and where-
fore of this.
Indeed, for Marxists, the whole issue of emancipation renders the Al-
thusserian framework problematic.37 Marxism, or at least much Marxism,
operates under the assumption that available within everyday life are op-
portunities for emancipatory changes in consciousness. But given the Al-
thusserian conception of ideology it is impossible to understand how this
could happen. Another way to make this point is to note that the Al-
thusserian account is functionalist; as such it offers a picture of the Ideolog-
ical State Apparatuses functioning well. But it seems that the account
presents the ISAs as functioning too well. The model appears to preclude
their malfunctioning, though surely the ISAs do malfunction, and knowl-
edge of that malfunctioning, which the Althusserian system denies its
followers, would, of course, be quintessentially useful knowledge for Marx-
ist praxis, not to mention good social science. Upon what basis can an
emancipatory praxis occur in the world of the ISAs? Note that trade unions
and political parties are listed as Ideological State Apparatuses. Earlier I
raised the question of whether this classification was apt? Perhaps now the
committed radical can see the point of the question. Also, the family,
ostensibly a private institution, is an ISA. But is this true of politically aware
working-class families, and the families of other oppositional groups?
We opened this chapter by noting that an abiding issue for Marxist
aesthetics is the identification of the emancipatory potential of art. But it
seems to me that art cannot be granted an emancipatory role that is
intelligible within the rest of Althusser's system. Of course, in his writings,
Althusser does try to assert that art is emancipatory. He claims, for example,
that the Brechtian alienation effect contests the myth of psychological
unity. He says that "in the theatrical world, as in the aesthetic world more
generally, ideology is always in essence a site of a competition and a struggle
in which the sound and fury of humanity's political and social struggles is
faindy or sharply echoed." 38 But where, given Althusser's picture of ideol-
ogy, does the spectator find the wherewithal to recognize, let alone join,
this competition? On occasion, Althusser waffles almost incoherently,
saying that real art is not to be ranked as an ideological agency, though
MARXISM AND PSYCHOANALYSIS 77

neither is it scientific, which, for him, means productive o f knowledge o f

reality.39 However, no clear explication o f this can be located intelligibly
within Althusser's essentially dualist epistemology wherein science and
ideology are the basic contrasting terms.
Consider this specimen o f Althusser's practical art criticism. Speaking o f
the painter Leonardo Cremonini, Althusser claims:

If Cremonini's faces are deformed, it is because they do not have the form of
individuality, i.e., of subjectivity, in which "men" immediately recognize that
man is the subject, the centre, the author, the "creator" of his objects and his
world. Cremonini's human faces are such that they cannot be seen, i.e.,
identified as bearers of the ideological function of the expression of subjects.
That is why they are so "badly" represented, hardly outlined, as if instead of
being the authors of their gestures, they were merely their trace. They are
haunted by an absence, that of the humanist function which is refused them,
and of which they refuse; and a positive, determinate absence, that of the
structure of the world which determines them, which makes them the anony-
mous beings they are, the structural effects of the real relations which govern
them. If these faces are "inexpressive," since they have not been individu-
alized in the ideological form of identifiable subjects, it is because they are not
an expression of their "souls" but the expression, if you like (but the term is
inadequate, it would be better to say the the structural effect) of an absence,
visible in them, the absence of the structural relations which govern their
world, their gestures and even their experience of freedom.40

But if the implications o f these paintings are to be found in what is not in

them, i.e., their absences, and if that, in turn, signals an entire anti-human-
ist theory, one must ask who the spectator is who can derive this from the
art works? It would have to be, at least, someone in command o f the anti-
humanist theory, as well as someone who could correlate the aforesaid
absences with that theory. And the likeliest candidate here would be an
Althusserian scientist. But art like this is hardly emancipatory in the sense
Marxist aesthetics has traditionally thought art could be; for it requires a
spectator who already knows the truth. 41
Previously, we noted that within the Althusserian-Lacanian paradigm
the construction o f subject unities is regarded negatively. Now we must
return to that issue. Althusser argued that the production of a unified
subject through discourse was simultaneously the production o f a subject,
one to be ruled. Here we have what might be thought o f as determinism
with respect to discourse; call it discourse determinism, wherein the
subject is an effect o f the discourse (of the signifier). Such a position is
fraught with the difficulties o f determinism plus some o f its own making.
Those of its own making derive from the primacy o f place that it accords
78 M A R X I S M AND P S Y C H O A N A L Y S I S

discourse: the postulated causal nexus between subject position followed

by subjection is implausible. Hearing does not necessitate believing, let
alone complying. One often rejects what one hears even if it is intelligible.
That is, even if a piece of discourse is intelligible enough to be said to
engender a unified subject position, that by no means indicates that it turns
its hearers into willing subjects. Unless we get an explanation o f quite
ordinary cases like these that fits with the rest of discourse determinism,
then we have no reason to believe that we are always subjected, in politically
significant respects, by the mere intelligibility of discourse. It may be urged
that whenever I reject one subject address it is because I have already been
positioned by an earlier, contradictory subject address. But why do I
choose one among rival subject addresses? Can an account be given that
decisively precludes all free action on the part of the individual?
Furthermore, if it is argued that the individual is unfree because he is
initially an empty organism completely shaped by some discourse, then it
would seem that the implicit definition of freedom here is too extreme. It
amounts to the claim that the only free individual would be sui generis. But
the relevant ideal of freedom is based on the belief that agents can do
otherwise and not that they ultimately create themselves.
Within the Althusserian-Lacanian paradigm, the individual is said to be
invested with the belief that she is autonomous, but this is a false belief in
the service of ideology. But why does the contemporary theorist deny
autonomy to the individual? Earlier we read Silverman offering as a reason
that "the subject's discourse is constrained by the rules of language; it can
only speak by means of a pre-existing linguistic system." However, the
assumption, in this argument, of what freedom would have to be, were
there such a thing, is too extravagant. For this argument appears to presup-
pose that no speaking subject is free unless it creates the language it speaks.
But this is absurd. If I have a hammer and I can use it to build a house, or a
hobby horse, or simply use it to pound the ground, then it seems to me that
I am free in what I hammer. And if I hammered someone who annoved
me—while certifiablv sane—I would be responsible for my act since it was
free. But Silverman's argument, by logical analogy, would have it that I am
not free because I did not invent hammers. This idea of freedom, however,
is unacceptably exorbitant, and any argument that uses it as a standard of
what freedom is is unsound.
As Silverman's argument exemplifies, there is a presumption among
Althusserian-Lacanians that if human actions have certain structural condi-
tions, these constrain human action in a way inimicable to autonomy.
Languages have both syntactical rules and semantical rules. But it is strange
M A R X I S M AND PSYCHOANALYSIS 79

to think of these as constraints that preclude autonomy. For these verv

features of language are what enable the speaker to speak—to, for example,
denounce capitalism. If the language lacked these structural conditions,
nothing could be said, which would in fact be a real blow to the possibility
of human autonomy.
Languages like English do not determine what can be said in them in any
politically significant way; nor do they challenge the individual's autonomy
in virtue of their rules. They, in fact, empower the individual. One way to
see this is to imagine what it would be like were there no semantical or
syntactical rules. Clearly this would not be a situation in which human
autonomy was increased, but one in which the capability of speaking would
be impossible, which, in turn, would represent a set back to that which we
ordinarily conceive to be human freedom. The problem with Silverman's
argument, in other words, is that it presumes as a limitation on human
autonomy that which is in fact something that facilitates autonomy. And
this can be seen readily by realizing that if we remove what Silverman
regards as constraints inimicable to autonomy, humans would lose the
capacity to speak, and, correspondingly, a whole realm of freedom.
Similar points may be made with reference to the way in which structural
conditions function in society in domains other than language. Society
presents the individual with a range of rules, roles, and ideals. Within these
conceptual networks, the individual pursues her goals. Without such con-
ceptual networks it would be impossible to formulate goals. Again, the
existence of a preexisting structure of roles, rules, and ideals is that which
makes autonomous action possible rather than that which blocks freedom.
The roles society affords, though well-specified, are not predetermined.
It is not the case, contra the earlier quotation from Althusser, that a social
role has antecedendy built into it every variation of behavior that will
eventuate in the pursuit of that role. The role may be connected to specific
rules, ideals, imperatives, and norms, but these must be interpreted by the
individual who must determine how and whether these factors apply to a
given case. This is not to say that role players will not often act routinely. It
is rather to say that even within a well-specified role there is ample room for
free action.
Another way to make this point is to observe that the bearers of roles
have ideas about the very roles and rules they execute; and these can change
the roles and rules in a way that is not predetermined by the roles and rules.
It would, of course, be absurd to suggest that humans are not shaped by
social rules, roles, ideals, imperatives, guidelines, precedents, and so on.
However, it seems equally problematic to think that society totally molds
80 MARXISM AND PSYCHOANALYSIS

the agent, since social roles, rules, ideals, and so on are transformed by the
interpretive activity of agents.42
According to contemporary theory, one reason that the individual is
possessed of the illusion of autonomy is because he is possessed of the
illusion of unity. The underlying reasoning here seems to be that unity is a
precondition for autonomy. And this seems plausible when we are consid-
ering the Lacanian developmental scenario. For the disunity of the child at
that point refers to a lack of motor coordination; and motor coordination
would seem to be connected to certain basic forms of autonomy. So, the
question arises as to why contemporary theorists think that the adult
subject's putative belief that he is in some sense unified is an illusion?
The problem for contemporary theorists seems to be that though I
believe I am unified, I am in fact a construction. But is this really a problem?
Unity and construction are not mutually incompatible terms. The Lincoln
Tunnel is unified, and it was and is a construction. Thus, the fact that
someone believes x is unified does not preclude the belief that x is a
construction.
Also, the fact that a subject believes he is a unified subject does not entail
that he holds the supposedly false belief that he is not a construction.
Furthermore, I may truly believe that I am a unity and yet be unaware of the
process by which I was constituted. The belief in subject unity, in other
words, does not, on its own, suggest any false or illusory beliefs about
whether or not the individual believes himself to be a construction.
Of course, an agent may believe both that he is a unity and that he is not a
construction. But it does not seem to be commonly believed, in either
modern capitalist or socialist society, that individuals are not constructed.
The operation of nurture and environmental conditioning on the individ-
ual, the popularity of practices like psychoanalysis and behavioral modifica-
tion (at least in America) as well as the admission that people change over
time are commonplaces rather than secrets in contemporary industrial
society. That is, if it is a common belief that people are unities under
capitalism, that belief is conjoined with the logically compatible belief that
people are, in various respects, constructions.
Since the conjunction of a belief in subject unity and subject construction
is not necessarily contradictory, the source of the falsity or illusion in the
belief in subject unity must lie elsewhere. The idea of subject construction is
putatively the main discovery of contemporary theory. So the belief in
subject unity must be the relevant false belief. Contemporary theory, then,
must offer some compelling argument to demonstrate the falsity of the
individual's confidence in being a unified subject.
MARXISM A N D PSYCHOANALYSIS 81

The attack on the belief in personal identity is long-standing in philoso-

phy; it has been advanced by Hume and Nietzsche.43 In the context of the
Althusserian-Lacanian paradigm, the unity of the subject is denied for two
reasons: first, because the person is always the production of the other, and
second, because of something to do with the unconscious. Let us turn to
the first proposal.
It is believed that there is no unified subject because the person is always
the production of the other. That is, the problem seems to be that the
boundary between the subject and the other is blurred in such a way that
the subject cannot be thought of as a unity. We are constituted from the
outside. We are not unified since not all the causes of our decisions are
internal to us. Effectively this implies that the subject (person), or anything
worth (metaphysically) calling a unified subject would have to be sui
generis. But this involves defining the subject out of existence; of course, all
we have left is the other. Nevertheless, one wants to know whether we have
got hold of a problem with the subject here or a problem with the defini-
tion.
Concepts like person, autonomy, and subject play a range of descriptive
and explanatory roles in everyday life, and in scientific and moral theories.
They are applied to a cluster of phenomena to which they refer. But being
sui generis does not appear to be a necessary condition for their application.
I may make an autonomous decision to be a lawyer, but this does not
require that I have invented the legal institution any more than my decision
not to steal requires that I wrote the Ten Commandments. I inherited the
possibility of making these decisions from the other, but that does not
entail that I do not behave freely and autonomously. Nor does this inheri-
tance show that I am not an integrated subject or a person, since to reach
this stage, given the conditions of human life, requires such inheritances.
The error of contemporary theory seems to lie in proceeding from the
correct recognition that the subject cannot be entirely disjunct from the
other, to a move in which the subject is totally swallowed by the other. This
translates into an impossibly demanding definition of the free, unified
subject rather than one that defines free, integrated subjects in virtue of the
options presented them; as Engels suggests, freedom emerges within the
bounds of necessity.44
The second line of argument against the existence of subject unity seems
to rely on the existence of the unconscious. At any point in time, the
individual is a mass of forces, many of which elude the purview of con-
sciousness. Some aspects of the individual, like the deep generative rules of
her grammar, are merely unknown to the individual, while others are
82 MARXISM AND PSYCHOANALYSIS

repressed; these latter are the contents of the unconscious. Subject unity, it
might be said, is a myth constructed by ignoring many of the properties of
the individual, including not only those that are unknown, but more
importantly, those that are unconscious. The individual always exceeds the
subject unity. The conflicting drives and forces of the unconscious are
sublated from our concept of ourselves as subjects. The subject's unity is a
construct. The subject's unity is an illusion.
Here the existence of the unconscious is being used to make a metaphys-
ical point. But exactly what point is being made or refuted? Is the fact of the
unconscious supposed to refute my belief that I have a personal identity
over time; that I am the same person today that I was two weeks ago? This
belief does not rest on an assumption that I have all the same properties I
had two weeks ago, nor does it assume that I or anyone else should be able
to enumerate every one of my properties of either two weeks ago or today.
My lack of knowledge about many of my properties does not seem perti-
nent to my belief that I am the same continuing subject today that I was
yesterday. If I suddenly learned to my surprise that one of my properties
was "the being who was almost killed by his infuriated third grade teacher
on December 1, 1956," I might be amused, but I would not feel that this
ignorance on my part challenged my belief that I was a unified subject over
time. In terms of the unconscious and its contents, even its conflicting and
contradictory contents, a belief in my personal identity over time commits
me to believing that I have had the same, continuing, numerically distinct
unconscious over time—that I have not had someone else's unconscious, so
to speak, during some intervening interlude—but I need not be committed
to knowing all of the contents of my unconscious in order to believe that I
am the same person over time.
The above argument is meant to show that the fact of the unconscious
does not confront the metaphysical question of personal identity where the
claim that there is a personal identity over time amounts to the belief that I
am a continuing subject, numerically distinct from other human subjects.
But perhaps the fact of the unconscious is supposed to address another
metaphysical issue, viz., the claim that a person is a self-conscious being.
This claim often arises when theorists atttempt to make a distinction
between persons and nonpersons. In making this division—in saying what
a member of the class of persons is—philosophers often include as a
characteristic of persons that they be self-monitoring or self-aware. But if
persons have an unconscious dimension, then there are necessarily some
things that they do not monitor and of which they are not aware. There-
fore, the idea of a self-monitoring person is a mythical construct.
MARXISM AND PSYCHOANALYSIS 83

The argument, of course, is too hurried. In defining persons partially in

terms of self-consciousness, it is not necessary to claim that they are con-
scious of everything about themselves. Self-consciousness amounts to the
ability to audit a number of one's own responses but not to understand
every mechanism and ongoing process—biological and psychic—of one's
own organism. The existence of the unconscious, since it does not exclude
the possibility of some areas of self-consciousness, does not directly con-
front the metaphysical claim that there are such things as persons and that
this concept is more than a fabrication.
Often one gets the impression that the real target of the denial of subject
unity in contemporary film studies is the transcendental ego. It seems to
be presumed by film theorists that Lacanian psychoanalysis leaves the
Kantian transcendental ego swamped in its wake. But it is difficult to see the
relevance of psychoanalysis to the claims made for the unity of appercep-
tion by Kantians.45 Psychoanalysis seems rather to pertain to what Kan-
tians would include in the realm of the empirical ego. This is not to assent
to the notion of a transcendental ego, but only to say that psychoanalysis
does not drive it away as neatly as a cross does a vampire.
Nor do I understand why ideologically motivated film theorists are so
preoccupied with attacking the transcendental ego. Other conceptions of
the subject, like Hume's bundle theory and logical behaviorism, have been
conjectured in the capitalist era and these seem to be theories that can be as
readily absorbed into the culture of capitalism as can the ideas of the subject
of Zen Buddhism and Catholicism. Capitalism does not seem to require
one ruling conception of the subject for its continuity. The transcendental
ego is not an indispensable element of capitalist mythology, nor is it clear
that it or any specific theory of comparable generality is a prerequisite for
capitalism. Film theorists, therefore, are not assured of finding the key to all
capitalist ideology through the attempted isolation of something like the
capitalist subject. Lacanian metapsychology, as it blends into metaphysics,
may just be the wrong starting point for the analysis of ideology in film.46
Indeed, I think that it is fair to say that contemporary film theorists, in
general, have not been very careful about their methodological commit-
ments. They defend their choice of the Althusserian-Lacanian paradigm on
the grounds that it situates human action in the realm of meaning rather
than mechanical causation, that it sees the individual as the product of
society, and that it regards language as a central factor in human life. But
that is not enough to justify the specific choice of the Althusserian-Lacanian
paradigm for there are other methodological frameworks which also satisfy
these broad requirements.47 Rather than assessing the relative merits of
84 MARXISM AND PSYCHOANALYSIS

rival methodological frameworks, contemporary film theorists have seized

upon their paradigm more in the spirit o f religious conversion than o f
rational enquiry.

THE EXPLANATORY POWER OF IDEOLOGY

This chapter opened with an interpretation o f the explanatory role that
the concept of ideology plays in Marxist art theory and cultural studies.
There we claimed that the operation o f ideology is thought to explain why
it is that the Western working class has not taken up its predicted role and
spearheaded the transition to socialism. That is, speculation has it that it is
through ideology, and especially through its dissemination via the mass
media, that capitalism has kept the working class quiescent by manipulating
its consciousness. The Althusserian-Lacanian paradigm is but one variant
of this explanation, which we may call the Ideology Hypothesis. But in
concluding this chapter it is useful to ask whether the Ideology Hypothesis,
in any o f its variations, is plausible, where by "plausible" we mean "Does it
even succeed in answering the question it poses?"
The question that animates the Ideology Hypothesis is "Why has the
transition to socialism been delayed in the West?" The answer is that
ideology has taken control o f the minds of the working class. Whether this
is a good answer depends on whether it handles the facts better than any
competing answer to the same question. That is, we should evaluate the
Ideology Hypothesis in the way that we would assess any other hypothesis
o f a rational or scientific nature. Namely, we would compare it to rival
hypotheses and weigh its comparative strength.
In Capital, Marx discusses the ways in which capitalism, once in place,
maintains its hegemony. He writes:
The advance of capitalist production develops a working-class, which bv
education, tradition, habit, looks upon the conditions of that mode of pro-
duction as self-evident laws of Nature. The organization of the capitalist
process of production, once fully developed, breaks down all resistance. The
constant generation of a relative surplus population keeps the law of supply
and demand of labour, and therefore keeps wages, in a rut that corresponds to
the wants of capital. The dull compulsion of economic relations completes the
subjection of the labourer to the capitalist. Direct force, outside economic
conditions, is of course, still used, but only exceptionally. In the ordinary run
of things, the labourer can be left to the "natural laws of production," i.e., to
his dependence on capital, a dependence springing from, and guaranteed in
perpetuity by, the conditions of production themselves.**
MARXISM AND PSYCHOANALYSIS 85

In the first sentence of this passage, Marx includes the operation of

ideology in his account of the maintenance of the capitalist order; educa-
tion, habit, and tradition make capitalism appear natural. However, Marx
goes on to note an even deeper force that keeps the working class in check
which he calls "the dull compulsion of economic relations." Put simply, if
the worker wishes to feed and support herself and her family, she must
submit to the capitalist system and its operation. There is nowhere else to
turn for one's livelihood. The capitalist has possession of the means of
production, and this possession is backed up by the military and the police
even if these forces are only mobilized exceptionally. Force, primarily
economic, but with arms standing in the wings, is the way in which the
capitalist system breaks down all resistance. The compulsion is a dull one but
it is compulsion none the less—if you want to eat, comply. Ideology may
have a subsidiary role in subjugating the masses, but the fundamental
mechanism of capitalist control is economic force.
One thing to note about this explanation in terms of economic force, in
contrast to the Ideology Hypothesis, is that it characterizes the submission
of the working class to capitalism in rationalist terms. The Ideology Hy-
pothesis tends to render the working class's behavior irrational—either
ideology is playing on an irrational, unconscious mechanism like the Imagi-
nary or it succeeds by means of repetitive conditioning. However, the
Economic Force Hypothesis explains the maintenance of capitalism by
means of the rational decisions of workers. Given the necessity of securing
one's daily bread, and given the socially available means at one's disposal, it
is a matter of rational choice, a simple practical syllogism, that the worker
complies with capitalism.
Marx, of course, was writing in the nineteenth century, whereas contem-
poraryfilmtheorists want to know how capitalism sustains itself today. But
there is no reason to think that the explanation has changed in its essentials.
We submit to capitalism, even though we may be aware of its injustices and
perhaps aware that economic arrangements could be otherwise, because we
must in order to survive. Of course, capitalism has changed since Marx
wrote Capital; many capitalist states have instituted welfare reforms which
to varying degrees supply us with the means to a livelihood even if we are
unemployed. But this change in capitalism makes our compliance even
more rational. That is, given the choice between compliance and revolt,
surely compliance makes more short-term sense. Revolt is both risky and
difficult. There is no guaranteed success, and success would require mo-
mentous effort. Compliance with the capitalist economic system, especially
where it is supplied with welfare nets, is frankly the line of least resistance,
86 MARXISM AND PSYCHOANALYSIS

even for those who realize that the distribution of economic goods could be
more equitable than it is.
It may be objected that the explanation just offered is too individualistic,
too centered in the rational decision making processes of atomic agents.
The real calculation should be made in terms of the prospects for revolt of
the workers as a group, united by class consciousness. Several things need
to be said here. First, the Economic Force Hypothesis as stated does speak
of the calculations of individual workers, but, then, the Ideology Hypoth-
esis also, one supposes, refers to the irrational processes of individuals. If
the Imaginary does play a role in ideology, then it must be recalled that
ideology is said to address individuals. Also, even if one supposes that the
working class could coordinate itself, one would still have to take into
account the risk and effort that revolt would entail. Compliance might still,
at this point in history, be the line of least resistance as determined by a
rational calculation of the options of the working class in the industrial
West.
Of course, so far we have spoken in terms of revolt. But the proponent of
the Ideology Hypothesis may wish to reframe the question in terms of
electoral politics. Why is it that the working class has not installed socialism
by parliamentary means? Of course, in some cases, there is socialism in the
West. But there are countries, like the United States, where, at present, the
prospects are slight. Doesn't it, it might be argued, seem likely that the
reason for this is the effectiveness of ideology? However, a more compel-
ling answer might be that the absence of viable alternative socialist politics
in places like the United States make compliance to capitalism the rational
alternative for each worker. Moreover, the effort required to organize a
viable socialist party and the risk of failure again make rationally compre-
hensible the reason most workers will choose to suffer the status quo.
What, of course, is needed in places like the United States are people who
are so committed to socialism that they are willing to spend the effort and
take the risks against formidable odds. But until they succeed in fielding a
socialist party that is a viable alternative to existing political parties, the
inactivity of the majority of the working class will be sensible and the
electoral road to socialism foreclosed. I do not believe socialism is impossi-
ble in the United States. But until a feasible socialist political alternative is
in place, the submissiveness of the working class is explicable in rational
terms.
The Economic Force Hypothesis, then, is a rival explanation to the
Ideology Hypothesis. Are there any facts that the one hypothesis can
handle better than the other? I think that there are and they can be summed
MARXISM AND PSYCHOANALYSIS 87

up in a sentence: the subjects of capitalism are often very aware that the
ideology disseminated by modern society is phoney. Surveys show that
workers believe that big business has inordinate power in society, that the
upper class controls government, and that rich and poor are treated dif-
ferently before the law; nor are these the only areas where people are
frequently aware that the pictures ideology offers are lies.49 Moreover, that
which statistical surveys have made abundantly clear, can easily be con-
firmed in one's own experience. Conversation with one's (non-academic)
acquaintances and family relations will quite often reveal that Uncle
Joachim, Cousin Frieda, and Barney the Oldsmobile mechanic know that
politicians lie, that the defense budget is padded with bribes, that hard
work is no substitute for being the boss's nephew, that landlords are legally
advantaged in their relations with tenants, that the tax laws favor big
business, and so on. This is not to say that every member of the under
classes has a crystal clear view of capitalism. Nor is it to say that one won't
meet a worker who buys everything William Buckley and Irving Kristol
write. Rather, the claim is that to a surprising degree ordinary people are
aware that capitalism is unfair and its ideological self-images shams.
These simple facts count heavily against the Ideology Hypothesis.
Clearly, they indicate at the very least that ideology is not working as
consummately as proponents of the thesis maintain. It cannot explain the
submissiveness of Western workers in terms of those workers' minds being
utterly inculcated with ideology, for the fact is that in large measure
workers reject the tenets of ideology. Under the Althusserian dispensation,
we are told that subjects always "misrecognize" the real conditions of
existence; we are never told why subjects must always necessarily misrecog-
nize; nor does the claim appear to fit the facts.50
On the other hand, the Economic Force Hypothesis is in a much better
position to accommodate these facts about workers' beliefs. It says that it is
perfecdy consistent for workers to believe that capitalism's self-images are
bunk, while at the same time they, the workers, think that in the interests of
survival and comfort submissiveness to capitalism is the rationally advisable
course. The Economic Force Hypothesis makes us cynics; the Ideology
Hypothesis makes us cyphers. The former, it seems to me, better squares
with reality.
As contemporary philosophy of science has shown, rational inquiry
involves the consideration of a hypothesis by comparing it to rival explana-
tions in order to determine which is stronger in the face of facts judged to
be settled by competant observers. When the Ideology Hypothesis is placed
next to the Economic Force Hypothesis—which also attempts to explain
88 MARXISM AND PSYCHOANALYSIS

the inaction of the working class—the latter is the more powerful explana-
tion because it can accommodate the facts that the under classes of capital-
ism do not, to an arresting degree, believe the capitalist ideology.51 The
Ideology Hypothesis just can't be true. If it were, we would expect every
Tom, Dick, and Mary to be parroting a combination of Horatio Alger,
Fortune, and Milton Friedman. But that is not the way it is; how can
ideology be holding capitalism in place if so many disbelieve it?
Of course, to deny that the Ideology Hypothesis can explain the in-
action of the under classes in capitalism does not deny that there is such a
thing as ideology. Nor does it deny that ideology plays a role in modern life.
What it rejects is the idea that ideology is as important as Western Marxists
have claimed it to be.52 And we have also contested the notion that it is as
overpoweringly effective as upholders of the Ideology Hypothesis pre-
sume. Very often, ideology just does not work. Models, like the Althusser-
ian-Lacanian paradigm, which make ideology appear an omnipotent force
from whose grasp escape is impossible, are off the mark.
The tendency to overrate the power of ideology is particularly rampant in
contemporary film studies. The ruling paradigm invests cinema, especially
mass entertainment movies, with too much responsibility and too much
effectiveness. It has movies performing a major role in positioning capitalist
subjects. But that can not be right. Even an Althusserian would want to
stress that the educational system is far more responsible for molding
subjects than the movies.53 And a non-Althusserian would go even farther,
suggesting that neither the school system nor the movies is as important as
the structure of economic relations backed by organized force for getting
subjects to accept their subjection.
Not only have contemporary film theorists given movies too much
responsibility; they suppose movies and the media to be far more effective
than they are. Even Americans know that their advertisements and mass
movies arc ridiculous; satires like Saturday Night Live could not exist
otherwise. And viewers of Dallas take J.R. neither as a role model nor a
reflection of reality but rather as entertaining.54 Spectators have far more
distance from the ideology flickering on their screens than contemporary
film theory allows. By overestimating how effective and convincing media-
made ideology actually is, theorists, in this respect, take their subjects, and
perhaps themselves, more seriously than they ought.
 3.

THE CINEMATIC
IMAGE

J . HE consensus among contemporary film theorists is that their central

task is the identification of the operation of ideology in film. In this they are
concerned not only with the way in which specificfilmspropound ideology
or have political effects. More importandy, they are interested in analyzing
the way in which the very structures of cinema—such as the cinemato-
graphic image, narration, and editing—are ideological. That is, they pre-
sume that these structures of representation are inherendy ideological.
There is, of course, no mistaking the obvious ideological slant of films like
Invasion USA. and Top Gun. However, one might think that there are
movies like The End of St. Petersburg and Mother which, though they
employ the same structures of representation as our more recent examples,
would not be chastized as ideological by a Marxist. Yet, the contemporary
film theorist maintains that any film that deploys the cinematographic
image, narration, and editing in what is called the realist style will be
ideological in nature.
In the next three chapters, we will listen to what contemporary film
theorists have to say, respectively, against the cinematographic image,
narration, and editing. I will attempt to defend a countervailing hypothesis.
Stated briefly, I deny that structures of representation, at the level of
abstraction discussed by contemporary film theorists, are essentially ideo-
logical. In my view, the ideological operation of films resides, roughly
speaking, in their content and its rhetorical inflection rather than in their
use, simpliciter, of cinematography, narration, and what is called classical
editing.
90 T H E CINEMATIC IMAGE

This chapter is specifically concerned with the cinematographic image as

it is typically employed in movies, though this will require some anticipa-
tory discussion as well of film narration, since many contemporary theories
see an intimate relation between the standard use of the cinematic image
and narrative filmmaking. When speaking of the "cinematic image," I will
talk about it at the level of the single shot, and about single shot images of
the sort that are described by such terms as verisimilitude. I will refer to
these images as mimetic images, images that, pre-theoretically, people are
disposed to say represent their referents by way of resembling them. Ab-
stract, nonrepresentational and/or nonmimetic images can be found in
cinema, such as Eggeling's Symphonic Diagonale. But this is not the use of
cinematography generally found in films. Most shots in most films and
movies are pictorial; we see a shot of a horse and immediately recognize its
referent in the moving picture before us. Throughout this chapter, it is this
type of example which we shall have in mind when speaking of the cinemat-
ographic (or cinematic) image.
The chapter is divided into three sections. The first section deals the
notion, popular among contemporary film theorists, that cinema is illu-
sionistic and, correspondingly, that illusionism is important in the explana-
tion of the ideological effect of film. In the following sections, I will
consider the views of contemporary film theorists in regard to the origin of
the cinematic image in photography and the use of perspective as a means
of organizing space in shots. Contemporary film theorists argue that both
the use of photography and perspective in film have ideological repercus-
sions. I shall challenge these claims.

ILLUSIONISM
Contemporary film theorists bear a marked hostility to a trans-art, trans-
media phenomenon which they call "illusionism" and which, effectively,
can often be taken as a cognate for something referred to as either "classic
realism" or just "realism" (which will be characterized in the next chapter).
Correlatively, contemporary theorists frequently endorse anti-illusionist art
for the ways in which that art undermines the illusions that realism pro-
motes. Brecht's theater is championed, for example, because it is "a reaction
against the perspective tradition of the post-Renaissance world; which
posited the eye (and the man behind it) as the centre of the world and art as
a window (therefore transparent) on that world."1 That is, the pernicious
effect that post-Renaissance, realist art imparts, and which Brecht seeks to
dispel, is the illusion of reality.
THE CINEMATIC IMAGE 91

Though reference to illusionism is more a trope of the seventies than of

the eighties, where the notion of transparency often replaces it, the idea of
illusionism is still worth discussing not only because it continues to crop up
but because its use indicates the degree to which contemporary theory is
motivated, often implicitly, by a commitment to modernist aesthetics.
Representation, in the sense of classic realism—which encompasses liter-
ature, theater, fine art, and film—is suspect because it is illusionist, a term
one might use to describe the deceptions of a magician. The initial source of
this criticism of representation, for contemporary theorists, seems to be
Brecht and his diatribes against the dramatic tradition of the West. Con-
temporary theorists were attracted to Brecht for two reasons: Brecht was an
important influence on Jean Luc Godard, an exemplary artist for contem-
porary film theorists, as well as a central reference point for Roland Barthes,
perhaps the exemplary cultural critic of contemporary theorists.
Though Brecht is hardly consistent in this matter, in his railings against
traditional theater, he, on occasion, accuses its representational practices of
instilling illusions of reality which, in turn, paralyze the spectator's capacity
to make judgments, notably judgments of a politically critical nature. That
is, naturalistic theater engenders the illusion that one is in the presence of an
actual, i.e., nonfictional, event, which illusion, in consequence, putatively
immobilizes our ability to criticize what we see, especially from a political
point of view. Speaking of the virtues of his Epic Theater, in contrast to the
vices of naturalism, Brecht writes:
Just as the composer wins back his freedom by no longer having to create
atmosphere so that the audience may be helped to lose itself unreservedly in
the events on stage so also the stage designer gets considerable freedom as
soon as he no longer has to give the illusion of a room or a locality when he is
building his sets. It is enough for him to give hints, though these must make
statements of greater historical or social interest than does the real setting.2

Restoring the theater's reality as theater is now a precondition for any

possibility of arriving at realistic images of human social life. Too much
heightening of the illusion in the setting, together with a "magnetic" way of
acting that gives the spectator the illusion of being present at a fleeting,
accidcntal "real event," creatc such an impression of naturalness that one can
no longer interpose one's judgments, imagination and reactions, and must
simply conform by sharing in the experience and become one of "nature's"
objccts. The illusion created by the theatre must be a partial one, in order that
it may always be recognized as an illusion.3
These passages indicate that Brecht, at least at moments, believed that the
techniques of visual mimesis in traditional stagecraft create illusions of
92 T H E C I N E M A T I C IMAGE

reality in spectators. Moreover, like Plato and Kant,4 Brecht takes a dour
view of what he assesses to be the illusions promoted by visual verisimili-
tude. Furthermore, not only are the techniques of visual verisimilitude
ranked as illusions in-and-of themselves; they also facilitate further illusions
or, more aptly, false beliefs. The belief that what is enacted is a "real event,"
supposedly stuns our critical faculties so that we believe "that's the way
things really are," with nothing else left to be said.
This latter effect is also enhanced by the "compelling momentum"5 of
plotting in classical theater which suggests not only that what was enacted
was a "real event" but that the way it was presented was the way it had to
happen, which implies, purportedly, that society, the social world, cannot
be changed. This, of course, has consequences of ideological significance
where the portrayal in question—say, of the impossibility of revolutionary
activity—serves political purposes; for it is tantamount to admitting and to
advertising that nothing (politically efficacious) can be done. Lasdy, Brecht
sees our critical faculties waylaid by the "magnetic acting" of traditional
theater which induces the illusion of identification between the spectator
and the characters. He writes:
Even today we willingly are happy to overlook such inaccuracies if we get
something out of the spiritual purifications of Sophocles or the sacrificial acts
of Racine or the unbridled frenzies of Shakespeare, by trying to grasp the
immense or splendid feelings o f the principal characters in these stories. 6

Thus we see that Brecht associates a cluster of illusion effects with the
practices of traditional theater. There are the visual illusions of verisimili-
tude which illusions trigger further acceptance of certain nonvisual illu-
sions, viz., false ideological beliefs. Next there is the illusion of compelling
momentum, achieved through plotting, which gives rise to the false belief
that things cannot be otherwise. And there is also the distracting illusion of
identification, which is thought to emotionally rivet us to the spectacle in a
way that renders our critical faculties inert. So, the representational prac-
tices of traditional theater cause several illusions—of reality, of inevit-
ability, and of identification—and these illusions each in its manner makes
us susceptible to ideological falsehood; one might say they pave the way or
are causal conditions for our acceptance of ideology.7
Though Brechfs criticisms of representation/illusion are developed with
regard to realist theater, the categories of illusion that he mobilizes are
broad enough to be extended by contemporary theorists to arts and media
other than theater. The illusion of reality, caused by techniques of ver-
isimilitude, is attributed, by some contemporary theorists, not only to
T H E CINEMATIC IMAGE 93

theater, but to fine art, photography, and cinema. Indeed, some theorists,
oddly enough, apply this category to non-imagistic media, such as writing.8
The illusion of inevitability, insofar as it is rooted in plotting, can be sought
by contemporary theorists in any narrative media, including literature, TV,
and film. And the illusion of identification might be thought to occur in any
art where there are characters with whom one could identify. The illusion
effects that Brecht deprecated with respect to tradition have been, in short,
generalized across the arts by contemporary theorists.
In later chapters, we shall be concerned with the purported illusionism of
the narrative. But presently we are concerned with the attribution of
illusionism to cinematic images. This involves two logically separable but
nevertheless connected issues: the illusionism of the cinematic image and,
then, the capacity of such illusionism to neutralize our critical capacities in
such a way that we become suckers for ideology.
As noted, contemporary film theorists have been prone to call mimetic
pictorial representations "illusions." This is a term which disdains such
representations in and of themselves, irrespective of what is being repre-
sented. For "illusion" is a label that signals deceptions, as practiced by
magicians and, before them, demons. Calling mimetic representations "il-
lusions," that is, already castigates them, for even nowadays, visual illusions
proper involve the deception of the percipient; and where such phenomena
are deliberately contrived to such an effect, one might think of them as a
species of visual "lies." "Illusion" is a designation with primarily negative
connotations, ones that indicate deceit. Since "illusion" is denigrating, we
understand that contemporary film theorists are contemptuous of mimetic
representations when they call such representations "illusions." However,
at the start, we must pause to ask whether it is appropriate to call mimetic
representations "illusions."
Consider the realist, mimetic representation of a tree; it might be a stage
flat, a snapshot, a painting, a statue, or, for our purposes, a strip of film, a
shot. Why would one say that a realist shot of a tree is an "illusion"—
indeed, does it even make sense to call such an image an "illusion?" Perhaps
we shall be told that the shot of the tree is an illusion because it is nothing
more than a substitute or proxy for the tree, at best it is an ersatz tree. But
this implies that what we really always want is the tree itself. Yet there
would be no point to having an institution of pictorial representation if
what we really always wanted is the referent of a representation rather than
the representation.
There are many purposes—including cognitive, emotive, and aesthetic
ones—that lead us to want representations of things rather than the things
94 T H E C I N E M A T I C IMAGE

themselves. We are able to derive information about the nature and qual-
ities of things by considering representations of them—e.g., the mechan-
ical drawing of a motor or a portrait of Isadora Duncan by Gordon Craig—
to which we are often insensitive when confronted with the real thing.
Also, certain of our interests in traditional paintings and, for that matter,
realist film images—such as the appreciation of their life/ifcmess—would
make no sense whatever if it were the real referents rather than their
representations that we were savoring. That is, we have reasons for the
institution of representation. And those reasons can be used to explain why
certain paintings, snapshots, film strips, etc. should be substitutes or
proxies—albeit not substitutes of their referents in every respect or role but
only in virtue of an institutionally restricted set of respects and roles. That
is, for the purposes of contemplation, a picture or shot of x, moving or
otherwise, it would seem, plays the role of x by means of presenting a
delimited range of recognizable similarities to or cues for x. But if certain
realist paintings, or snapshots, or film images should serve as proxies for
their referents, why disparage that fact about them by calling them "illu-
sions"?
Indeed, I question whether it is logically correct to call realistic represen-
tations "illusions." To call them "illusions" suggests that they comport
themselves convincingly as something they are not meant to be—trees,
motors, and Isadora Duncan. But realist representations are meant to be
partial proxies within the institution of representation. That is, in our
society, realistic images, including film shots, are cultural currency. They
are disseminated and recognized as the kinds of proxies they are. In fact,
they must be so recognized in order to perform many of their cognitive and
aesthetic functions.
Perhaps it will be urged that paintings, sculptures, photographs, stage
settings, and film shots are illusions of real things because they are some-
how incomplete. They are always less than the things they stand for—thev
do not have all the properties of the things they stand for (and, in fact, they
have some properties that their referents lack). This seems to imply that the
only representations that would not be illusions would be perfect replicas,
that is, facsimiles that differ from their referents only in such numerically
individuating features as spatial location. But if representations had alwavs
to be perfect replicas of their referents, we would, again, have strong reason
to wonder why we have an institution of representation. For it seems likely
that, in most cases, anything we could learn from or appreciate about the
perfect replica, we could have gotten straightaway from the original. There
seems little point in criticizing a realistic picture as an illusion because it is
THE CINEMATIC IMAGE 95

not a perfect replica. It is not an illusion because it is incomplete. Rather it is

a representation, the very sort of thing that does not typically propose a
perfect replica of its referent. Moreover, it is this feature of incompleteness,
in part, that enables it to perform its various tasks within the frameworks of
the art world, science, etc. Just because, sensibly enough, representations
are not, generally, perfect replicas, there is no reason to conflate the idea of
realistic representation, as a viable institutional practice, with "illusion."
Of course, the ultimate reason not to equate mimetic representations
with "illusion" is that anything properly called an illusion either deceives or
is liable to deceive the percipient of the putative illusion. To be a full-
fledged illusion, the percipient would have to be fooled into believing that a
picture of a tree is a tree, or, at least, a window onto a tree, or the percipient
would have to be deceived about some "tree-derived" property of the
painting or vice versa—if we think, for example, that a mimetic picture of a
tree is fifty-five feet high, or that the tree pictured in a postage stamp is
actually an inch high, we are suffering an illusion. And, of course, it is
probably this sense of illusion, the deception sense, that contemporary
theorists have in mind when they chide the illusionism of mimetic represen-
tation. But are mimetic representations—stage flats, photos, paintings, and
film shots—convincing illusions in this sense?
Needless to say, it may be possible to stage a situation in such a way that a
percipient might be deceived by a mimetic representation of x so that he
reports that he sees x. Certain psychological tests where the monocular
station point is fixed, where the edges of the array are occluded and the
percipient is stationary suggest this effect can be achieved. That is, under
very special circumstances a mimetic representation might be presented in
such a way as to deceive normal percipients. However, such cases have little
bearing on our own normal interactions with mimetic representations. For
normal viewers in standard conditions are not deceived.
Contemporary film theorists extend the Brechtian characterization of the
illusionistic nature of mimetic stagecraft to all forms of mimetic representa-
tion, including the cinematic image. But what can "illusion" mean here?
Clearly it suggests that these things lead to false beliefs. But this claim must
be even further specified because anything probably can be used to lead
someone into a false belief at some time. An illusion, rather, must be some-
thing that has a high probability of leading normal spectators into false
beliefs. But what false beliefs do snapshots and still lifes, on the one hand,
and naturalistic staging and cinematic images, on the other, encourage
normal viewers to embrace? One supposes that with snapshots and movie
close-ups of apples, spectators—like Pliny's famous birds—believe that
96 T H E C I N E M A T I C IMAGE

edible apples confront them. And with theater and cinema, the illusion of
reality amounts to viewers somehow mistaking the events depicted by these
means of representation for actual events, transpiring here and now. But
this is just false. No one thinks that downtown Tokyo is in the screening
room during Mothra—let alone Mothra in person; how could they be?
Likewise, theatergoers do not literally believe that Birnam Wood, moving
or otherwise, grows in their local theater. Indeed, it would be perplexing if
people were deceived by representations. How would we account for the
very different cognitive and aesthetic responses, not to mention the be-
havioral responses, that we make to representations of things versus the
things themselves? Who would sit by as buffaloes or the Light Brigade
charge at us or while Romeo fatally misjudges Juliet's slumber? Typically,
we must know that we are viewing mimetic representations in order to
respond properly to them. And typically we have this knowledge and do
respond properly. We are not deceived by some illusion of reality which
would, in any case, interfere with the proper functioning of mimetic repre-
sentation.
The mistaking of plays and films for actual "real-life" events by benighted
yokels is a standard, universally appreciated and age-old gag of both theater
and film just because one would have to be so dim-witted to make such an
error. Such yokels, contra contemporary film theory, are not paradigmatic
spectators. Plays and films, when seen in standard viewing situations, do
not look enough like the events and locales outside of the theater to be
mistaken for them. Things like the monocular station point, scale variation,
and black-and-white cinematography work against film being taken for
reality, while the missing fourth walls should persuade uninformed the-
atergoers that something is awry. But, of course, the telling point is that
film viewers and theater viewers are informed; theatergoing and filmgoing
are institutionalized activities; and the normal spectator in the normal
viewing circumstance is there to see representations and not the real thing.
The apocryphal yokels are funny because of their extraordinary ignorance.
They are not normal viewers. They are such stereotypical comic butts just
because everyone else, the normal viewers, can feel superior to them.
There is, perhaps, a sense in which "x is an illusion of v" simply means "x
looks like y" (if we opt for resemblance theories of pictorial representation).
A stage flat is an illusion of a house if it looks like a house in certain relevant
respects. This does not mean that anyone is fooled by the stage flat into
believing there is a house on stage. Nor does this mean that the stage flat
looks exacdy like a house. Indeed, it might not look like a house at all in the
sense that it could be taken for a house, and yet it might look like a house
T H E CINEMATIC IMAGE 97

across a limited number o f dimensions—it has a door and windows while

also being obviously a piece o f canvas. In this sense a spectator who calls the
stage flat an illusion might only mean that she thinks that it looks like a
house because it has certain representationally relevant similarities to a
house; but saying this does not commit her to believing that the stage flat is
a house.
Let us call this latter use o f "illusion," the epistemically benign sense o f
illusion. The Brechtian sense o f illusion, as it appears to be advanced by
contemporary film theorists, however, regards mimetic stage flats and film
images in a way that suggests that normal spectators are deceived or
ensnared by verisimilitude. Their argument seems to proceed by initially
describing its objects by means o f the benign sense o f "illusion" and then
fallaciously switching to the deception sense o f "illusion" mid-proof. In
respect to its capacity for equivocation, perhaps the benign sense of illusion
is not so benignant, and, if there is such a sense, maybe it would be best to
drop it altogether.
Plays and films are often spoken o f as if they were optical illusions—
presenting misleading appearances, like straight sticks in streams. This talk
may prompt theorists to regard plays and films as deceptions. But such
theorists are victims o f their own metaphors. For optical illusions only
become full-blown illusions when they provoke the formation o f false
beliefs in the face o f obvious counter-evidence or information. And as
psychologists never tire o f pointing out, human organisms do not simply
rely on what is imprinted on their retinas for their beliefs but also depend
on other evidence and information that surround the moment o f percep-
tion.
I do not parlay the image in my bathroom mirror into a belief that my
doppelganger is hungover before me not only because I know about
mirrors (though that is very crucial) but also because the doppelganger
hypothesis is refuted by my other senses—just as my hand corrects my eye
when dealing with straight sticks in water. The likelihood that a normal
perceiver could be deceived by a film image drops precipitously as soon as
he can perform physical operations in relation to the screen; I can change
mv position enough (even in my theater scat) to realize that the visual array
is not three-dimensional, if I am ever in doubt about the film's oddly
glowing "realities." For example, I can move my head and note that such
shifts do not result in gaining new information about the edges of objects—
that is, the visible edge o f the object stays the same irrespective o f my spatial
relation to the visual array; thus, the array cannot be three-dimensional.
And with theater and film, even if there were not ordinarily, in fact, other
98 THE CINEMATIC IMAGE

perceptible cues, my knowledge that these visual arrays are presented

within the institutions of theater and film would be enough to block the
occurrence of (epistemic) illusion, just as I would reject these arrays as "real
life" because they do not fit coherently with the rest of my knowledge of the
way the world is physically, architecturally, and socially (do couples usually
squabble in kitchens with views fronting on fifteen-hundred seat auditor-
iums?). And, of course, though plays and films may employ optical illusions
as constituent elements, entire plays and films are not optical illusions. Thev
are marked and disseminated as what they are—plays and films—and they
trick no one except theoreticians.
Furthermore, there are certain facts about the normal viewing of mimetic
representations that lead to logical conundrums for the theorist, artist, or
critic who propounds an illusion or deception theory of representation. For
example, representational pictures—films, photos, and paintings—sup-
port irregularities like scratches, grain, uneven applications of paint, thick
impastos and varnish, markers for reel changes, etc., which the spectator
must "see through" in order to comprehend the represented scene. Like-
wise, the playgoer must "see past" the linoleum stage floor to apprehend
Arden forest. But "seeing through" and "seeing past" such surface distor-
tions and perceptual noise presupposes that the percipient knows she is
looking at a painting, a photo, a play, or a film, and that she knows the
proper way of attending to the objects so identified. However, the illusion
or deception theory contends that the viewer believes that the representa-
tion is its referent. But if the normal practice of "seeing through" surface
distortions entails that the viewer knows it is, say, a film, then it follows that
the normal viewer believes she is viewing a film, and not its referent. That is,
combining the deception theory with certain commonplace facts about
painting and other forms of visual representation along with the notion
that knowledge entails belief, we derive a contradiction, viz., that the
viewer both believes and does not believe that she is viewing a film.
Many other factors about the normal practice of mimetic representation
count against the illusion of reality theory. For example, the use of framing,
as a convention, emphasizes the discontinuity of the subject of the represen-
tation and the adjacent physical environment, thereby forestalling the al-
ready unlikely possibility that the knowing viewer will be traduced. And, of
course, to appreciate many of the aesthetic features of mimetic representa-
tions—such as pictorial depth—one must know that one is looking at a
certain form of representation. But, again, if a percipient knows she is
looking at a mimetic representation then she believes it is a representation
at the same time the illusion-of-realitv theory says she believes that it is not a
representation, but a reality.
T H E CINEMATIC IMAGE 99

The illusion of reality plus the mundane facts of interacting with repre-
sentations result in an embarrassing inconsistency. The spectator is said to
both believe and not believe that the representation is a representation. At
this point, we must consider whether one should abandon one wing of this
contradiction, or accept that the spectator is in a contradictory state and
postulate some psychic mechanism, like Metzian disavowal, to explain how
the spectator tolerates this contradictory state. Most contemporary film
theorists appear to prefer the latter tack. But I, instead, advocate that we
simply abandon the illusion of reality thesis, i.e., that we deny that the
spectator takes the representation for reality, thereby believing it is some-
thing it is not. Undoubtedly some such process as Metzian disavowal
would have to be operative if viewers did simultaneously believe and not
believe that mimetic representations were their referents. But if we say that
viewers believe their representations to be representations and if we reject
the illusion theory, then we have no need to conjecture a mechanism of
disavowal. Not only, I submit, is my recommendation more economical
but it squares with the facts better than the alternative approach.
The reasons for rejecting the notion that viewers mistake mimetic repre-
sentations for reality are of two types. First, what I have referred to as the
ordinary facts of attending to representations militate against the claim that
viewers misapprehend the nature of the visual arrays before them. And
second, the accounts of the peculiar state that spectators, supposedly vic-
timized by illusions of reality, are said to be in are uniformly unconvincing.
For example, spectators are said to "suspend disbelief." 9 This suggests
that when confronted by a mimetic representation, I do something to
myself or something happens to me over and above my recognizing that
which the representation is a representation of. Phenomenologically, I have
no sense of such an internal process. Nor am I inclined to believe that it is
within my power to convince myself that the painting or film before me is
not a representation but its referent. That is, contra Descartes, 10 my beliefs
and disbeliefs are not things that I will. I cannot will to believe that
" 2 + 2 = 4 " nor that my chair has just dematerialized. Believing and dis-
believing are not things that I do to myself. My beliefs and disbeliefs, so to
speak, just happen as I consider matters. Of course, one might say that I can
pretend that my chair has dematerialized. But to pretend that would require
that I know the chair has not dematerialized. That is, the concept of
pretending is of no use to theorists who wish to claim that spectators of
mimetic representations are deceived by illusions of reality.
Of course, it might be said that the willing suspension of disbelief is not
something that I do but rather something that happens to me. But I must
protest that I, at least, have no recollection of ever having taken, for
100 T H E C I N E M A T I C IMAGE

example, a portrait or a close-up for a person. Indeed, I have always

supposed that what "suspending disbelief' comes to is my decision not to
criticize the improbability of the events in certain stories—e.g., Tarzan
teaching himself to read in the novel of the same name. That is, "suspend-
ing disbelief' really means bracketing criticism of narrative improbabilities
(rather than, say, immoralities) in certain appropriate circumstances. Re-
fraining from criticizing or being bothered by implausibility is something
that is within my power, but it does not affect my beliefs. In fact, if this is
the correct interpretation of the "suspension of disbelief' then, in order to
do it, I must be continually aware that I am in a certain appropriate con-
text—such as viewing a fiction film about life on Mars.
One might, following a suggestion of Gombrich's,11 attempt to charac-
terize my mental state before visual representations, such as pictures, on the
model of "seeing-as," which Wittgenstein applied to ambiguous figures
such as the duck-rabbit. How would this work? Well, it might be claimed
that we see the traditional realist picture as flat (as a canvas) and then it
appears to have depth (as a scene in nature)—first it is a picture plane, and
then it is seen as a three-dimensional expanse. But the experience of ambig-
uous figures does not appear to be a proper analog to my experience of
viewing traditional pictures or films. I do not flip-flop between seeing
pictures as flat, then seeing their referents as three dimensional in the way in
which I shift from first seeing a duck and then a rabbit. There are not
complete transformations of my visualfield.With ambiguous figures I must
see the figure first one way and then in another, mutually exclusive, way.
But when I see a picture I both recognize it is a picture while recognizing
what it is a picture of. These two recognitions are simultaneous and not
sequential in the manner of ambiguous figures. Nor are there heuristic
promptings or pointings, as there are with ambiguous images, which will
induce a comparable flip-flop between two ways of seeing mimetically
representational pictures.
One contemporary film theorist, Dudley Andrew, has attempted to
employ the "seeing-as" account of mimetic representation in terms of an
oscillation—not between two-dimensionality and three-dimensionality—
but between belief and doubt. Andrew writes:
Quite simply the oscillation at the heart of all instances of "seeing-as"
becomes in the cinema a vacillation between belief and doubt. That is, as
participating in the familiar world of our ordinary experience yet then slip-
ping into its own quite different screen world. Only an unusually strong act
of attention enables us to focus on the light, shadow and color without
perceiving these as the objects they image. And, on the other side, only an
T H E C I N E M A T I C IMAGE 101

equally strong hallucinatory mode of attention can maintain from beginning

to end the interchangeability of what we perceive and the ordinary world,
negating all difference of image and referent. Cinema would seem to exist
between these two extremes as an interplay between "the real and the image."
The film experience in general and in every instance of viewing a film can be
analyzed in terms of a ratio between realistic perceptual cues and cues which
mark an effort and type of abstraction.12

The first thing to note about this analysis is that its expansion of Gom-
brich's use of "seeing-as" is somewhat strained, if not frankly illicit.
Gombrich suggested a scenario with respect to oscillating visual experi-
ences. Andrew has turned it into an oscillation between epistemic states—
belief and doubt. Why is this a problem? Obviously, the visual oscillation
Gombrich invokes could occur—as it does in the case of the duck-rabbit
figure—with no epistemic repercussions for the percipient. To find the
interplay of belief and disbelief at the heart of "seeing-as" is nothing but a
non sequitur, as is Andrew's later equation of Gombrich's oscillation with an
oscillation between alternately taking the cinematic image to be real and
then unreal. I have already questioned the reliability of a seeing-as theory in
the preceding paragraphs. But even if that theory were viable, Andrew's
derivative theory would be no better off, since despite Andrew's assertions,
the play of belief and doubt docs not follow from seeing-as in the case of
ambiguous figures.
Viewed independendy of the purported connection with "seeing-as,"
Andrew's claims of an oscillation between belief and doubt appear al-
together unsubstantiated. Who recalls suffering this vacillation while view-
ing a film. First I believe Dorothy and the Tin Man are in front of me; then I
doubt it; then the oscillation starts up again. One would have litde oppor-
tunity to attend to the story if one were on such a merry-go-round. Perhaps
this oscillation goes on unconsciously. But that is a pretty elaborate piece of
psychic machinery to postulate; what behavioral or even introspective
evidence motivates positing such a psychic process?
So far I have questioned the intelligibility of the states attributed to
spectators by deception theories of mimetic representation. But what of the
producers of such artifacts? Is it plausible to attribute to them the desire to
fool audiences into the belief that the spectacle before them is a "real event,"
an event transpiring in the spectators nonfictionally existing "here and
now?" Imagine the chagrin of the makers of Red Dawn had the opening of
the film not been reported in the arts section of the New York Times but on
the front page as a news item under a banner headline reading—"World
War III Breaks Out on Broadway." Furthermore, if movie makers ever
102 T H E CINEMATIC IMAGE

succeeded in convincing audiences that they were witnesses to "real

events," such filmmakers would probably alienate our affections. Who
wants to be bombarded by the Empire's laser cannons? In short, the idea
that the practice of mimetic representation can be explicated by reference to
an epistemically pejorative concept of illusion is thoroughly confused.
Defenders of the position may claim that there is some special formulation
of the illusion effect that I am insensitive to. If so, then it is up to them to
produce it.
Undoubtedly, it will be argued that my account of attempts to theorize
the illusion effect is too harsh. For, it may be urged, that whatever failings
such theories have had so far, said theories are attempting to explain certain
observable phenomena that call for explaining. That is, illusion theorists are
struck by the way in which we are riveted to mimetic representations. And
they attempt to explain this by saying that we must be responding so
intensely because we believe that the events represented are "real events."
Why would audiences scream and cry, why would they sweat and cheer,
unless they believed that the events before them were really occurring?
But I do not think that it is useful to postulate illusion effects to explain
our intense reactions to mimetic representations. For example, we may be
just as engaged in and riveted to live basketball games where there is no
question of representation or illusion. We can explain this simply in virtue
of our interest is what is going on, even if this occurs in a "ludic" ontological
space rather than "reality." The reasons accounting for our high level of
interest can be further specified sociologically and with reference to our
personal psychology. I see no reason to treat the intensity of our engage-
ment with mimetic representations as different in kind than our responses
to the live sporting events we attend. Our interest explains the degree of
intensity of our response, not our beliefs about ontological status.
Also, we generally behave very differently toward the objects and events
we encounter in our ordinary, non-illusory daily life. When we take the
events to be "real," speeding cars do not thrill us, butcher knives do not
appear ominous, nor bickering couples melodramatic. If mimetic represen-
tations were taken to be illusions of reality, we would respond to them as
we ordinarily respond to the events we witness and the objects we observe
in "reality"—that is, generally, with circumspection, reserve, and often
obliviousness, nonchalance, or an utter lack of interest. Thus, proposing
that mimetic representations cause illusions of reality will not explain why
we are riveted to such spectacles, since reality, so to speak, is so rarely
riveting.
Instead of using the concept of illusion to explicate our relationship to
T H E CINEMATIC IMAGE 103

the visual arrays of mimetic representation, I would prefer to speak of

recognizabilitv. Afilmimage, a photo, or a portrait represents x and, when
it is successful, the spectator recognizes x in the representation, i.e., recog-
nizes that the representation is a representation of x. The factors involved in
securing this recognition are in the domain of perceptual psychology and
we shall discuss pictorial recognition at greater length later in the chapter.
What I wish to stress here is that recognition rather than illusion supplies us
with a perfectly adequate framework for characterizing the spectator's
apprehension of mimetic representations.
It might be countered that recognition is not as viable a concept as
illusion for discussing our responses to mimetic representations, for recog-
nition does not explain the intensity of our responses to such images
whereas illusion does. I have already challenged the notion that illusion is
helpful here. Furthermore, I think that by using recognition as our key
concept, we can accommodate the intense responses we have in regard to
mimetic representation. For if we discard the notion of illusion, we can
characterize our intense emotional, aesthetic, and intellectual responses to
mimetic representations, like films, by saying that we know, via recogni-
tion, that a given film represents x (or represents x as a so-and-so) some-
thing that wefindgripping, involving, exciting, boring, funny, anachronis-
tic, apt, insightful, eerie, engrossing, and so on.
If I feel uplifted watching a spectacle like Gandhi, this need not be
explained by claiming that I believe I am in India rubbing elbows with the
Mahatma. Rather I am uplifted because I recognize that Gandhi depicts the
kind of courageous life that I find admirable and exciting. Explaining the
intensity of my response in further detail would hinge on explaining why, if
I find such lives exciting, I do. Admittedly, recognizing what an array
depicts, does not, in and of itself, explain why I am moved by it. But a
recognition account, supplemented with sociological and psychological,
and perhaps even sociobiological, expositions of why the kinds of events
depicted are found moving, will explain the intensity of response.
Thus far, we have concentrated our objections on the notion that mi-
metic representations, in and of themselves, promote illusion effects, nota-
bly the illusion of reality. But it will be recalled that contemporary film
theorists, inspired by Brecht, not only claim that mimetic representations
are illusory but that they are also conducive to the formation of further
illusions, viz., reactionary beliefs. That is, mimetic representations abet the
audience's acceptance of the ideological falsehoods that are stated in the
plot or implied by the plot structure of a movie. Mimetic representations
accomplish this by putting spectators in a sort of trance which paralyzes
104 T H E CINEMATIC IMAGE

their critical faculties in respect to social falsehoods. Believing that what is

before us is reality, we putatively accept the ideological conception of what
we see as a matter of fact. Mimetic representation is rather like the immo-
bilizing sting of the spider. It stuns our powers of criticism—with illusions
of reality—consequendy inducing us to endorse propaganda. We see black
people represented as shiftless and we accept this as a fact about black
people in general. In terms of artistic production, Brecht advocated reflex-
ivity—i.e., calling attention, within the representation, to the fact that the
representation is a representation—as a means of short-circuiting this pro-
cess. For if the actor in a play announces he is an actor, the illusion of reality
is purportedly dispelled and, putatively, our critical faculties are again
thrown in gear.
The postulated connection between mimetic representations and the
demobilizing of the spectator's powers of criticism is a causal hypothesis.
As armchair psychology, it is reasonable. But when weighed against the
facts, it appears insupportable. Think of all the critical faculties that were
not paralyzed by an exercise in visual verisimilitude like Fort Appache: The
Bronx. Moreover, one must ask of those viewers who thought that the film
was truthful about communal life in the South Bronx whether it was the
mimetic cinematography that got to them rather than the fact that they
already believed certain myths about the unrelenting violence of black social
life. And wouldn't such racist viewers continue to believe such myths if the
film were shown out of focus or if the production had been a stick figure
cartoon.
The Birth of a Nation presents a similar case. Despite its pictorial ver-
isimilitude, blacks saw it as the vilification it was, while the whites who
embraced it were already racist. Likewise, that Red Dawn is a mimetic
representation in no way prevents me from seeing it as a piece of jingoistic
warmongering, nor does it persuade me that a Soviet invasion is imminent.
Nor does Rambo: First Blood Part II convince people that the Soviets and
Vietnamese are presendy electrocuting POWs on wire mattresses. It may in
fact reinforce this belief in those viewers who already accept the idea that
American POWs are currendy being tortured in Southeast Asia. But note,
it is not the mimetic representation, in this case, that dampens the viewer's
critical faculties; rather the viewer's critical faculties were in abeyance long
before the film.
Furthermore, the lack of verisimilitude does not correlate with the opera-
tion of the viewer's critical faculties. Jordon Belson's mystical abstractions
may evoke a sense of cosmic inevitability but without resorting to mimetic
representation. Nor does the practice of reflexivity—revealing, for exam-
T H E CINEMATIC IMAGE 105

pie, that the film is a film—guarantee that the production will be viewed
critically. Both Roma and Clowns are about their own process of production
and about their author, but Fellini's reflexivity only subserves the propaga-
tion of his world view, a kind of life-affirmative pluralism. Indeed, one
suspects that Fellini's intrusiveness in these films in fact enables him to get
away with his shameless exploitation of shopworn, universalist (clown as
man; city as life) imagery. In short, reflexivity does not result in heightened
critical awareness, nor does mimetic representation cause lowered critical
awareness.
One might attempt to support the view that visual "illusionism" impedes
criticism by arguing that such illusionism works by masking "real" unsee-
able forces by means of a phenomenal presence that implies that what we
see is "all there is." Such a move appears to underpin a good deal of
contemporary film theory's rejection of empiricism and phenomenology.
This might be called the "camouflage" notion of illusion; the phenomenal
detail in the image obscures insight into the actual laws of society. But it is
not clear to me that the camouflage theory fits well with other things
contemporary theorists wish to say. For in many examples of the illusion
effect, picturing a relation—say of a heterosexual couple—is said to frame
the pertinent relation as natural and universal. That is, the problem here is
not just that real social generalizations are obscured, but that alternative,
putatively false, social generalizations are promoted. But if false generaliza-
tions—e.g., heterosexuality is universal—are said to issue from the image,
then it is not the case that mimetic images restrict us to brute, visual
phenomena.
Of course, it is difficult to discuss the putative engendering of the
illusions the Brechtians have in mind solely with reference to mimetic
representation and without reference to narration. So, though narration is
properly the topic of the next chapter, we will conclude this section with
some anticipatory comments on it in order to round out the discussion of
cine-Brechtianism.
Contemporary film theorists see the illusionism of mimetic representa-
tion working in tandem with the illusionism of narrative in the production
of ideologically false beliefs. The suspect feature of narrative structure for
contemporary theorists is its compelling momentum. The narrative, with
its propulsive force and closure, makes us feel that what has been depicted
had to happen as it did in the story. Narrative has an aura of inevitability.
And this has ideological repercussions because it suggests that society
cannot be changed. Moreover, this effect is attributed to all realist represen-
tations that employ traditional narrative devices.
106 T H E C I N E M A T I C IMAGE

But surely this is open to easy counter-examples. Gorky's Mother and

Remarque's All Quiet on the Western Front imply that human relations—
and society—ought to be otherwise, while Typee attempts to show that
things could be otherwise. These works are classical, realist narratives no
less than a reactionary item such as Lord of the Flies. And turning to films,
works such as Rules of the Game, Kammeradschaft, and La Terra Trema,
though all classically narrated, suggest that society should be otherwise. In
brief, it is within the resources of narrative, realist representation to convev
the possibility of social change. Reactionary fatalism is not an invariant
effect of traditional, narrative representation. Indeed, recall that the theater
that upset Brecht, whom contemporary film theorists follow, was often
referred to as Aristotlian. And the tragedies that Aristotle used as his data
were, thematically speaking, fatalistic. But there is no reason to think that
independent of the details and themes of particular stories, that stories as
such are fatalistic, ideologically or otherwise.
In fact, identifying the closure and structure of expectations of realist
narration with a sense of ideologically pernicious inevitability is unsound.
The reasoning seems to be: realist narratives suggest inevitability; a sense of
inevitability promotes the illusion that society can not be changed; realist
narratives discourage social change.
The first step in this argument is a questionable generalization, if "inev-
itability" means that things could not be otherwise. Back to the Future
suggests that everything is open to change, while at the conclusion of The
Purple Rose of Cairo, one can hardly help thinking how different things
could have been if the heroine had chosen her celluloid lover. Moreover, it
seems that in the assertion that inevitability causes the illusion that society
cannot be changed, the meaning of "inevitability" has shifted. At first we
take "inevitability" to refer to the sense of closure and structure of expecta-
tions in a narrative; but, then, this reference is expanded to encompass the
causal structure of society. This is nothing but equivocation. My feeling of
the narrative aptness of the disappearance of the ghost and the marriage at
the end of Kiss Me Goodbye has nothing whatever to do with my beliefs
about the prospects for social change. 13

P H O T O G R A P H Y A N D T H E C I N E M A T I C IMAGE
Though film images need not aspire to verisimilitude, the cinematic
image characteristically is a mimetic representation, that is, a representation
that people are prone to think represents its referent by way of resembling
it. The purported illusionism of the cinematic image, which so vexes con-
temporary theorists, is a function of the photographic nature of film.
T H E C I N E M A T I C IMAGE 107

Again, photography need not be employed mimetically, though in cinema

it standardly is. And, of course, mimetic reproduction was the purpose for
which photography, and then cinematography, were developed. So there
is, at the very least, a strong historic bond between cinematography and
verisimilitude which, for contemporary film theorists, also involves a bond
with illusionism. Thus, the photographic nature of the cinematic image is
of key concern for contemporary film theorists who attempt to isolate the
ways in which the photographic aspect of the cinematic image contributes
to the production of ideology. In this section we will review some of their
arguments.
However, before turning directly to what contemporary film theorists
contend about photography, it is helpful to situate their position within the
context of the history of film theory. The photographic nature of film is an
issue around which the major movements of film theory have taken shape.
Indeed, probably the most comprehensive viewpoint with which one can
organize the history of film theory is to focus on the alternative accounts of
the role of photography in successive theoretical formulations. That is, the
contesting, attempted clarifications and alternative assessments of film's
relation to photography provides us with a continuous line of controversy
which makes what most would agree are the major movements in film
theory stand out in stark relief.
Broadly speaking, these movements are:
A. The Creationists, such as Arnheim, Balasz, the Soviet montagists, and
such later day proponents as Lindgren. Roughly, the creationists hold that
the appropriate role of film is to manipulate and to reconstitute reality,
overcoming, thereby, photography's inclination toward mechanically re-
cording it.
B. The Realists, such as Panofsky, Kracauer, most notably Bazin, and, in
some respects, Cavell. For the realist, the cinema, in various ways, should
record and reproduce reality in accordance with its photographic origin.
C. The Psychosetniotic Marxists who comprise the vast majority of practic-
ing film theorists today and to whom we have consistendy referred as
contemporary film theorists. This group is committed to isolating the
ideological resources available to cinema, and for many of them, photogra-
phy, insofar as it is said to impart the impression of reality, naturalness, or
transparency, is a major mechanism of ideological persuasion.
These three groups of theoreticians can be characterized contrastively by
what, on the surface, initially appears to be their radically different assess-
ments of the photographic element of cinema. These assessments, more-
over, emerge in the course of a neatly plottable, historical dialogue.
The creationists can be thought of as beginning the dialogue about
108 THE CINEMATIC IMAGE

photography in film theory. But the creationists themselves were reacting

to views about photography that were common in the nineteenth centurv.
These latter views centered on the idea that a photograph was merelv a
mechanical reproduction of reality. What was often at stake in these asser-
tions was whether or not a photograph could be art. By maintaining that a
photograph was merely a mechanical duplication o f reality, the enemy of
photography was denying that photographs could project the imaginative
and personal viewpoint of an artist. The photograph was not an expression
of a viewpoint but a piece of nature ground out by a machine. The
underlying contrasts in this debate were photography versus art, prosaic
imitation versus expression, the machine versus temperament, a sheerly
causal process versus intelligent action. Emile Zola displays this depreca-
tion of photography when he writes:

The individual element, man, is infinitely variable; as much as in his creation

as in his temperaments. If temperament had not existed, all paintings would
have of necessity to be simple photos. 14

This view continues into the twentieth century and is echoed by Croce
who claims:
if by imitation of nature be understood that art gives mechanical reproduc-
tion, more or less perfect duplicates of natural objects, in the presence of
which is renewed the same tumult caused by natural objects, then the propo-
sition (that art is the imitation of nature) is evidently false. . . . if photography
have in it anything artistic it transmits the intuition of the photographer, his
point of view, in the pose and grouping which he has striven to attain. And if
photography be not quite art, that is because the element of nature in it
remains more or less unconquered and ineradicable. Do we ever indeed feel
complete satisfaction before even the best of photographs? 15

The consensus of turn of the century aestheticians to Croce's rhetorical

question was " n o . " Photos were not art because they were merely mechan-
ical duplications of reality. Thus, when Clive Bell searches for a way in
which to cashier a painting from the order of art, he likens its achievement
to the slavish imitation available in "cinematographic pictures." 1 6
In order to show that film could be an art, and this was the first historic
task of film theory, creationists had to demonstrate that film could be more
than a mere mechanical recording of reality. However, the manner in which
they pursued this goal often involved them in sharing at least two assump-
tions with the opponents of film and photography, viz., (1) that if film was
to be an art, it would have to be more than merely imitative—it would have
to afford expressive or formal values over and above verisimilitude, and (2)
T H E C I N E M A T I C IMAGE 109

that, in some sense, it is feasible to think that there is a zero degree of visual
style, conceived of as the mere reproduction of reality, to which a film or a
photo might descend.
Creationists thought that film should not be used to mechanically re-
produce reality but to interpret it. But in regarding mechanical reproduc-
tion as a danger to be overcome, they, in effect, were agreeing that in some
respect it is plausible to think that in certain lamentable cases film and
photography might be truly describable as mere mechanical reproductions.
That is, in accepting mere mechanical reproduction as the pertinent foil in
the film-as-art argument, they acknowledged the feasibility of the idea that
film could be meaningfully said to mechanically reproduce reality.
The strategy that creationists adopted to demonstrate that film is an art
revolved around showing the ways in which the cinematic image diverged
from a perfect duplication of reality and, furthermore, how such diver-
gences could be employed for expressive purposes. High and low angle
shots were frequent examples in this regard. The distortions in scale that
resulted from such camera positions could be used expressively to signal,
alternatively, power or weakness, among other things. Editing, of course,
represented a major departure from a perfect recording of the space-time
continuum and received special attention from creationist film theorists,
especially the Soviet montagists.
Through editing and other devices, the creationists urged it was possible
for filmmakers to not slavishly imitate reality but to interpretively recon-
stitute it. The filmmaker need not merely record the world; he or she could
create worlds—worlds of works of art. The importance of the creative
potential of cinema was seen as standing in an inverse ratio to the impor-
tance of photography, i.e., the recording element in cinema. Montagists
were especially dismissive of the weight to be accorded photography in
cinematic representation, arguing instead that the significance of any shot
depended on its place in an edited array. But even outside of editing, the
potential for distortive yet expressive stylization within the shot was em-
phasized. The road to film art lay in working against the recording poten-
tials of photography for the sake, most often, of heightened expressivity.
Photography was the bête noire of the creationist. It raised the possibility
that following the least and worst line of resistance, cinema would degener-
ate into mere recording.
Of course, the very notion of a mere mechanical recording is a complex
and probably, ultimately, confused one. At the very least it superimposes
several not necessarily related ideas. Film is said to be mechanical because it
is machine-based. But this should have caused no problems for cinema's
110 T H E C I N E M A T I C IMAGE

prospects in regard to either expressivity or art. For surely a piano is a

machine, but that has never led anyone to deny that one can express oneself
bv means of it nor that art can be made by it. "Mechanical" obviously must
mean more than "machine-produced" in this argument. In fact, it appears
to have meant unimaginative, uninspired, inexpressive, banal, prosaic, etc.
And probably it was thought that insofar as a camera is a machine, it made
such results more likely. But there was never really any reason to think this
since, again, pianos, and even paint brushes, are machines of sorts, involved
in predictable physical processes once set in motion. That is, where "me-
chanical" really means "uninspired" we should blame the artist not the
machine. If photography was ever mechanical in this sense, the problem
was with the photographers and not the medium.
Moreover, if mere mechanical reproduction was taken to mean a perfect
(because automatic and unmediated) representation of reality, then it has
never been clear that there are such things; for there is no camera that is
fully automatic, no representation in any medium that is unmediated, and no
representation of reality that is "perfect" except in terms of the human
interests, uses, and subtending viewpoints the representation is intended to
serve. Creationists might have been better advised to reject the idea of "a
mere recording of reality" as absurd, rather than acknowledging it as an
albeit negative potential of cinema. However, they suffered the myth to
flourish, perhaps as a scarefying bogeyman, and the myth continues to be
given credence in various forms to the present day.
Creationist film theory correlates roughly with the period of silent film
and remained powerful into the late thirties. Moreover, it reflects the values
of the filmmakers of that period, who, in an effort to make their medium
into an art, emphasized overt stylization. This is evident not only in the
development of assertive styles of editing but in movements such as what is
called the German Expressionist cinema. Indeed, the sight gag perfected by
American comics can be seen as part and parcel of the anti-recording
animus of silent film since it is based on the principle of not presenting
reality as such but in terms of presenting an event from two simultaneous
but diverging viewpoints: the false perspective of the comic butt and the
accurate state of affairs as seen by the audience. Sight gags, that is, do not
re-present reality, but rather present conflicting views of reality. By the end
of the silent period, many of the assertive styles developed in the twenties
were melded into one grand international style in such films as Sunrise,
which unified the self-declaiming, overt artifice of German Expressionism
and its penchant for camera movement with American glamor photogra-
phy and sight gags, self-indulgent special effects, and editing strategies
derived from the Soviets.
T H E C I N E M A T I C IMAGE 111

With the emergence of sound, many of the pyrotechnics of silent film

receded, at first due in part to expediency but also gradually by choice. New
approaches to film exposition began to appear along with new tastes. In
short order, in the late thirties and early forties, a new approach to film
theory began to emerge as a way of better rationalizing certain tendencies in
cinema—such as the work of Renoir, and later of Welles,Wyler, and the
Italian Neo-Realists—than was available to creationist theory. This was
realist theory.
Whereas the recording aspect of photography was the aesthetic embar-
rassment of the creationists, the realists made it the fulcrum of their ap-
proach to film. The relation between creationists and realists is classically
dialectical. The creationists "no" to photography is met by a realist affirma-
tion. In terms of the creationist anxiety over the art-versus-photography
antithesis, the realist responded "too bad for art." For iffilmis an art, then it
must be greeted as an utterly unique one, not one that refers to reality by
proposing images that resemble it, but one that literally records reality, one
that literally re-presents it. That is, due to its photographic base, the
relation of a film image to its referent is one of identity, not resemblance—
the film image is identical to its referent due to the special causal processes
involved in photography. The film image does not just resemble its referent
after the fashion of a painting but is exacdy identical to its referent—it is a
precise tracing in light of something that actually existed before the camera
lens. The cinematic image is taken as a mechanical reproduction of reality
bv the realist and this is regarded positively.
Moreover, the realist takes this analysis of the cinematic image to imply
that due to the unique character of cinematic representation—best thought
of as cinematic re-presentation (presentation again)—cinematic styles that
approximate recording are to be endorsed in opposition to the assertive
stylizations of silent film. It is as if in a gesture of ingratitude—after
creationist theory won legitimacy for film—it was no longer necessary to
adhere to aesthetic standards predicated on expressive artifice. Instead, the
photographic-recording dimension of the medium could be indulged with-
out recrimination, the battle for film art having already been decisively
won. Various realist styles of filmmaking, employing long takes and mini-
mizing editing, were favored from the forties into the sixties. The realist
style of visual exposition was purportedly grounded in the belief that due to
its photographic component, cinema, ontologically, was an essentially real-
ist art, that is an art somehow distinctively rooted in reality.
This position was most powerfully stated by André Bazin. For him
cinematic representation was utterly unique. The cinematic image, due to
photography, docs not merely look like its referent, as one might say of a
112 T H E C I N E M A T I C IMAGE

drawing, but the photographic image is identical to its referent because

photography can be seen as a mold or a tracing literally generated by the
object standing before the camera. The image we see on the screen literally
re-presents reality because that is how the camera operates. If film is a
mechanical recording that is its strength—no other medium can literally
duplicate reality. Similar views about the nature of cinematic and pho-
tographic representation can be found in the writings of Cavell, Barthes,
and Sontag, while somewhat less extreme explications of the essentially
realist nature of cinematography are available in Panofsky, Kracauer, and
Perkins.
The strongest statement of the realist theory of film, one predicated on
the thesis that photographic representation is uniquely re-presentational, is
nothing short of logically incoherent. The theory states that each cinematic
image literally re-presents the object that caused it to be by being identical
to a mold or like a tracing to the object or person that gave rise to it. The
close-up of Gloria Swanson literally re-presents Gloria Swanson because in
terms of patterns of light it matches those rebounding off Gloria Swanson,
contour for contour, and Gloria Swanson was the object that caused the
photo to be. For any photo or cinematic shot, there exists an object that it
matches as a tracing. But this theory quickly leads us into insufferable
perplexity.
Imagine that we take three shots of the self-same person, assuring that
the image of the person is the same screen size in each of our shots.
However, also imagine that we take our respective shots with three dif-
ferent lenses—a wide-angle lens, a telephoto lens, and a normal lens. Now
the theory of re-presentation tells us that each of these shots will be
identical to a mold with its subject. Furthermore, identity is a logically
transitive relation—if x is identical to y, and y is identical to z, then x is
identical to z. However, though each of our shots is putatively identical to
its subject, even the untrained eye can see that our shots are not identical to
each other. Thus, the re-presentational theory of the unique identity rela-
tion of photographic images to their referents, when added to some easily
established facts about lenses, leads to a contradiction.
Nor will it help matters to argue that the theory only pertains to what are
called normal lenses. For that has changed over history in a way that will
allow us to reproduce our counter-example with three candidates for the
tide of normal lens from three different periods of film history. Moreover, it
will not strengthen the theory to say that what is re-presented is not a
precise tracing of light but an imprint of light. For a flash pan records and
re-presents an imprint of light but one that need not be recognizable and,
therefore, need not be representational (mimetically, that is). But a realist
T H E CINEMATIC IMAGE 113

theory of film, where the images in question are not representational, is

clearly the antithesis of realism a la Bazin, however much it may gladden the
hearts of modernists.
Most contemporary film theorists reject the realist theory of film. They
do not do so on the grounds of the logical considerations just sketched but
more on the grounds that, contra-realism, it is difficult to conceive of a film
image as an automatic re-presentation of reality in any persuasively charac-
terizable sense of the word "automatic." A camera is generally adjusted by
an agent with some purpose in mind and that purpose guides the character-
ization we encounter of the referent of the image. The camera does not
simply duplicate reality; it is not a mere mechanical recording of reality in
any sense that suggests its results lie outside human practices and purposes.
We are not faced by chunks of unmediated reality when we gaze upon
photographic images but intentional objects. That is, where the re-
alist theorist speaks as though the cinematic image were some sort of
natural object, the contemporary theorist would hold that the cinematic
image is always the product of a signifying practice. Thus, most contempo-
rary film theorists reject the most extreme ontological claims of realist
theorists. The psychosemiotic Marxist maintains that film is not an auto-
matic or mere mechanical reproduction of reality. Indeed, insofar as the
contemporary film theorist holds that film is necessarily a signifying prac-
tice, she disagrees with everyone—both of the creationist and realist
camp—who thinks the mere recording of reality to be a possibility.
However, here the plot thickens. For though most contemporary film
theoreticians reject the realist ontology of the cinematic image, many psy-
chosemiotic Marxists accept what can be called the realist phenomenology
of the cinematic image. That is, contemporary theorists reject the idea that a
film image pristinely re-presents reality—after all it is part of a process of
signification. But, at the same time, they believe thatfilmsgive the impres-
sion of being re-presentations of reality to spectators. The realist theory is
ontologically incorrect but nevertheless it does, for certain contemporary
film theorists, afford an account of how spectators regardfilmimages. That
is, film images do not really duplicate reality, however they do give the
impression that they duplicate reality. Photography is, of course, crucial
here. And furthermore it is this impression that is key to an understanding
of the ideological machinations of film.
For the impression of reality imparted by the film image is what is
conducive to the ideological persuasiveness of the film. That is, a pho-
tographically mimetic characterization in a film will, according to many
contemporary film theorists, promote the impression in spectators that
what is portrayed is the way the world is, becausefilmimparts the impression
114 T H E CINEMATIC IMAGE

of re presenting reality. The film in all its ideological extravagances is

accepted by spectators as an unmediated reproduction of reality. Put crude-
ly, psvchosemiotic Marxists reject Bazin's theory of the ontology of film for
themselves but, so to speak, attribute a belief in it to ordinary spectators.
Whereas Bazinians applauded the realistic effect of cinema, contempo-
rary film theorists find this phenomenon and its exploitation for the pur-
poses of ideology morally abhorrent, and they dedicate themselves to
disabusing the world of the myth of cinematic realism. Yet, they do agree
with people like Bazin that film causes this reality impression in spectators.
Where psychosemiotic Marxists diverge from Bazin in terms of cinema's
impression of reality is in their evaluation of its moral worth. Thus there is
much agreement as well as disagreement within the history of film theorv.
Creationists accept the possibility that film may merely mechanically re-
produce reality but find this aesthetically abhorrent. Realists, in turn,
respond by finding the mechanical duplication of reality aesthetically and
morally commendatory. Psychosemioticians come at the debate obliquely,
stressing, like creationists, the intentionalist nature of photography, but
hold that this is characteristically masked in such a way that viewers in their
naive state erroneously endorse something like the realist re-presentation
theory of photography. Contemporary theorists agree with the realists,
then, that cinema imparts an impression of reality, but they find in this the
shame rather than the glory of film insofar as it is this supposed impression
of reality that facilitates ideology.
Where the realist claimed that photography re-presents reality, the con-
temporary film theorist psychologizes that ontological claim by asserting
that the photograph and the cinematic image cause the impression of reality
(here understood not in Baudr/s sense of charged reality but in the illu-
sionistic sense). This, of course, is a claim that requires some explanation
since, among other things, normal percipients in normal viewing condi-
tions appear to have no problem distinguishing films from their referents.
Thus, the contemporary theorists must explain the origin of this pho-
tographic impression of reality on the way to explaining the way in which
once this impression is in place it reinforces the ideological import of such
cinematic images by implying that what is photographically represented
claims for itself the authority to assert "And that's the way it is."
Recent attempts at offering such explanations are to be found in Ideology
and the Image by Bill Nichols and Visible Fictions by John Ellis.17 It is to
those accounts that we now turn.
Nichols' approach relies on buying into a certain view of the psychology
of perception. He holds that the mind deals with its sensory input by means
of organizing it in accordance with a code which then permits the input to
THE CINEMATIC IMAGE 115

be translated by further processing operations. Moreover, Nichols believes

that the code that is operative in perception is an iconic c o d e — o n e in which
the relation o f the signs in this code to their referent in the world is one o f
resemblance or analogy. T h e c o d e in question is rather like an internal
picture. Nichols says:

Our perception of the physical world is also based on codes involving iconic
signs. The world does not enter our mind nor does it deposit a picture of itself
there spontaneously. Perception depends on coding the world into iconic
signs that can represent it within our mind. The force of the apparent identity
is enormous, however. We think that it is the world itself we see in our
ordinary "mind's eye" rather than a coded picture of it. (p. 11)

M o r e o v e r , these codes are essentially cultural in nature.

The variability of the everyday world becomes translated by reference to less

variable codes. The environment becomes a text to be read like any other text.
T o not know the codes maintained by a given culture is tantamount to being
an illiterate infant wandering through an unintelligible world, (p. 26)

Perception relies u p o n cultural codes that are iconic and that are em-
ployed unconsciously. W h a t has this theory o f perception t o d o with film?
Well, film, via p h o t o g r a p h y , is like perception in that it relies o n iconic
codes. Nichols says, " I t is n o t reality u p there on the silver screen but iconic
signs that re-present reality" (p. 11). A n d since cinematic images operate by
way o f iconic i m a g e s — j u s t as perception is said t o — N i c h o l s believes that it
is possible to c o n f u s e o n e for the other. H e writes:

Since images bear an analogous or iconic relationship to their referent (a

relationship of resemblance), it is easy to confuse the realms of the image and
the physical world by treating the image as a transparent window (especially
the photographic image), or by treating the physical world idealistically by
assuming that something like its essence has been transferred or reproduced
in the image. Many films employing realist styles encourage such a confusion,
and yet it is essential to remember that a film is not reality any more than an
image is what it re-presents, (p. 24)

Indeed,

We might even say, metaphorically, that realist images are an objectification,

or projection o f the normal perceptual process. What our nervous system
initially encountered as unorganized sensory input is now encountered as the
organized or signifying output of these objectification, or images, (p. 24)

S o the impression o f reality is caused by photography because photogra-

phy works in iconic codes, not only reproducing our cultural way o f seeing
116 T H E C I N E M A T I C IMAGE

but presenting it as natural. This, in turn, not only reinforces cultural wavs
of seeing, but the impression of reality of the cinematic image effectively
imputes some veracity to whatever it shows.
There are a surprising number of things wrong with Nichols' account,
but luckily there is a short way to dispose of his theory. Nichols holds that
perception depends on coding the world into iconic signs that can be re-
presented within our mind. This coding is reinforced by cinema which also
exploits it for ideological purposes. But what of the underlying notion of
perception—will it wash?
Nichols' account of perception suggests a three stage process: first, there
is the object in the world which, second, is transformed into an iconically
coded sensum which, third, the mind translates. But this seems radically
implausible.
The sensum and the referent look alike. Therefore, if the mind can
translate the sensum why can't it process the object directly without the
intervention of the sensum? That is, why postulate the iconic step? But, of
course, if Nichols loses the iconic step, his theory of the cinematic image
disappears. On the other hand, if Nichols responds that the mind can't
process the object directly, then we can ask how it can process the iconically
coded sensum, since the sensum looks like the referent. It seems, then, that
there will have to be another coding phase postulated in order to decipher
the second stage sensum. But if that third stage code is iconic, it too will
resemble the object which we are unable to perceive without iconic coding.
So a third stage, iconically coded sensum will require a fourth stage coded
sensum for translation, and so on.
One may say that these intervening coded sensums are not iconic, and
that this will stop the regress. But that leads us to ask why the possibility of
non-iconic processing was not considered at stage two of Nichols' scenario.
In short, it seems that Nichols' theory of perception founders because it will
commit us to an infinite regress of iconic sensums, or it will have to admit
the possibility of non-iconic processing or even of direct perception. But
either of these alternatives will deprive Nichols of the cornerstone of his
theory of the ideological operation of the cinematic image.
Another way to make this point is to note that Nichols' theory of
perception, like so many of the psychological models propounded by
contemporary film theorists, is an homunculus theory. The mind that
translates the iconically coded sensum is just like the person perceiving the
object at one remove. This must be the case, since the sensum looks like the
object. But nothing of theoretical value is gained by postulating this "in-
ner" perceiver; for if there is a problem explaining how the person perceives
the object, all those problems will be, so to speak, brought inside, when it
T H E CINEMATIC IMAGE 117

comcs to the internal perceiver's perception of the iconically coded sensum,

since it looks just like the object. Indeed, as I have suggested, it seems likely
that this way of characterizing perception will land us with an infinite
number of homunculi, each requiring its own inner perceiver to translate
the preceding, iconic sensum. And, of course, if Nichols theory of percep-
tion is wrong, then his theory of both the impression of reality imparted by
the photographic element of film and the consequent theory of film's
ideological operation fail.
Like Nichols, John Ellis attempts to account for the impression of reality
in cinema by characterizing the film situation as analogous to our normal
practice of perception in such a way that we easily misrecognize our
perception of film for a perception of reality. And this, in turn, leads us to
accept what is on the screen as "the way things are." He writes:
The ideology of vision that we inhabit is one that thinks of the eyes as
projecting a beam of light, like a torch-beam, that illuminates what we look
at, making it visible and perceptible. The arrangement of projection in the
entertainment cinema that developed in the West exacdy parallels this ideol-
ogy of vision: it too presents a beam of light, coming from a source behind
the spectator's head, which widens out to illuminate a scene for perception.
The beam of light from the projector parallels the beam of light from the eyes.
On this basis, it can be said that the first identification that takes place in
cinema is with the apparatus of projection. A profound homology is experi-
enced between the cinematic spectacle and the activity of perception of it.
They are identified, rather than perceived as complementary activities, or
opposed forms of activity.18

Here Ellis is speaking of what Baudry called the apparatus rather than the
photographic image per se. However, like Nichols, he thinks that the
projection of the cinematic image affords a probable confusion of screen
perception and natural perception. This argument depends on postulating
what might be thought of as a searchlight theory of vision in which beams
of illumination shoot out of our eyes after the fashion of the first posters
advertising Universale Frankenstein in the thirties. But one wonders
whether this conception of perception is really widespread. It is true that
Arabic astronomer A1 Hazen criticized something like the views Ellis at-
tributes to all of us in his eleventh-century treatise The Elements of Optics,
but the proponents of those theories are long dead and their ideas rather
recherche by now. Nor does it seem appropriate to attribute arcane philo-
sophical theories to contemporary spectators. We do not confuse the pro-
jector beam with our "perceptual beams" because we do not think we have
"perceptual beams."
What Ellis and Nichols are both trying to supply is a mechanism that
118 T H E C I N E M A T I C IMAGE

accounts for why we do not notice the difference between a cinemato-

graphic projection and perception. Both attempt this by postulating analo-
gies between photographic processes and some imputed dimension of
perception. The match is then said to be powerful enough to block us from
noticing the difference. However, the mechanisms Nichols and Ellis offer
are absurd, though absurd in different ways.
In defense of what Nichols and Ellis are attempting, it may be said that
they want to call attention to the fact that, in watching cinematic images,
we do not characteristically dwell on the fact that they are photographic.
Though such a recognition may always be available to us, we choose, for
the most part, to bracket it from the forefront of our attention. We attend
to what such images represent and not to the fact that they are representa-
tions. This type of overlooking might be described as regarding the image
as natural, a key term for contemporary film theorists. And that pho-
tographic images are perceived as natural—it might be claimed—is unde-
niable, even if Nichols and Ellis have misdescribed the psychic mechanisms
that induce this.
But I wonder if it is appropriate to describe ourselves as regarding such
images as natural. Suppose we look at a photo of a black soldier saluting a
French flag. We do not regard this as natural in any literal sense—i.e., as a
product of nature divorced from culture. If what is meant by saying that we
regard the photo as natural is that it, like millions of advertisements, does
not strike us as a bizarre juxtaposition of elements, then that is a special
sense of "natural." It might be better to say that we are familiar with such
photos and juxtapositions, and even complacent about their meanings.
But if it is complacency that we are talking about, then the explanation
for that is probably not located in photography, for we are also complacent
about non-photographic images. Instead, we should attempt to explain our
complacency in terms of how we came to have that enduring disposition
which undoubtedly involves an account of our participation in a culture
which conditions us to respond to the massive bombardment of images in a
complacent way. That is, there is no special photo effect in our example;
rather there is a type of conditioning that is relevant not only to the visual
artifacts of our culture. We happen to be complacent about that with which
we are familiar. This accounts for our failure to batten upon both ideologi-
cal messages in images as well as various other details of everyday life.
Complacency is not peculiarly engendered by photography. Rather fa-
miliarity is the major lever. But if the contemporary film theorist attempts
to explicate the naturalization effect in terms of familiarity, the brief against
photography will be displaced. For most contemporary film theorists are
T H E C I N E M A T I C IMAGE 119

conventionalists as regards representation. Thus, they can imagine repre-

sentational systems radically different from the mimetic picture—systems,
for example, in which something that looked like large Pollock drip paint-
ings supplied the vocabulary of portraiture. We might become very familiar
with this type o f representation and with familiarity would come natural-
ization. So understood in terms of familiarity, naturalization can have no
peculiar connection with photography. Indeed, one wonders why, given
their conventionalism, contemporary film theorists want to stress a priv-
ileged role for photography.
However, contemporary film theorists seem bent on finding a special
naturalizing effect that is special to the cinematic image. Ellis' primary
proposal is an example that can be thought of as essentialism psychol-
ogized. Beginning with an analysis o f the impression o f photography much
like Bazin's, Ellis goes on to suggest that this accounts for film's naturaliz-
ing impression. He writes:

From the point of view of the spectator (the point of view for which cinema
exists) the cinematic illusion is a very particular one: it is the illusion of
something that has passed, which probably no longer exists. The cinema
image is marked by a particular half-magic feat in that it makes present
something that is absent. The moment shown on the screen is passed and
gone when it is called back into being as illusion. The figures and places
shown are not present in the same space as the viewer. The cinema makes
present the absent: this is the irreducible separation that cinema maintains
(and attempts to abolish), the fact that objects and people are conjured up yet
not known to be present. Cinema is present absence: it says "This is was."
(pp. 5 8 - 5 9 )

Ellis continues by explaining that

the present absence of the cinematic image and sound enables cinema to
adopt a particular mode of narration which can be called "historic."
The historic mode of narration involves the use of forms that do not
explicitly acknowledge the presence of the viewer.
Such forms as direct address (where a character speaks direcdy "out" of the
screen as "I" to a "you," the audience) are extremely rare in cinema. . . .
Direct address makes explicit the relationship between the viewer and the
subject of the look; the historic mode does the opposite. Events take place as
though they came from nowhere. Events told in the historic mode of narra-
tion are told by no one, they have no origin, no motivating intelligence.
"Who is telling this story?" is not a question that is addressed in a classic
entertainment film. . . . Instead, the story unfolds: it seems as though reality
itself is telling itself, almost unaware that it is being watched. This charac-
120 T H E CINEMATIC IMAGE

tenses the historic mode of narration: reality unfolds itself. . . . The effect of
historic narration depends upon the photo effect. It depends on the sense that
the photo effect brings of the completeness of the actions before the projec-
tion of the film begins. The film is always already complete, a record of
something that predates the projection. The historic mode depends upon the
fact that it has a story to tell, a story which is completed at the outset, yet
unrolls as though it were in the present. Present absence thus creates a sense
of the presence, yet self-containedness of the cinematic fiction, taking place in
its spaces and times, in a separation from the viewer who nevertheless can
experience the spurious presence of those events and people.
The historic mode, characterised by its sense of reality narrating itself, can
thus be seen to rely upon the photo effect. The historic mode is one of the
lynchpins of the notion of cinema-as-reality because it enables discussion of
cinematic events to take place without any acknowledgment of their status as
cinematic fictions. . . . Cinema, in the historic mode of narration, can be
discussed as though it were events and characters alone. This is the product of
one theory of realism, the theory of the transparency of cinema as a window
on the world. However, the historic mode of narration is not this simple. It is
not an absence of narration, but the appearance of the absence of narration. It
crucially depends upon the photo effect, that is, on the absence of the events
represented as much as on the illusion of their presence, (pp. 59-61)

In this passage, Ellis claims that cinematography reinforces the trans-

parency effect of the sort of narrative characteristic of entertainment cinema
which is called "historic narration." In order to assess Ellis' claim, then, we
must have some notion of what this transparency effect is supposed to
involve. The discussion of this transparency effect is generally introduced
by a discussion of the linguist Benveniste's distinction between two modes
of expression which are referred to, in the literature, as discourse and as
history (or, sometimes, simply as "story"). 1 9 In speech or narration that is
dubbed "history" or "story" or "historic narration," there is no mention of
" y o u " or " I . " "After May '68, a mighty film theory arose" is an example of
history or historic narration. In discourse, on the other hand, reference to a
speaker is included in the utterance: "Verily I say unto you, after May '68 a
mighty film theory arose" is discourse. Narration in the mode of history is
said to impart the impression that it is authorless, that it is an instance of
events or "reality" appearing, deceptively by all accounts, to narrate them-
selves. Realist representation, including entertainment films, are supposed
to exemplify the historic mode of narration.
Why is this transparency effect ideologically significant to contemporary
theorists? They believe that the lack of reference to a speaker can be thought
of as the suppression of the acknowledgment of the representational appa-
THE CINEMATIC IMAGE 121

ratus and of the situation of utterance (the production of the representa-

tion) which includes the suppression of reference to or of the acknowledg-
ment of the interests that motivate the utterance (or representation). The
historic mode of narration is ideological because it imparts the impression
that the representation is authorless which implies that it results from no
interests whatsoever, ideological or otherwise, or that the production, since
it appears authorless, lacks any variety of point of view.20 Historic narra-
tion, that is, masks its motivations, a very serviceable gambit for disseminat-
ing ideology.
Though Benveniste introduced the history/discourse distinction with
reference to language, contemporary film theorists like Ellis and Metz are
willing to extend it to film.21 Ellis also holds that the photographic aspect of
cinema enhances the effectiveness of historic narration by giving the movie
the aura of presenting a past, and therefore a completed action, thereby
enhancing the apparent self-sufficiency of the narrative and its apparent
disconnection from a present production of meaning. In literature, the
historic mode of narration corresponds to the use of the "past definite"
tense; in film, the pastness, i.e., the re-presentational nature, of the pho-
tographic image takes up this function. Metz says of the function of tense in
historic narration that "the narrative plenitude and transparency of this
kind of film is based on a refusal to admit that anything is lacking . . . ,"
which when connected to Ellis' discussion suggests that photography calls
into action the operation of the Imaginary in order to yield the appearance
of reality narrating itself, sans author, sans motive, sans everything.
It is hard to know where to begin listing the problems with this account;
let us start with the idea of transparency, understood as the appearance of
reality narrating itself. Can it really be the case that by adopting the historic
mode of narration, a realist representation, or text, brings its audience to
believe that the representation is none other than events or "reality" narrat-
ing itself? If the realist text in question is a novel, one wonders why the title
page is not enough to establish that the book has an author? And in theater,
there are program notes, on paintings there are signatures, and in movies,
of course, there are credit sequences at the beginning and end of the film.
Moreover, if we arc to extend the transparency effect to advertisements,
then there are the featured brand names that make the interests motivating
the representation quite apparent. These devices certainly should dissuade
audiences of normal intelligence from taking these items to hail from
nowhere.
Indeed, what would it take for a realist representation to have the effect
of appearing as an event narrating itself? The very notion sounds peculiar,
122 T H E CINEMATIC IMAGE

even mystical. It is even more esoteric than a representation appearing, via

illusion, as a "real event." At least we have a grasp of what is meant by "real
event," and can suggest what it would take to feign the appearance of one.
But what is involved in "reality narrating itself," over and above the illusion
of reality? A close encounter of the third kind with a Hegelian Zeitgeist?
And what would that be like?
Is the notion of an authorless narrative even comprehensible? If a realist
representation is intelligible, we irresistably infer that some sentient agency
executed it. A literally authorless narrative would rend our conceptual
scheme. If, when reading a novel, we were suddenly taken with the convic-
tion—albeit obscure to define—that the events that comprise the book
were narrating themselves (!), we would be as dumbfounded as we would
be by the sight of the Pentagon levitating and turning orange as the result
of Yippie chanting. The notion of an intelligible narrative as sui generis
violates our sense of natural order. Nor do we have any idea of what it
would take for an intelligible novel or film to override this natural prejudice
(if indeed it is merely a "prejudice"). Certainly omitting references to an
authorial "I," and especially a fictional "I," will not turn the trick. Instilling
the impression that a novel is authorless, I submit, is stricdy speaking just
impossible. Normal audiences can not be deluded into believing that novels
and films are authorless, nor do I understand what plausible interests the
creators of novels and films could possibly have in convincing us that their
works are authorless. Moreover, if contemporary theorists complain that I
have taken literally what is meant metaphorically, then I must point out that
this metaphor is spoken of as if it described a state that has certain effects—
in promoting ideological acquiescence—whose sequence makes no sense if
"authorlessness" is, in fact, only a metaphor.
But perhaps it will be felt that I am too quick to reject the notion of an
authorless representation. What if, in an act of statistical wonderment, the
waves at Jones Beach threw up an assortment of sea shells whose arrange-
ment—astoundingly—"spelt out" a replica of Les Misérables, or better yet, a
hitherto unwritten short story. Would not that be an authorless narrative
and could that not serve as the model of authorless narration: The answer, I
take it, to the former query, is no; for if we knew the way in which the
arrangement of sea shells came about, we would, I think, just deny that it is
a representation—so deep is our conviction that representations and narra-
tives have authors. Thus, the likely answer to our second query is that these
shells are not plausible models for authorless narration, for thev do not
compose a narrative at all.
Ellis appears to offer us behavioral criteria for authorless narration, viz..
T H E CINEMATIC IMAGE 123

when people speak of the events in the film as if they were a sequence of
events rather than a sequence of representations. But then if this is a
criterion of authorless narration, then such narration does not appear to be
a consequence of the historic mode. For people describe the events in
Murder My Sweet as they report a work in the historic mode even though
Marlowe's discourse is included on the sound track.
Contemporary theorists use the distinction between history and dis-
course as a means of identifying realist texts across media, including film.
Realism is associated with the historic mode of narration, and this, of
course, is why it is ideologically dangerous. For with the historical mode
comes the transparency effect. But it does seem rather mistaken to believe
that the corpus of realist representation, even conceived in the way the
contemporary do, can be assimilated to the historic mode of narration.
Aren't Goethe's Young Werther, Smollet's Roderick Random, epistolary
novels and so on, in general, full of discourse and if we extend the his-
tory/discourse distinction beyond literature, aren't Rembrandt's self-por-
traits and Frank Perdue's chicken advertisements discourse too? And Carlos
Fuentes' Aura, a rather predictable and standard ghost story, incessantly
addresses us in the second person.
Of course, many traditional films—e.g., I Remember Mama, DC)A, Dou-
ble Indemnity, How Green Was My Valley, Sunset Boulevard—have narrators
who refer to themselves, while the original Invasion of the Body Snatcbers
("You're next!") and the Creation ofthe Humanoids address the audience in
the second person. Are such films to be considered nonstandard—tradi-
tional films that are discourse rather than historic narration? But surely these
are standard films, and their use of the first and second person is perfectly
within the norms of Hollywood-style filmmaking. So if Hollywood film-
making is supposed to be realist according to contemporary theorists, then
it cannot be the case that realism is characterized by historic narration.
It may be argued that these counterexamples miss their mark because in
most of the preceding cases what is referred to are fictional narrators, not
actual authors. But I am not sure that this distinction is relevant to the
history/discourse distinction as Benveniste draws it. Nor is it clear why, if
reference to the actual author is key, that such reference is only deemed
important through the deployment of first person pronouns. For surely
works of classic realism may indeed contain references to the actual author,
and, for that matter, to the audience of the representation. Thackeray refers
to himself jokingly in Vanity Fair, Fielding acknowledges himself and the
reader in Tom Jones (Bk I, chap. 4), while many televised, used car salesmen,
if we are to take their ads as examples of realist representations, include
124 T H E C I N E M A T I C IMAGE

actual dealers addressing us ("You can't beat this anywhere!") unabashedly.

And in film, it is hardly nonstandard in a travelogue or a documentan,' to be
addressed by an author directly, or by someone reading the author's lines.
The notion of authorless transparency as well as attendent claims about its
correlation with the historic mode of discourse just do not fit the data
contemporary theorists wish to describe.
Nor is it clear why discourse, the mere inclusion of first and second
person reference, enjoins a reminder of the author's role, or why lack of
such references cause the impression that the work is authorless. How
would the lack of the use of such reference overcome our presumption that
someone, generally someone whose name is available to us, contrived the
narrative. Of course, even greater difficulties arise when the putatively
linguistic distinction between history and discourse is extended to nonver-
bal aspects of visual media such as painting and film. For surely in these
cases the symbol system in question lacks a system for demarcating person
(as well as tense indexes apart from verbal ones). And, of course, most
mysterious of all is the notion that somehow inclusion or exclusion of first
and second person reference, and of whatever their supposed pictorial
equivalents are taken to be, would result, respectively, in the awareness or
unawareness of the motives and interests, especially the ideological ones,
that underlie a given representation. Of course, it is true that most films that
have ideological commitments do not come out and say "The ideological
commitments of this film are thus and so." But if this is the rather obvious
and well-known point that contemporary film theorists wish to make, they
have no reason to advert to the purported linguistic distinction between
discourse and historic narration. That ideologues do not identify them-
selves has nothing to do with abstract linguistic categories.
Ellis claims that the cinematic image, in virtue of its photographic dimen-
sion, underpins the historic narration of the popular, realist film. Perhaps in
this he feels that the imputed "pastness" of the cinematic image serves a role
comparable to that which Metz, following Benveniste, feels the past defi-
nite tense plays in historic narration in literature. However, it must be
stressed that film does not, strictly speaking, have a system of tenses. Nor
does it make any sense to say that film is always in the past tense (or, as is
often also claimed, that film is in the present tense). For the phenomenon of
tense is only comprehensible within a system, and a language or media
putatively possessing only one tense would not be a system. That is, the
claim that film is always in the past tense is as good as the claim that film is
always in the present tense, which is to say, it is not very good at all.
Attempting to identify film with a single tense is tantamount to admitting
that the category of tense is inapplicable to film.
T H E CINEMATIC IMAGE 125

Film is often said to be an art that manipulates time. So perhaps we

should be willing to metaphorically extend the notion of tense to film.
Flashforwards are future tense;flashbacksare past tense. Two things must
be noted about such a metaphorical expansion. First, that these temporal
articulations are not grounded in film, i.e., the cinematic image, but in the
dialogue and the narrative. And, second, if this metaphoric expansion is
permitted, it will be the wreck of the theory that film is always in the past
tense (or always in the present or future tense). For if we grant that film
avails itself of a system of tenses through dialogue and narrative, then we
are committed to film, or at least specific films, having as many tenses as
language has. That is, a film will have as many "tenses," or temporal
articulations, as the dialogue and narrative develop. Thus, it will not be
solely in the past tense. And if individualfilmsare said to have more tenses
than the past tense, it makes no sense to say that film, or even popular film,
is always in the past tense.
Of course, Ellis might wish to defend his hypothesis about the cinematic
image without reference to tense. He might say, instead, that all films, in
virtue of their photographic elements, impart an aura of pastness. A cine-
matic image, as such, enjoins a sense of pastness. This is virtually to
attribute Bazin's theory of the photographic image, as a phenomenological
description, to everyfilmspectator. Since Bazin's theory is insupportable, I
am loathe to attribute it to spectators. Ordinary viewers usually have better
sense about such things than film theorists.
I suppose that it is true that most spectators, being aware in some rough
and ready way about howfilmsare made, know that thefilmbeing projected
for them was completed in the past, though, of course, it needs to be said
that someone reading a printed novel also knows that the book was written
prior to the reading. But, despite this knowledge, I am not sure that film
viewers have the strong association with pastness, that Ellis attributes to
them, with every photo and every cinematic image. Though I know that
veal is made from young calves, this is not in the forefront of my attention
when I eat it. But in Ellis' case, one would assume that the sense of pastness,
that he associates with photography, would have to be somewhat effective
on the spectator, if it is to reinforce the historic mode of narration.
Crucial to Ellis' argument is that every photo is a record. But even if it is
true that, in principle (what about trick shots?), every photo could be used
as a record of the state of affairs that gave rise to it, not every photo or
cinematic image is used as a record. The photo of me standing next to a
cardboard likeness of Reagan is used as a joke, not as a record of my
afternoon on Times Square. Similarly, filmed fictions, and not just ones
with trick shots, are not primarily records, and anyone who views them that
126 T H E C I N E M A T I C IMAGE

way cither has got things wrong or has some pretty special interests (like
figuring out whether so and so was a bit player in such and such a film). And
where films and photos are not being used as records, there seems little
reason to presume that attentive spectators are struck by their "pastness."
Needless to say, a particular film or image may exploit the association of
photography with pastness, as was popular in Westerns, like Butch Cassidy
and the Sundance Kid (e.g., the last image), a decade or so ago. But, in such
cases, it is the internal structure of the film that mobilizes a sense of
pastness. And, of course, most films do not emphasize the pastness of their
images; the cinematography in 2001, after the prologue, does not empha-
size pastness but futurity. That is, even if pastness is one of our cultural
associations with photos (though it remains less clear that pastness is a
cultural association with movies, save, today, possibly silent ones), it re-
mains for the individual work to take advantage of this association. Most
films do not, and to that extent I am suspicious of Ellis' claim that an
invariant sense of pastness accompanies every cinematic image in a way that
bolsters the transparency effect of historic narration (if there is such a thing
in film).
Ellis also associates a sense of completeness to the cinematographic
image. This completeness purportedly contributes to the self-contained-
ness of the film, which sort of completeness some contemporary film
theorists might want to connect with the engagement of the Lacanian
Imaginary by film. However, clearly several different, fallaciously equiv-
ocating, senses of "completeness" would have to be in motion for this
argument to slide by. The cinematic image, if it is said to be complete, is
complete in the sense that the moment, or string of moments, photo-
graphed are over or finished. On the other hand, the kind of "complete-
ness" that is relevant to the impression of "self-containedness" afilmaffords
pertains to whether or not the narrative has closure. The two senses of
"completeness" are quite distinct, as one can see by stringing cinematic
images of randomly selected events together. First, we see a glimpse of a
1911 soccer match, then a close-up of a dirty pot, then a shot from Keys of
the Kingdom, followed by one from a Danish home movie, and so on
randomly. Everything shown will be "finished," but will not contribute an
aura of "self-containedness" to the melange. Individual photos and shots
themselves are not necessarily complete in the sense of "self-contained."
Also one can have a shot of an incomplete action such as my knocking over
my glass when I reach for a drink. To the extent that Ellis' theory of the
cinematic image links completeness in the sense that x is finished with
completeness in the sense that x is "self-contained" or has closure, the
THE CINEMATIC IMAGE 127

theory is a pun. Moreover, anyone who wishes to extend the theory and
claim that the completeness Ellis associates with the cinematic image to the
tvpe of completeness that would affect the Lacanian Imaginary has merely
compounded the hijinks, for an image that portrays something that is over
need not portray anything that is whole in any sense; recall the shot of the
bridge breaking apart in Bruce Conner's A Movie.22

PERSPECTIVE
Linear perspective is built into the motion picture camera so that when
the latter is standardly used the result is a cinematic image perspectivally
arranged. Linear perspective is based on the insight that parallel lines,
receding into depth, project onto the human eye as convergent lines. So in
"central" or "Albertian" perspective, parallel lines verge on a centralized
vanishing point. When a picture of x follows the rules of perspective, it is
said that the picture delivers a sheaf of light rays to an appropriately
positioned viewer which is in pertinent respects, regarding orthogonals,
isomorphic to the visual array that x would deliver to the fixed monocular
station point of a viewer of x in the world. Linear perspective was devel-
oped intensively during the Renaissance and it is grounded in an optical
theory of vision.
Perspective is a pictorial cue which engenders a strong impression (rather
than illusion) of pictorial depth. It is not the only factor in a picture that
affords a sense of depth; others include the comparative size of the objects
portrayed, texture-density gradients, edge and overlap phenomena, the
portrayal of shadows, the tendency for objects to lose detail at a distance, as
well as the tendency of distant objects to turn bluish due to atmospheric
impurities. However, though perspective is not the only means of convey-
ing pictorial depth, it is a particularly powerful means. Moreover, perspec-
tive affords a high degree of pictorial fidelity. A snapshot of Times Square
contains a great deal of accurate visual information. Looking at one you
could derive accurate information about the relative locations of the various
theaters and their distances from each other; you could tell how long it
would take you to walk from the theater showing Karate Kid, Part II to the
one showing Cobra.
Nevertheless, contemporary film theorists are very suspicious of perspec-
tive and regard it as an important factor in the propagation of ideology in
cinema. Jean-Louis Comolli, explicating an argument of Marcelin
Pleynet's, of which he approves, discusses the impregnation of the basic
128 T H E CINEMATIC IMAGE

(cinematic) apparatus by the Quattrocento code of perspective, whose

significance he unpacks in terms of
the hegemony o f the eye, "visual ization," and the ideology o f the visible
linked to the Western tradition o f systems centered on a single point. The
image produced by the camera could not fail to confirm and reinforce "the
visual code defined by renaissant humanism" which placed the human eye at
the center o f the system o f the other organs of senses: the eye (Subject)
enthroned in the place of the divine (humanism's critique o f Christianity). 2i

Here we see that perspective is thought of as a code or convention, and

that it is complicit with something called "the ideology of the visible"
which has nefarious metaphysical overtones. Moreover, contemporary film
theorists see perspective as a mechanism that operates, along with the film's
narrative, to position subjects in the sense discussed in the previous chapter
on the Althusserian-Lacanian paradigm. Stephen Heath, speaking of
G. Ten Doesschate's description of perspective, writes:
The component elements o f that account should be noted: the possible exact
match for the eye o f picture and object, the deceptive illusion; the center o f
the illusion, the eye in place. What is fundamental is the idea o f the spectator
at a window, an "aperta finestra" that gives a view on the world—framed,
centered, harmonious. 24

The organization of space by perspective here is coincident with illusionarv

subject unity. Perspective implements subject positioning, literally suggest-
ing a correct place for the spectator. Heath maintains:
The film poses an image, not immediate or neutral, but posed, framed and
centred. Perspective-system images bind the spectator in place, the suturing
central position that is the sense o f the image, that sets its scene (in place, the
spectator completes that image as its subject). 25

This view concerning perspective is dogma among contemporary film

theorists; it can be and is asserted without argument in film analyses. Here
is an example of Bill Nichols applying it to a scene from The Birds in which
Melanie is looking out a window:
This metaphorical dimension to the violence poised at the window in part
turns on the principle o f central perspective basic to the photographic lens
(and its precursor: linear, renaissance perspective). The installation of the
viewer as subject depends upon reserving a singular place for him or her, the
reciprocal in front o f the image o f the vanishing point "behind" it, the point
o f origin from which the camera "took" its view and where we now take ours.
The two-dimensional image, from the point of view o f central perspective,
stands in for the world it re-presents as would an ordinary window if the view
T H E C I N E M A T I C IMAGE 129

beyond it could somehow be imprinted on its surface. This, incidentally, is

precisely the metaphor elaborated by Leonardo da Vinci to describe how
central perspective is achieved by a painter (and strikingly similar to the
registration of a latent image on a film strip).
The window that admits and frames the scene confirms the camera in its
ascension to the fixed position of source or origin which we, in turn, are
invited to assume. T o launch an assault at the window is, in turn, to assault
the place of the viewer; it is an act of aggression against the eye of the
beholder and the " I " of the self-as-subject insofar as that " I " originates in the
realm of the imaginary. 26

The contemporary film theorist, then, believes that perspective is a code,

that it is complicit with the "ideology of the visible," that it functions to
produce subjects (in the Althusserian-Lacanian sense) by positioning spec-
tators, and that it is appropriately described in the language of deceptive
illusion. Later we shall also discuss its putative special relation with narra-
tive. But let us now turn to an examination of the claims about perspective
reviewed so far.
Since we have already spent much space discussing illusionism, we can be
brief about the application of that concept to perspective images. Perspec-
tive causes an impression of depth but no deceptive illusion. One way to see
this is to recall that the perspective image is drawn from afixedmonocular
station point. But we characteristically view perspective pictures including
cinematic images with two eyes. As well, we often change our position
while perusing paintings and we can rotate our heads while viewing a
movie image. These factors alone (though there are others) normally guar-
antee that we realize the projection is a flat surface and not really deep,
though it may have pictorial depth, a technical term not referring to real, i.e.,
three-dimensional depth, but to the impression of depth a two-dimensional
surface can impart. In fact, David Bordwell argues, successfully I think, that
"Under normal viewing conditions, the greater a picture's perspectival
depth, the less likely we are to be fooled."27
Contemporary film theorists are wont to say that perspective is a code,
implying that perspective is a culturally rooted, pictorial convention, akin
to using a halo as a sign of sainthood. For contemporary theorists, given
their commitments to semiotics, a convention or code is arbitrary; it can
always be otherwise, though it may appear to users of the cultural code that
it is "natural" (rather than cultural and arbitrary). This propensity to appear
natural, moreover, has important ideological potentials not only because it
masks the cultural origin and signification of the phenomenon in question,
such as perspective, but because it imbues whatever it organizes, however
ideologically charged, with an aura of naturalness, the property of "that's-
130 T H E C I N E M A T I C IMAGE

the-wav-it-is-ness." But before perspective can figure in an argument about

its naturalizing potentials, it must be established that it is, in fart, coded or
conventional.
Why presume that perspective is conventional? One argument might be
that though perspective presents itself as a replica of vision, it is not a
perfect replica of ordinary vision. And one might go on to say that if
perspective does not replicate ordinary vision, then not only is it not w hat it
pretends to be, but it is not accurate; it is only a (culturally) arbitrary
picturing of the world. But I am not convinced that a case can not be made
for perspective's accuracy over and above its cultural rootedness. Nor am I
persuaded that, given the purposes of perspective, it can be regarded as
arbitrary.
In order to defend the accuracy of perspective, we must clarify that about
which perspective is supposed to be accurate. If we say that a perspectival
picture, in order to be accurate, must afford a perfect replica of normal
vision, then perspective is not accurate. But the question is whether this is
the right requirement for accuracy. Of course, contemporary film theorists
can find theoreticians and polemicists who will make claims that perspec-
tive replicates vision, but they may have been overly enthusiastic and, as a
result, they may have missed the particular grounds upon which claims that
perspective can, in fact, make to nonarbitrarv, non-culturallv determined
accuracy can be based.
When I say that a perspective picture of a scene in nature is accurate, I do
not mean that it is a perfect replica of anything, including human vision.
Rather, I mean that it provides accurate information about certain aspects
of the appearance of the natural scene. Of course, the information perspec-
tive drawings excel in providing concerns the appearance of the relative
positions of things in space and the distances between them. No other
system of pictorial representation is as good at giving this kind of informa-
tion as is the perspective system. Other systems, like the ancient Egyptian,
had other purposes than the perspective system does. But it cannot be
denied that for the purposes of pictoriallv projecting the placement of
things in space, informatively, the perspective system is superior—that is,
contains more accurate information—than does the Egyptian system. In
fact, the perspective system is more accurate, transculturally, in terms of
affording certain spatial information than any other mimetic pictorial sys-
tem.
Two points in the preceding argument need to be extended. First, there
is a difference between the claim that a system provides a perfect replica of
the world, or of vision, and the claim that it provides an accurate represen-
tation. For accuracy is a matter of degree. My claim is that in a nonconven-
T H E C I N E M A T I C IMAGE 131

tionalist sense, perspective, transculturallv considered, is the most accurate

mimetic pictorial system of representation. Some dismiss such claims by
saying that perspective is not "true." But I am not claiming that it is "true,"
a term of assessment which I, by the way, think is more aptly applied to
propositions than to pictures. So rather than claiming that perspective is
"true," I claim it is accurate. But how do we measure accuracy across
representational systems? We can not, unless we stipulate the meaning of
accuracy in respect to some specific dimension.
This leads to my second point: when I claim accuracy for perspective, I
am saying that no competing, mimetic representational system is as accu-
rate as perspective in rendering information about the appearances of the
relative disposition of objects in space. This, it seems to me, is incontest-
able. Furthermore, perspective accomplishes this because it has been
grounded in the laws of vision. No other system of representation is based
on scientific laws in this way. And the laws from which perspective derive
are in no sense conventional, arbitrary, or adopted by fiat. They are laws.
This is the source of what gives the system the accuracy it has in recreating
the spatial appearance of the world. Moreover, the laws of the scientific
theories in question were not adopted as a result of a social compact or
decision, in the way of codes and conventions; they were discovered.
The skeptic, considering the preceding argument, will believe that he has
located a slip up. For to make my argument, I must stipulate that the
relevant sense of accuracy is accuracy in respect to the disposition of objects
in space. This is correct, but it does not cut against my argument since it is a
conditional one—if we want pictorial accuracy or fidelity about the ap-
pearance of the disposition of objects in space, then perspectival systems of
picturing are best suited to the purpose, no matter what culture you hail
from. Of course, I admit that we may not be especially concerned with this
kind of accuracy, but this is beside the point. If we want to saw through
metal, then we do best to use a hack saw; if we want accurate spatial
information about the appearance of the layout of the world from our
pictures, then we use perspective. In neither case are we playing by the rules
of society. We are adapting to the structure of the world, including that of
human biology.
Of course, it is true that there are many different types of representational
systems. And many of these systems are not devoted to giving accurate
information about the way the world appears—for example, abstract ex-
pressionism. These nonmimetic, nonperspectival systems may interest us—
especially from an aesthetic point of view—because they have features,
qualities, and purposes besides those of delivering accurate spatial ap-
pearances. And for purposes other than delivering accurate spatial informa-
132 THE CINEMATIC IMAGE

tion, these other systems may be more desirable than and superior to
perspective systems of the sort developed in the Renaissance. But if one is
committed to pictorial fidelity in respect to spatial appearance, then per-
spective, as a result of its scientific origins, is the best means to that end.
We may make our point in a related but slightly different way by drawing
a distinction between the genus, representational systems, and the species,
mimetic pictorial systems. Representational systems include practices that
refer to the world but which do not do so in such a way that percipients are
able to recognize the referents of the pictures without any special training.
Mimetic, pictorial representations, or pictures, for short, are symbols
whose referents can be recognized, in the picture, by viewers who have not
been trained in any special method of "reading" pictures. Some explain this
phenomena in terms of resemblance relations between the picture and its
referents, while others speak of the sharing of invariant perceptual cues.
Which explanation is superior is not germane to our discussion at this
point. What is key is that perspective pictures are a subclass of pictorial
representation.
Now there are undeniably many types of visual representations, such as
American stop signs, that do not involve pictorial recognition. And the
existence in a culture's representational system of representations that work
by way of pictorial recognition may be the result of a social contract. But
the fact that there exist nonmimetic representational practices does not tell
against the accuracy of perspective; it only shows there can be representa-
tional systems for which the kind of accuracy perspective excels in is
irrelevant.
. Moreover, there are pictorial practices—such as the Japanese floating-
eye style, the ancient Egyptian frontal-eye style, and what Deregowski calls
split-type drawings 28 —which do not afford the degree of spatial accuracy
involved in Western perspectival picturing. But, of course, this may only
indicate that spatial accuracy may not be the central concern of every
pictorial mode of representation. However, we still may compare these
various pictorial practices in terms of spatial accuracy. And when we raise
the question of spatial accuracy—which is a question of a specific variety of
accuracy—in the context of pictorial systems we observe that the subset of
pictures that employ Western perspective arc the most spatially accurate
ones.
Is the preceding claim just an example of the imperialistic hauteur of the
West, as many contemporary film theorists might argue? No. In the early
eighteenth century, the Japanese taught themselves the Western perspec-
tive system by comparing the perspective illustrations of Dutch scientific
books to their own illustrations, and they judged the Dutch illustration to
T H E CINEMATIC IMAGE 133

be more "lifelike" (which I would interpret as more accurate for the pur-
pose of spatial mimesis). This resulted in the genre of Uki-e, or "relief
pictures," pictures of three-dimensional effect, using perspective and shad-
ing, made by such masters as Hokusai, Kunizoshi, and Hiroshige, among
others, and easily marketed.29
If we arc attempting to debate about whether perspective is the aesthet-
ically best pictorial system, it might make sense to argue that perspective
systems and alternative methods of picturing, like the floating-eye style, are
somewhat incommensurable because they are subtended by different cen-
tral purposes. Certainly paintings with "floating points of view" can be
aesthetically superior to perspective paintings, and vice versa; it may be-
come difficult to weigh the comparative value of a masterpiece in the
perspective style with a masterpiece of the floating point of view style. We
can make the prosaic observation, nevertheless, that, all things being equal,
the perspective painting will give us more information about spatial ap-
pearances than will the picture with the floating point of view. This is the
sense in which perspective can be said to be "true" to the world, though it is
better to express this by saying that it has the highest available pictorial
fidelity. Of course, saying the perspective system has the highest degree of
accuracy does not entail that it may not involve some inaccuracies. It only
claims that perspective is the most accurate pictorial means of rendering
information about spatial appearance. These observations about the ac-
curacy of perspective count against not only the idea that perspective
involves deceptive illusion, but also against the notion that perspective is a
code, for codes are by definition arbitrary. But "if you want your picture to
convey information about the relative locations, sizes, and dimensions of
objects in a measurable space as seen from one spot, then scientific perspec-
tive is not as arbitrary as other systems."30
When I have previously defended perspective's relative superiority vis-à-
vis spatial accuracy, critics have cited maps as counter-examples to my
hypothesis insofar as maps may be more accurate about spatial layouts than
perspectival pictures.31 But clearly maps are not suitable counter-examples.
For I claim that perspective is more accurate in terms of affording spatial
information than any other mimetic pictorial system. And maps are not
pictures. One must be taught to read a map; the mountains on a map are
not recognized perceptually as mountains but are coded, often by color.
Maps are read, not recognized.
Another complaint about my formulation is that by speaking of ap-
pearances in my account of perspective, I have inadvertently stumbled into
relying on the notion of "illusion," thereby contradicting myself, given my
objections to correlating perspective and illusion.32 However, critics over-
134 T H E C I N E M A T I C IMAGE

look the crucial fact that the word "appearance" need not mean "illusion,"
nor is the "illusion" sense of "appearance" being used in the preceding
formulations. When a police officer questions witnesses about the ap-
pearance of a bank robber, the last thing she wants is a report of an illusion.
Except for certain contexts, such as Platonic dialogues, where the operative
dichotomy is Appearance versus Reality, an appearance is the outward
aspect of any physical thing. There is no reason to suspect that in ordinary
language ' T h e milk appears spoiled" means "The milk is not spoiled but
deceives us into thinking it is." That is, talk about appearances can be talk
about how things are, not about illusions.
Nor need talk of appearances refer to a particular instance of vision, but
may refer to what is visible. That is, a perspective picture described as
affording spatial information about appearances is not being characterized
as providing a phenomenological replica of a specific act or kind of act of
vision but as providing information about the structure of ambient light in
an optic array which, in turn, is the sort of thing from which humans derive
reliable information about the layout of things in space. I use the word
"appearance" to signal that the spatial information in perspective pictures is
visual. But my commitment to perspective as affording information about
the structure of light in optical arrays hardly commits me to the belief that
perspective is a point perfect replica of the experience of normal seeing. It is
not a representation of seeing; it affords spatial information from the
optical structure—from the appearance—of the layout of the environment.
Moreover, the information afforded by perspective is more accurate than
that available from any competing, mimetic pictorial system.
Contemporary film theorists move from the notion that Renaissance
perspective is an arbitrary code or convention 33 to the attribution of similar
coding to the cinematic image as such. This is doubly wrong—wrong not
simply because perspective is regarded as a code or a convention, but also
because not every cinematic image is governed by perspective. Consider
soft: focus shots without vanishing points and the capacity of such things as
telephoto lenses to violate perspective laws. 34
Reigning doxy has it not onlv that perspective is a code or convention,
but that it functions in cinema as an instrument for positioning subjects in
the Althusserian-Lacanian sense of subject. Recall Heath's contention that
"perspective-system images bind the spectator in place, the suturing central
position that is the sense of the image, that sets its scene (in place, the
spectator completes the image as its subject.)"
We can give Heath's first use of "in place" some literal sense. Namely, a
perspectival image has an optimal station point, a place in space. The place,
that is, is the implied physical vantage point of the perspectival image. In
T H E CINEMATIC IMAGE 135

fact, when viewing a movie, few of us ever occupy this precise station point.
Yet the image may be said to have this point, even if our inhabiting it is
rarely germane to our still deriving accurate spatial information from the
image (I can be sitting far off center in a side aisle and still know that
Humongus is about three feet behind his next victim). Heath would have it
that the perspective system somehow enjoins us to believe that we are at the
monocular station point o f the perspectival image—that we identify our
position, wherever we are in the theater, with the location o f the camera
lens in relation to the profilmic visual array. We are thus seduced into
believing that we have a certain position in relation to the film, a unitary
place in the center o f the action. This sense of "in place" leads to another—
the Althusserian-Lacanian sense of "in place" in the above quotation.
Identifying with the visual center of the action causes a sense of psychologi-
cal subject unity—the illusion of the "self" as a single, unitary, placed,
centrally positioned center of experience—which presumably, in turn,
heightens the illusion of the image ("completes" it). That is, Heath main-
tains a connection between the centrality of perspective's physically locata-
ble station point and the centered position of the subject/spectator as
coherent and stable. And this type of subject positioning is, in turn, seen to
be a key ideological effect, creating subject supports upon which capitalism
can drape its sundry roles.
In the previous chapter we have already had cause to question whether
the abstract subject supports believed to be created by the aforesaid type of
positioning really supply any clue to the formation of the kinds of person-
alities and economic role players required by capitalist society. Also, in our
discussion of Metz, we began to review some of the difficulties involved in
maintaining that spectators identify with the camera. If spectators did
identify with the position of the camera, they could only do this by ignor-
ing a great deal of visual information which the human organism has at its
disposal (if only subliminally). For example, cameras do not "see around
corncrs." So when I watch a film from the extreme left side of the theater, I
will not see the same thing that I would see if I occupied the same seating
arrangement in relation to the profilmic event. This does not prevent me
from getting accurate information about spatial appearances from the
screen, for, if I know about cameras, I know the disposition of objects is
relative to the lens and not to my seat. But if Heath wants to claim
identification with the camera he must explain how the organism represses
all the available spatial cues at its disposal.
Heath attempts to do this through reference to an unconscious mecha-
nism that is especially responsive to perspective. In order to assess Heath's
explanation we must scrutinize the supposed internal workings of this
136 T H E C I N E M A T I C IMAGE

unconscious mechanism. Heath describes perspective in terms of a central

position, a unified place. Words like central, position, unified are key words
in the Althusserian-Lacanian psychology. They are words used to describe
subject construction. Indeed, subject construction, in the conceptual
framework of contemporary film theory, is synonymous with subject posi-
tioning and subject unity. Thus, the explanation we are being offered of the
spectator's identification with the camera appears to be that given the
central (single, unified) position (place) of the station point in perspective,
the spectator is lured (and deceived or trapped) into identifying with the
camera lens (and, presumably, filtering out all the available, contrary spatial
cues) because the structure of perspective reinforces the illusion of subject
unity (in the psychological-ideological sense) and, putatively, because spec-
tators are prone to accept reinforcements of the illusion of subject unity
above countervailing cognitive and perceptual stimuli. Is this a persuasive
explanation?
If we agree that there is a literal sense in which perspective involves a
position, then the sense in which it "positions spectators" is nonliteral; it
does not, for example, move them around the theater. Furthermore, the
sense in which spectators are positioned as subject unities is also not literal,
but rather is an explanatory metaphor, one that pertains to a conception of
the self without any reference to actual points in space. But what do these
different phases of the identification have to do with each other besides the
fact that they are described by a similar vocabulary?35 Perspective does not
require that we be positioned at the literal monocular station point of the
image in order to work effectively as a mimetic, pictorial representation, so
it does not literally position us—though it may metaphorically position us
in the sense of establishing the orientation of the action. But even if
perspective did somehow literally dictate a central position, how would
that cause the belief that each of us is a unitary subject? After all, the idea
that we are the (nonspatial) centers of our experience is another metaphor.
How do we or our psyches parlay a (doubtful) spatial position of cen-
trality into a belief about the centrality of the subject? The latter is a very
different phenomenon from the former, even if it is described in similar
words. The Althusserian-Lacanian position does maintain that visual stim-
uli have an especially powerful role in the formation of identity. Neverthe-
less, the equivocation on the word center in Heath's account cannot be
bolstered by a simple allusion to the mirror stage since literal, physical
centering is not a necessary component of that phenomenon—couldn't a
child be standing at the left corner of her looking glass? Rather it seems that
Heath will have to postulate an etymologicallv playful psychic mechanism
T H E C I N E M A T I C IMAGE 137

that moves from perspective's station point to subject unity along a series of
puns.
That, in and of itself, is not, in principle, an insurmountable problem,
since it has been established that the unconscious is fond of literalization
and verbal images. The problem is more specifically whether Heath can
postulate unconscious receptivity to the puns he relies on in his explana-
tion—position, for example. That is, either the receptivity to such puns is
part of the psychic mechanism or it is an equivocation on the part of
Heath's attempted explanation. The former, without further argumenta-
tion, appears unlikely, especially when one recalls that certain of Heath's
subject unity metaphors are not common parlance, but esoteric jargon of a
very recent, marginal school of psychoanalysis. Are we to believe the
unconscious knew these puns as as early as the Quattrocento? Also, are we
to believe that the normal viewer unconsciously knows the rhetoric of the
theory of perspective? On the other hand, a weakness for equivocation
saturates contemporary film theory and, thus, it seems likely that it is in
operation in this explanation too.
Earlier we learned that perspective is supposedly complicit in the ideol-
ogy of the visible. A major difficulty in assessing this accusation is knowing
what in the world the "ideology of the visible" comprises. If it reduces to
the production of positioned (centered) subjects, then it is liable to the
foregoing objections as well as the objections made in the preceding chap-
ter. But perhaps the ideology of the visible involves more than subject
positioning (and centering). For example, it is often noted that much
modern epistemology is rooted in questions of visual perception and em-
ploys primarily visual metaphors for knowledge. Maybe this feature of
Western philosophy is the locus of or at least a part of the ideology of the
visible. However, if this conjecture of mine is correct, two points should be
made here. First, though epistemology is obsessively concerned with vision
and visual metaphors, perhaps to its own detriment, perspective is not a
central metaphor in its discussions, nor does perspective figure in the major
debates in the philosophy of perception. Thus, it would seem off the mark
to regard perspective as a major contributor to the ideology of the visible if
that ideology is identified with the place of perception in epistemology.
Second, if philosophy's emphasis on visual perception is really the culprit
here, one wants to know why its possibly skewed research program is
ideological? In what system of social domination does it participate?
In our quotation from Comolli we were told that the camera reinforces a
code which "placed the human eye at the center of the system of representa-
tion, thereby excluding other systems and assuring the domination of the
138 T H E C I N E M A T I C IMAGE

eye over the other organs of senses." Does this mean that the relevant form
of domination in the ideology of the visible is that of the eye over the other
senses and that any representational system that "privileges" sight, as per-
spective does, is ideologically complicit? This is a strange sense of ideologi-
cal domination. I find it hard to think of my eyes as subjugating my nose.
Human visual capacities are more highly developed than our other senses,
but it is difficult to see the way in which that fact plays a significant role in
any system of social domination. Nor does it seem appropriate to think of
the fact that we rely so heavily on vision as a matter of privileging it. Do we
think bats privilege the sonic or that they foster an ideology of the sonic?
Would bat art be complicit in the ideology of the sonic? And who would be
the oppressed in this ideological arrangement?
Last, perspective is said to naturalize the ideologically charged images it
organizes by passing itself off as something other than the code or conven-
tion it really is. We have explored reasons for disputing part of this assertion
in our denial that perspective is code. But apart from the issue of whether or
not perspective is a code, it is still possible that it performs some naturaliz-
ing magic on its images. However, the argument against this hypothesis
would be of a piece with our arguments against illusionism's purported
naturalizing effect. That an image is perspectivally organized in no way
enhances the credibility of its ideological message. This can be seen quite
easily by considering advertisements. If you don't believe that drinking
wine cooler will make you a zesty, fun-loving, beach bum, then a perspec-
tively shot ad will not, in virtue of its perspective, convince you.
Indeed, there is another way in which ads show that the contemporary
approach to perspective is misguided. If ads are thought of as effective
disseminators of ideology, then it cannot be their use of perspective that
accounts for their success. For many ads are not perspectival. Many, for
example, are photomontages, full of virtually cubist spatial contortions.
These ads seem as successful as perspectival ads. Therefore, the presence or
absence of perspective in ads seems irrelevant to their ideological effect.

AN A L T E R N A T I V E VIEW O F T H E C I N E M A T I C IMAGE
In this chapter we have been concerned with the cinematic image, which
we have thought of as the image projected by a single shot—a close-up of a
hero's face, or a long-view of Casde Dracula. We have been highly critical of
the approach to the cinematic image of contemporary theorists. But do we
have an alternative account and can that alternative account tell us anything
that is theoretically informative? Thus, I shall attempt to sketch an alterna-
T H E CINEMATIC IMAGE 139

tive account and to show that this alternative account can serve as part of an
answer to a question that contemporary theorists have about movies.
Contemporary theorists, it seems to me, are often motivated by the ques-
tion of why movies—the commercial entertainment films of Hollywood
International—are so powerful, that is, so popular cross-culturally as well
as across the boundaries of race, gender, and class. The following account
of the cinematic image provides an inroad into answering that question.
Cinematic images, in their standard use in movies, are, for the most part,
representational, but, more important, they are pictorial representations.
They refer to their referents by way of picturing them. This process of
picturing has been explained in various ways. Some hold that pictures refer
by displaying or manifesting a delimited range of resemblance to their
referents; or some prefer to say that pictures operate by displaying certain
of the same perceptually invariant cues as their referents do. Which of these
psychological theories, if they are ultimately competing theories, is
stronger is not my present concern. For whether by means of similarities or
shared cues, pictorial representations are the sort of representations whose
referents—whether the picture depicts a man, a horse, or a house—the
spectator recognizes simply by looking.
Given that the typical movie image is a pictorial representation, what has
this to do with the popularity of movies i.e., with their accessibility across
nations, creeds, races, and classes? Well, a picture is a very special sort of
symbol. Psychological evidence strongly supports the contention that we
learn to recognize what a picture stands for as soon as we have become able
to recognize the objects, or kinds of objects, that serve as the models for
that picture. Picture recognition is not a skill acquired over and above
object recognition. Whatever features or cues we come to employ in object
recognition, we also mobilize to recognize what pictures depict. A child
raised without pictorial representations will, after being shown a couple of
pictures, be able to identify the referent of any standard picture of a kind of
object with which he or she is familiar.36 The rapid development of this
picture-recognition capacity contrasts strongly with the acquisition of a
symbol system such as language. Upon mastering a couple of words, the
child is nowhere near mastering the entire language. Similarly, when an
adult is exposed to one or two representational pictures in an alien pictorial
idiom, say a Westerner confronting a Japanese image in the floating-eye
stvle, she will be able to identify the referent of any picture in that format
after studying one or two representations of that sort for a few moments.
But no Westerner, upon learning one or two linguistic symbols of the
Japanese language, could go on to identify the reference of all, or even
140 T H E CINEMATIC IMAGE

merely a few more, Japanese words. Moreover, as we have already seen, the
Japanese were eminendv able to catch on to and replicate the Western
system of perspectival picturing by examining a selection of book illustra-
tions; but they could never have acquired any European language bv
learning the meanings of just a few words or phrases.
Pictorial representation thus differs radically from linguistic representa-
tions. The speed with which the former is mastered suggests that it does not
require special learning, above the realization, perhaps, that flat surfaces are
being used to stand for three-dimensional objects. Rather, the capacity to
recognize what a picture depicts emerges in tandem with the capacity to
recognize the kind of object that serves as the model of the picture. The
reciprocal relation between picture recognition and object recognition, of
course, explains how it is possible for us, having acquired detailed visual
information from pictures, to recognize objects and places we have never
encountered in real life. And, of course, the fact that pictorial recognition
does not require any special learning process would also explain how
movies, whose basic constituent symbols are pictures, are immediately
accessible to untutored audiences in every corner of the world. These
audiences do not need any special training to deal with the basic images in
movies, for the capacity to recognize what these images are about has
evolved part and parcel with the viewer's capacity to recognize objects and
events.
The technology of film could be adapted in such a way that the basic
images of a film genre or film style were not pictorial representations. One
could imagine a morion picture industry of changing abstract forms, after
the fashion of Hans Richter's Rhythmus 21, or one of spectacles of color,
such as Stan Brakhage's Text of Light. But that was not the road taken by the
movies. Movies became a global phenomenon—and a lucrative industry—
precisely because in their exploitation of pictorial recognition—as opposed
to symbol systems that require mastery of processes such as reading, decod-
ing, or deciphering in order to be understood—they rely on a biological
capability that is nurtured in humans as they learn to identify the objects
and events in their environment.
The basic images in movies are not simply pictorial representations; they
are, standardly, moving pictorial representations. But just as an audience
need not go through a process of learning to "read" pictures, neither is its
perception of movie "movement" learned. Rather, it is a function of the
way stroboscopic or beta phenomena affect the brain's organization of
congruous input presented in specifiable sequences to different points on
the retina. O f course, following a movie involves much more than the
T H E CINEMATIC IMAGE 141

capacity to recognize what its moving images represent. But we should not
overlook the crucial role that the relative ease of comprehending the basic
symbols of movies plays in making movies readily accessible.
M y remarks thus far are apt to displease the majority of contemporary
film theorists. For the contention that pictures (and, by extension, motion
pictures) work by enabling their referents to be recognized in virtue of the
way in which our perceptual system is keyed, goes against contemporary
received wisdom that pictures, like any other symbol system, are matters of
codes and conventions. Undoubtedly, some contemporary theorists will
recall an anthropology class in which he was told that certain non-Western
peoples were unable to understand pictures shown to them by missionaries
and other field workers. However, the evidence here has never been entirely
decisive. Complaints about the fidelity of the photographs involved have
been raised, along with the more serious objection that what the subjects
failed to understand, and then only initially, was the practice of using flat
surfaces to portray three-dimensional objects. 37 Once they got the hang of
that, they had no trouble in recognizing what hitherto unseen pictures
referred to—assuming they were familiar with the kinds of objects dis-
played in the pictures. 38 And Margaret Hagen writes:
There is no reliable evidence that pictorially naive people are incapable of
either perceiving or drawing pictures of isolated objects. For the perception of
spatial layouts in pictures the findings are less consistent and subjcct to some
heated debate. It is safe to conclude that it often takes a short period of time
for the pictorially naive to learn to attend to the depicted contents rather than
to the medium (paper, canvas, whatever), and the accuracy andflexibilityof
their pictorial spatial perceptions are often limited in apparently surprising
but understandable ways. What is important is that complete failure of spatial
perception in pictures by the naive is almost unknown in the literature.
Unless one wishes to argue that a completely arbitrary code of depiction can
be acquired, learned, and understood in a truly astonishingly short period of
time, then the hypothesis that pictures function as representations without a
basis in resemblance is untenable.39

Also, on the non-conventionalist side of the scale, we must weigh the

psychological evidence of the child's acquisition of pictorial practices, and
the zoological evidence that certain animals have the capacity for pictorial
recognition, 40 against exotic anecdotes that are meant to demonstrate that
the practices of picturing are cultural conventions that must be learned in
the fashion of language acquisition. For when would children have time to
enter our pictorial convenants, and how could animals?
Moreover, we can consider our own cases. We all recall our own lan-
142 THE CINEMATIC IMAGE

guage acquisition and we know how to go about helping youngsters to

learn to speak and to read. But who remembers undergoing a similar
process in regard to pictures, and what techniques would we employ to
teach a youngster "pictorial literacy"? Yes, we may show a child a few
pictures, and then say the name of the object portrayed and point to parts of
the object. But very shortly the child just sees what the picture is of; the
child does not "read" the picture or decode it or go through some process
of inference. And from a meager set of examples, the child can proceed to
identify the subjects of a plethora of pictures, because there is a continuum
between apprehending pictorial representations and perceiving the world
that does not depend upon learning anything like the conventional, arbi-
trary correlations of a vocabulary, or the combinatory principles of a gram-
mar.
There is undoubtedly a temptation to think that picture recognition
involves some process of decoding or inference because of the contempo-
rary influence of the computational metaphor of the mind. We think that
computers supply us with powerful insight into how the mind works. And
if we were to build a computer to simulate pictorial recognition, it would
require a complex information-processing system. But it does not follow
that if computers employ complex information-processing systems in pic-
torial recognition, then humans must likewise possess such systems. It may
rather be that our neurophysiology is so constructed that when stimulated
by certain pictorial arrays, we see what the picture is of. John Searle notes
that balance is controlled by the fluids of our inner ear. Were we to
construct a robot, balance would probably be governed by some complex
computational program. But, for us, balance is a matter governed by our
fleshy hardware.41 A similar case might be made that biology—rather than
information processing—may have a great deal to tell us about the work-
ings of object recognition and picture recognition. And to the extent that
pictorial representation is a matter of the way in which humans are made, a
practice rooted in pictorial representation—such as the movies—will be
widely and easily accessible to all humans made that way.42
Many contemporary film theorists, due to their inclinations toward
semiotics, resist approaching pictorial representation in the movies in the
preceding fashion. Their resistance rests on a confusion, or rather a confla-
tion, on their part of the ideas of code, convention, and culture. In film
studies these terms are treated as if they were equivalent. If something is
coded or conventional, then it is regarded as a cultural production. This
seems fair enough. But it is more problematic to presume, as film theorists
do, the inverse; that if something is a cultural product, then it is an example
T H E CINEMATIC IMAGE 143

of coded or conventional phenomena. Thus, if pictorial representations,

including moving pictorial representations, are cultural productions, which
they certainly are, then they must be conventional. The difficulty here lies in
the assumption that everything that is cultural is necessarily conventional.
Consider plows. They are cultural productions. They were produced by
certain agricultural civilizations that had culturally specific needs not
shared, for example, by hunter-gatherers. Is the design of a plow a matter of
convention? Recall, here, that for semiotically influenced film theorists,
arbitrariness is a key defining feature of a convention. That is, a group
creates a convention—like the color of a police cruiser—when there are a
number of alternative ways of dealing with a choice situation, and where
the decision between these alternatives is a matter of indifference—i.e.,
each of the alternatives is equally satisfactory. The choice between these
alternatives is arbitrary and is reached by fiat. But the adoption of the
design of the plow could not have been reached by fiat. The plow had a
purpose—digging furrows—and its effectiveness depended on its being
accommodated to the structure of nature. It would have to be heavy
enough and sharp enough to cut into the earth, and it had to be adapted to
the capacities of its human users—it had to be steerable and pullable by
creatures like us with two arms and limited strength. A device such as a
plow had to be discovered; it could not be brought into existence by
consensus. We could not have elected pogo sticks to do the work of plows.
The plow was a cultural invention, not a cultural awpention. It was adopted
because it worked, because it met a cultural need by accommodating
features of nature and biology.
The point of introducing the concept of a cultural invention here is, of
course, to block the facile identification of the cultural and the conven-
tional. Applied to the sort of pictorial representations found in movies, this
concept suggests that pictorial representations may be cultural inventions,
inventions that, given the way people are built, cause spectators who are
untrained in any system of conventions to recognize what pictures stand
for. The structure of such images is not determinable by mere decision.
Given the constraints of the human perceptual apparatus, we cannot decree
that anything is recognizable as anything else, though we may decree that
anything can stand for anything else. It seems cogent to suppose that this
limitation is in large measure attributable to human biology. And insofar as
movies are constituted of a mode of representation connected to biological
features of the human organism, they will be generally more accessible than
genres in other media, such as the novel, that presuppose the mastery of
learned conventions such as specific natural languages. Also, if the recogni-
144 T H E C I N E M A T I C IMAGE

tion of movie images is more analogous to a reflex process than it is to a

process like reading, then following a movie may turn out to be less taxing,
less a matter of active effort, than reading. Perhaps this can be confirmed by
recalling how much easier it is to follow a movie when one is fatigued than
it is to read a novel in the same condition.
The claim has so far been advanced that one crucial factor in the popu-
larity of movies is the fact that movies usually rely, in terms of their basic
imagery, on pictorial representations that allow masses of untutored specta-
tors easy access to the fundamental symbols in the system, due to the way
humans are constructed. But is this not just a reversion to the kind of realist
explanation we dismissed earlier? Not at all. The Bazinian claims the specta-
tor somehow regards the film image to be identical with its referent, while
contemporary film theorists hold that the typical cinematic image imparts
the illusion of reality, of transparency, or of naturalness. My contention,
though, has not invoked any of these realist, psychological effects, nor
anything like them. It has instead claimed that the untutored spectator
recognizes what the film image represents without reference to a code; I
have not claimed that the spectator takes the pictorial representation to be,
in any sense, its referent. Human perceptual capacities evolve in such a way
that the capacity for pictorial recognition comes, almost naturally, with the
capacity for object recognition, and part of that capacity is the ability to
differentiate pictures from their referents. Thus, we are not talking about a
realist, psychological effect—the taking of a representation for its refer-
ent—but only about the capacity of movies to exploit generic, recogni-
tional abilities. Another way to see the difference between this approach
and that of the realists is to note how often in the realists' account of the
effect of movies they emphasize the importance of the fact that movies are
photographic, whereas in the account offered here, the important technol-
ogy for explaining the effect of the image is the non-cinema-specific tech-
nology of pictorial representation.
I have characterized the cinematic image as a type of pictorial representa-
tion. Pictorial representations, in turn, are those whose referents are recog-
nized by untutored spectators simply by looking; that is, they are recog-
nized by spectators who have not been trained in some process of reading
or decoding, nor do they identify the referents by inference. When a viewer
looks at a picture, the viewer recognizes the kind of object the picture
depicts—a tree, a gun, a dog—where the viewer is already familiar with the
kind of object depicted. The viewer can also recognize particulars portrayed
in pictures, where the viewer is knowledgeable of the particular. For exam-
ple, I can recognize the referent, George's house, in the picture, if I am
T H E CINEMATIC IMAGE 145

already acquainted with George's house. The viewer, of course, may also be
told what the unfamiliar particular portrayed in the picture is—the space
shuttle or the Jolly Green Giant—and then go on to recognize it in succes-
sive pictures. This capacity for pictorial recognition, moreover, it seems
must be at least partly explicable in terms of the structure o f the human
perceptual apparatus.
Some contemporary film theorists may object to my presentation of the
preceding theory as a contrast to their own. For some theorists, as we have
noted, regard the cinematic image as iconic, as having a motivated relation
of resemblance to the referent. However, they still regard iconic symbols as
coded or conventional, and as basically explicable in terms of culture. The
preceding theory, however, not only remains neutral on the issue of resem-
blance (though I am not entirely unsympathetic to it) but also stresses that
cinematic images are cultural inventions, not conventions, and that consid-
erations of the structure of the perceptual apparatus are germane to an
account of pictorial representations in general and cinematic images in
particular. (In terms of the previous section on perspective, it should be
recalled that we classified perspective pictures as a subcategory of pictures).
Another charge that contemporary film theorists will level at my theory is
that it reduces pictorial representation to biology and therefore denies it a
history. This is a red herring. The preceding account does not deny that
there are many forms o f pictorial representation and that pictorial represen-
tation in a given culture can undergo development. One can profitably
investigate the variety and vicissitudes o f pictorial representation. We have
only claimed that reference to the biological structure of the human percep-
tual apparatus has some important role in the account of pictorial represen-
tation and the cinematic image. Furthermore, I think that contemporary
film theorists have failed to acknowledge this. The preceding theory is a
corrective, then, but it does not preclude historical studies of pictorial
representation.
It is also true that I think the calling to mind of the role biology plays in
the communication of the cinematic image can, additionally, serve an
explanatory role. For if one wants to know how it is possible for movies to
be so popular to audiences of different cultural backgrounds, political and
religious affiliations, and class interests, then we should look to features of
movies that address the mass movie audience in terms of what those
audiences have in common. That is, we expect that if movies can engage
peoples of widely different cultural backgrounds successfully, then it is
plausible to suppose that movies are connected to fairly generic features of
human organisms. And identifying pictorial recognition, conceived as
146 T H E CINEMATIC IMAGE

partly a biological capacity, as one lever of the movie's popularity, certainly

connects movies with a fairly generic feature o f human organisms. This
undoubtedly does not account for why specific movies or specific types of
movies are popular with target audiences. But it does offer the start of a
theoretical account of the mass appeal of movies (to be expanded in ensuing
chapters).
One theoretical bonus, then, of our account of the cinematic image is
that it suggests part of an answer to the question of why the motion
pictures called movies, in which category I would include commercial,
fictional TV, are so popular, i.e., such a dominant mass art form. Thus, this
account offers an explanation of something that I think contemporary film
theorists, such as Metz, want explained. Moreover, it does this without
resorting to psychoanalysis, and, indeed, if it is compelling, it shows that
psychoanalysis, at the level of the cinematic image is beside the point, for
we can explain the mass appeal of the cinematic image in cognitive terms.
At the same time, however, I will be undoubtedly reminded that this
account does not offer anything by way of an explanation of the manner in
which the cinematic image as such participates in the dissemination of
ideology, an issue o f perhaps highest concern to contemporary film theo-
rists. In response, I must repeat that I am not convinced that the cinematic
image as such has any ideological effect; pictures, cinematic or otherwise,
may or may not be ideological depending on their internal and contextual
articulation. This is not to deny that movies and cinematic images cannot
be vehicles of ideology, but only that pictures as such are not ideological.
We will have to look at each image43 and each film individually in order to
determine its ideological content and not to the facts of cinematic represen-
tation itself.
Theoretically, the contemporary film theorist can claim to have a more
elegant explanatory framework than mine. The contemporary theorist can
explain both the popular allure of the cinematic image and its ideological
effects by means of the same set of reigning concepts, whereas my approach
is messier, explaining the mass appeal of movies with one set of concepts
and the ideological effects of given films ad seriatim. This is not entirely
correct, since I think we can make various generalizations about the opera-
tion of ideology in film, but it is true that the contemporary film theorist's
approach is more of a unified theory than mine. However, since I think that
the very notion that the cinematic image per se has an ideological effect
grievously mistaken, I prefer accuracy to elegance.
 4.

N A R R A T I O N

C O N T E M P O R A R Y film theorists appear to hold that representation

is inherently ideological, i.e., possessed of some politically discreditable
effect independently of what is represented. Narration is a form of repre-
sentation insofar as a narrative describes, and in that sense represents, a
sequence of events, generally human events. And, of course, film, in its
most salient use in our culture, is narrative. Thus, in this chapter, we shall
explore the case of contemporary film theory versus narrative.
Though undoubtedly there are some contemporary film theorists, like
Peter Gidal, who are opposed to narrative in general, the target of many
contemporary film theorists seems somewhat more narrowly, if obscurely,
circumscribed. They lodge their criticisms most frequendy against some-
thing called "realism," or "classic realism," or "the classic realist text," where
a film is also considered a "text." In regard to cinema, "classic realism"
generally involves at least the use of narration in coordination with what are
thought of as visual codes of verisimilitude.
As the term classic realism implies, it is applicable across media. Its
corpus includes things like objective (rather than nonobjective) photos,
traditionally plotted novels and plays, paintings proposing a purported
relation of similitude to their referents, TV soap operas and police shows,
advertisements for Calvin Klein underwear, Hollywood films like Death
Wish and Sixteen Candles, Neo-realist films like Bicycle Thief, a piece of
Soviet Realism like The General Line or of deep-focus realism like Munifi-
cent Ambersons.
Classic realism does not simply refer to a particular art movement such as
148 NARRATION

the one associated with Courbet, though Courbet is a classic realist. Nor is
realism simply Naturalism: The Golem, Warning Shadows, and The Band-
wagon are just as much a matter of realism in this sense as is Greed.
Rather, it would seem that "classic realism" is to be contrasted, in visual
media, with abstractionism, especially reflexive abstractionism, and/or with
reflexive subversions of the so-called codes of verisimilitude. In temporal
media, classic realism appears to be contrasted with the subversive struc-
tural ploys of the New Novel à la Robbe-Grillet and Duras, and, perhaps,
the disjunctive intertextuality of someone like Kathy Acker.
As we shall see, contemporary theorists are not very helpful in defining
"classic realism," forcing us, in the main, to rely on our ordinary intuitions
about what we pre-theoretically consider to be realist representations when
we read their assaults on the ideological effects of classic realism.
Roughly, I suppose, one can indicate the extension of the ruling concept
of classic realism by means of some loose historical boundaries. Generally,
classic realism comprises the practices of representation falling in line with
the dominant artistic sensibility from the Renaissance to the advent of
Modernism. In terms of film, classic realism involves at least the vast
majority of fictional films made from around nineteen-thirteen or so until
the crystallization of a Brechtian-Modernist cinema in the mid-sixties (with
notable préfigurations such as Man With A Movie Camera and Kuhle
Wampe) as well as fiction films after Godard et al. that continue in the pre-
Brechtian-Modernist, that is to say classic realist, style. Moreover, though
pictorial representation is a central ingredient in this style, narration is the
component that, for contemporary theorists, is first among equals in the
securing of the effect of classic realism.
In Visible Fictions, John Ellis attempts to clarify the operative notion of
realism as it is applied to film by contemporary theorists. He believes that
realist films are those that offer probable portrayals of characters and events
where the probability in question has not to do with generalizations about
actual behavior but with norms of what is plausible. This invocation of
probability, of course, shows the high degree to which classic realism is
bound up with narrative.
Second, Ellis holds, realist films aspire to a kind of surface accuracy in
costume, setting, props, etc. And next, a realist film is said to approach an
ideal of coherence "in the sense that events should always be seen as having
explicable causes and being related to each other within the representation,
rather than coming out of the blue."1 Of course, this can result in diver-
gences from Ellis' criterion of probability; that a newspaper reporter like
Clark Kent can fly is hardly probable, but it is perfectly coherent given the
NARRATION 149

premises of Superman. Ellis is aware of this tension and writes: "This

criterion (coherence) can be regarded as quite anti-realist by those who
believe that 'realism' equals the attempt to portray things as they are or
were. The demand that a representation should explain itself adequately to
its audience cuts across the desire that it should show things 'as they
were.'" 2
Last, Ellis includes under the rubric of realism, films such as Rossellini's,
that break with conventions in order to "get to a new sense of reality."3 The
kind of convention relevant here would include the requirement for expli-
cability. Thus, Ellis lumps together what might be broadly thought of as
Hollywood realism with its antipode, Neo-Realism, including radical en-
deavors in that form. This undoubtedly is disturbing for anyone who thinks
that the former has no claim to the title realism, whereas the latter in
subverting the former in the ways it does is a paradigm of realism.
Ellis is prepared for this charge, and acknowledges that he is characteriz-
ing different senses of realism; he says he is speaking of realism;. And yet,
unaccountably, he goes on to hypothesize the essence of realism. He writes:
In essence, realism is a regime of unified portrayal: every criterion of realism
aims at the same objective, to combine all the elements of the representation
at any point into a harmonious whole. This prevents the reading of the image,
scanning it to see its different elements and their possible conflicts or com-
binations, which is a central feature of modernist tendencies in the other arts.
(P- 9)

This characterization, however, is hardly adequate. It does not work on

Ellis' own terms. Rossellini's Poison is not harmonious in Ellis' sense, while
Rules of the Game encourages scanning for conflicts and combinations of
elements, thereby encouraging alternative interpretations. Also, this char-
acterization, with its heavy emphasis on the "harmonious whole" (that is,
on the criterion of coherence), suggests that visionary films such as Ken-
neth Anger's Inauguration of the Pleasure Dome are examples of realism.4
Clearly this will not do. Ellis has not given us a viable criteria of realism, nor
to my knowledge, has any other contemporary film theorist. However, Ellis
has more or less earmarked or itemized what he and other contemporary
film theorists have in mind when they speak of realism—viz., mainstream,
narrative films, i.e., mass movie entertainments, and realist films (such as
those of Soviet realism, deep-focus realism, Neo-Realism) that are not
reflexive or not modernist. Moreover, it is the narrative element of these
films that is especially central in imparting their "harmonizing effect," since
it is narrative that unifies the portrayal of events over time. Thus, it is to
150 NARRATION

these films and their narrative structures which we now turn in order to
explore the claim that realism is inherendy ideological.

ENUNCIATION
Narration, according to contemporary film theorists, has ideological
ramifications insofar as it contributes to the appearance of the harmonious
wholeness of a film. This impression of harmonious wholeness, in turn, is
ideologically suspect both for the ways in which it conspires in the produc-
tion of unified subjects, in the Althusserian-Lacanian sense, and in the
conjuring up of the politically dangerous naturalizing effect—i.e., the
impressions that a film is a sample of reality narrating itself—a prospect
whose dire consequences for contemporary theorists we discussed in the
last chapter. A crucial feature of classical realist narration, in respect of its
ideological effect, is that within this style, it is said that the narrative effaces
all marks of enunciation. That is, by effacing the marks of enunciation,
realist films, through their narrative form, induce impressions of harmo-
nious wholeness which abet the production of Althusserian-Lacanian sub-
ject unities and/or the sense that the film is reality narrating itself. In order
to get a handle on that of which realist films are here accused, some notion
of enunciation and the relation of realist narration to enunciation is neces-
sary.
Like the distinction between history and discourse, to which it can be
correlated, the notion of enunciation comes from the linguist Benveniste.
The notion of enunciation figures in the contrast that he draws between the
speaking subject (le sujet de l'énonciation) and the subject of speech (le sujet
de l'énoncé).5 Benveniste writes:

I signifies "the person who is uttering the present instance of the discourse
containing I." . . . I can only be identified by the instance of discourse that
contains it and by that alone. It has no value e x c e p t . . . in the act of speaking
in which it is uttered. There is thus a combined double instance in this
process: the instance of I as referent and the instance of the discourse contain-
ing I as the referee. The definition can now be stated precisely as: I is the
individual who utters the present instance of discourse containing the linguis-
tic instance J. 6

In order to approach the contrast Benveniste has in mind, suppose that I

say "I went to the store yesterday." The subject, "I," inside the quotation
marks is the subject of speech (le sujet de l'énoncé). The subject outside the
quotes, the I speaking the sentence in the present tense—the I say .. .—
NARRATION 151

could be seen, heuristically, as the equivalent of the speaking subject (le

sujet de l'énonciatùm). Or, to be more precise in terms of what Benveniste is
getting at, the simple utterance "I went to the store yesterday" really has
two subjects, the "I" (the subject of speech; le suject de l'énoncé) who went to
the store, and the "I" that speaks the sentence (the speaking subject; le sujet
de l'énonciatùm). The realm of enunciation is the process of making the
utterance which includes the speaking subject as well as the listener and the
context of the utterance. The utterance itself comprises the énoncé. Clearly,
every utterance will be enunciated, will herald from a speaking subject and
some context of speech, but not every utterance will call attention to its
process and context of enunciation and/or to its speaking subject.
The distinction between the énoncé, the utterance or specimen of speech,
and the enunciation can be connected with the distinction, examined in the
last chapter, between history and discourse. For some utterances manifest
their enunciation, their speaking subject, e.g., "Tell me a story" or "J like
Ike." Such utterances, as we have seen previously, are discourse. On the
other hand, certain utterances do not show the fact of enunciation on their
face; in the utterance "Something changes between 22 March and 28
December 1895," reference to the speaking subject is absent. The utterance
is an example of the technical category of history in which the language is
thought to be severed from a concrete communicative situation, the con-
text of enunciation. Benveniste says of the historic mode that "there is in
fact no longer even a narrator. The events are set forth as they occurred in
the story. No one speaks here; the events seem to recount themselves."7
And this, as we know, contemporary theorists regard as a prime means for
the dissimulation of ideology for when a communication appears as reality
narrating itself it fosters the belief that it is "true,"8 no matter how distor-
tive it actually is.
Though Benveniste himself did not develop the énoncé! énonciation and
history/discourse distinctions at great length, they have been taken up by
contemporary theorists of literature, painting, and film with vigor. How-
ever, contemporary theorists have often been very loose in expanding the
application of Benveniste's categories, using them, for example, to categor-
ize not only discrete utterances, as Benveniste does, but whole texts, and as
a result they have derived very different conclusions from it—different
from each other's as well as from what one might have predicted. Catherine
Belsey sees as a mark of the classic realist text the effacement of enunciation,
i.e., the text's own status as discourse, whereas Jenny Simonim-Grumbach
thinks all fiction is communication in the historic mode.9 And David
Bordwell has pointed out that some contemporary literary theorists have
152 NARRATION

found ways to identify elements of discourse in texts that one might

otherwise think were examples of the historic mode. 10
As one might expect, matters become even more confused when one
turns to film, since apart from dialogue and titles, such as intertitles, movies
do not have the formal machinery of grammatical personhood necessary to
make the distinctions between history and discourse and between the
énoncé and the énonciation. Thus, some take the effacement of enunciation
to be the mark of all realist film, while others find marks of enunciation in
films which one would have thought consistent expansion of the distinction
metaphorically would have categorized as belonging to the realm of the
énoncé. For example, Hollywood camera movements that call attention to
themselves are said to suddenly admit enunciation because they signal the
presence of the author in the way grammatical self-reference in a sentence
does. 11 This extension of the linguistic category of enunciation is both
strained and far fetched as well as being quite remote from the purposes for
which Benveniste introduced the concept as is Metz's notion that a film is in
the historic mode if it does not look at me and discourse if it looks at me.12
Though it would be entertaining to trace the vagaries, linguistic contor-
tions, and fanciful extrapolations by which contemporary film theorists
have attempted to impose Benveniste's categories to film, the job has
already been admirably done by David Bordwell and needs no repeating.
Contemporary film theorists often use Benveniste's categories in ways that
differ from both Benveniste and each other. 13 However, there are certain
views of enunciation that are repeated often enough that they can be
regarded as the majority opinion of the matter among contemporary film
theorists and that bear criticism.14
Generally it is held that traditional films, such as mass movie entertain-
ments, appear in the historic mode of narration since they efface all marks of
enunciation. This appears to mean that such films do not internally repre-
sent the fact that they have an enunciator—an author, director, and so on.
This gives them the attribute of appearing authorless which is one reason
spectators are said to regard them as specimens of reality narrating itself. At
the same time, the effacement of enunciation appears to involve the failure
of the film to represent other contextual factors about the process of its
production, such as the institution of cinema and the ideological determi-
nants of its making. Indeed, sometimes it is said that the film effaces the fact
that it is a production.
Of course, part of the context of enunciation is not only the speaking
subject but also the receiver of the communication. In films in the historic
mode of narration the lack of reference to enunciation somehow induces
NARRATION 153

the spectator to think of herself as the enunciator. So not only is reference

to the viewing spectator suppressed, but the spectator supposedly mis-
recognizes herself as the enunciator. Traditional film, Metz writes, "Insofar
as it abolishes all traces of the enunciation . . . succeeds in giving the
spectator the impression that he is himself the subject."15
Contemporary film theorists see a welter of ideological effects issuing
from this purported phenomenon. Seen as authorless, the film is taken for
reality narrating itself, thereby imbuing what may be ideological with the
aura of truth, transparency, and naturalness. As well, the film induces the
spectator to regard herself as the enunciator, which, presumably, not only
prompts her to accept the ideological falsehoods of the film as her own
beliefs, but confirms the Althusserian- Lacanian sense of subject unity by
placing her in the position of the "unified" enunciator of the spectacle. At
the same time, the spectacle is not really unified or complete because all
trace of the enunciation is absent. By identifying with the enunciation of the
film, the spectator fills in the missing enunciator and thereby completes the
film. This action of the Imaginary, in turn, is reinforced by the impression
that the film is now whole, thereby further encouraging the ideologically
false belief that the spectator is a unified subject. That is, the spectator by
projecting herself via identification as the missing enunciator gives the film
the illusion of being complete, whole, and harmonious, which illusion, in
consequence, reconfirms the spectator's faith in being a unified subject, an
effect which we have learned that Althusserian-Lacanians hold to be deeply
ideological.
The problems with this cluster of hypotheses are legion, and we have
encountered several of them already in our discussions of the Althusserian-
Lacanian paradigm and the notion of the transparency of the historic mode
of narration. Though contemporary film theorists proceed as though they
are straightforwardly applying well-founded distinctions of linguistic the-
ory, it is neither clear that Benveniste's categories are well-founded, nor
their application by film theorists straightforward. Of course, the question
of the adequacy of Benveniste's linguistics is really an issue for experts in
thatfield,nor need we detain ourselves over Benveniste'sfindingssince film
theorists have so changed their purport. Film theorists think that discourse
masquerades as the historic mode of narration, a prospect unknown to
Benveniste, nor does Benveniste seem to hint that I identify myself as the
enunciator when I hear a third person utterance.
Of course, the wildest expansion upon Benveniste is the notion that his
distinctions can be applied to film, which, as noted, has no system of
personhood, save in virtue of its linguistic elements, like dialogue and
154 NARRATION

intcrtitles. But o f course contemporary film theorists are not interested in

applying the distinction to film dialogue—which would contain a lot o f
utterances that are discourse—but to the visual elements o f film. In some
cases, this involves personifying the film so that it looks at me or it doesn't look
at me (Metz) in such a way that in the former case the film seems to be
interpreted as saying "I'm here," and in the latter case, "It's here." (Or
maybe "Nobody's here but these images—reality narrating itself again.")
But this is just a bit o f perfectly arbitrary word play that has nothing
determinate to do with grammatical personhood. If I thought a film wasn't
looking at me, couldn't I just as easily "translate" its address as "I'm not
looking at you," which would not only be discourse, technically speaking,
but would probably be true (of typical films, though maybe not interactive
video). And, of course, if the film does give the visual impression that is
somehow equivalent to "it doesn't watch me," then it is discourse in virtue
o f its first person reference. O f course, one might, as I do, think it theoret-
ically advisable to forget about personifying the film and attempting to
"translate" the pronominal reference o f its purported utterances. 16 But that
just might spell the end of the use o f history, discourse, and enunciation by
film theorists.
O f course, one reason to reject the notion that film images are describable
in terms of person is that grammatical personhood is a system involving
three types of indexes whereas there is no corresponding system in film.
Stretching things so that point-of-view shots and subjective images (e.g.,
dream images) are called first person images and the rest third person, one
is still left with the problem o f identifying second person images as well as
the problem that you have just stipulatively redefined the matter in such a
way that traditional realist films can be discourse and can contain discourse
insofar as they have point-of-view schemas and subjective images. More-
over, if you identify second person images as those that implicitly address
the audience, all o f the images in film will be second person. In order to
remedy this you might say that the second-person images are only those in
which the special assertiveness o f the style refers to the presence o f the
audience, but then, of course, those are exactly the images that some
theorists say bear traces o f enunciation such that one might wish to categor-
ize them as first person images. 1 7 Indeed, one might argue that every film
image except maybe some shots into mirrors are third person. But, of
course, the point here is not to argue about the person o f this or that image
but rather to indicate that there is no determinate way to lay the system o f
personhood on the cinematic image track and thus little theoretical profit or
precision to be derived from attempting it.
NARRATION 155

Moreover, if we use the linguistic elements of traditional films—the

dialogue and the intertitles—to determine the person of saidfilms,then we
will find first, second, and third person utterances, but this will destroy the
thesis that traditional realist films are all in the historic mode. Not only is
much dialogue that is discourse in evidence in realist films, but, as was
noted in the previous chapter, there are numerous examples of traditional
films, such as The Cabinet of Dr. Caligari, that are narrated in the first
person. Furthermore, if we determine whether or not a film is in the
historic mode on the basis of whether it bears traces of enunciation, then all
films with beginning and/or end titles should count as discourse.
Our objections so far have hinged upon the inadequacy of distinctions
based on grammatical personhood to do the work the contemporary film
theorist wants done. However, the contemporary film theorist may wish to
recast the notion of the effacement of enunciation in a way that makes this
process literally independent of issues of grammatical personhood. Drop-
ping the linguistic window dressings, the contemporary theorist might
wish to say that realist works efface their enunciation by masking the fact
that they are productions or constructions. That is, traditional films impart
the impression of transparency by effacing reference to their conditions of
production, thereby appearing authorless, and, consequendy, without mo-
tive, interest, or point of view, ideological or otherwise. Taken literally, this
scenario is false for the kinds of reasons discussed in the last chapter. We are
always aware that representations are the products of human agency; a
necessary condition tor the cognitive assimilation of a representation, in-
cluding a realist film, is the recognition that it is a human product. If we fail
to grasp the ideological implications of afilmit cannot be because we take it
to be authorless. A work that appeared authorless—though it is hard to
know how it might do so—would probably call more rather than less
attention to itself.18
But perhaps what is meant by the suppression of the fact of production in
this context is not the effacement of the artifactuality or the production
history of the film but rather some implicit denial or masking of the
spectator's interpretive response. That is, the contemporary film theorist
may hold that it is certain factors of the process of the production of
meaning in the reception of realistfilmsthat are the facts of production that
the realist film represses. A communicative exchange requires an active,
interpretive response from the spectator. In order for afilmto be processed,
the spectator must do such things as comprehend conventions, supply
presuppositions, and engage in a range of interpretive acts. The realist film,
it might be claimed, encourages a spectator in the impression that this
156 NARRATION

interpretive activity is not occurring, and that the meaning of the work is
"self-sufficiently" reflected in the representation. The impression of trans-
parency here is that the communicated meaning of the film somehow is
completely internal to the work, ready to yield itself automatically.
That the audience is unaware of its interpretive activity is implausible. In
contemporary American society it is unfortunately a commonplace that
everyone has his or her own interpretation of everything, especially of
representations such as movies. Americans, with their exorbitant claims to
personal interpretations of every domain of life, err in a direction that is
exactly opposite that which contemporary film theorists attribute to
them. 19 It is outlandish to claim that ordinary consumers of representa-
tions, such as realist films, are not aware that interpretations are involved in
responding to them—ordinary spectators are the first to admit it, almost
nihilistically—while it is also apparent that everyone knows that a com-
pleted communication requires a receiver. Nor is it easy to see what is
involved in the supposed impression that the communicative meaning of a
work is self-sufficiently contained in it. The notion is so obscure that
contemporary film theorists should refrain from ascribing it to anyone.20
Contemporary film theorists hold that by effacing enunciation, realist
representations appear transparent, which, among other things, make it
seem as though they are not ideologically motivated. That is, by effacing
enunciation they appear transparent in that they do not appear ideologi-
cally colored. Obviously the story about the way in which a realist represen-
tation appears authorless—that is, author-transparent—is meant to sug-
gest a causal explanation of this ideological transparency. If a work appears
uncolored by an authorial viewpoint—i.e., appears authorless—then what
would an ideological viewpoint be attached to? We have, in both this
chapter and the previous one, dismissed the idea that a realist representa-
tion of film appears authorless. However, even if this hypothesis is dis-
carded, it still might be the case that an effect of realist representations and
realist films, qua their representational and narrational apparatus, is to
promote the impression that they always lack ideological motivations.
Undoubtedly, once we reject the causal hypothesis that in realist repre-
sentation there is an impression of authorlessness, which calls forth an aura
of ideological neutrality, it becomes hard to understand how and why
simply by being a realist representation, something could give the impres-
sion that it was bereft of ideological affiliation, that is, where it does have
said affiliations. Aren't there works of realist representation of unmistakable
ideological affiliation—West's The Death of General Wolfe, Gros's Napoleon
atArcole, Huxley's Brave New World, Riefenstahl's Triumph of the Will, and
NARRATION 157

Wayne's Green Berets. If realism erases obvious ideological tincture, what

has gone wrong in these cases?
Or perhaps the contemporary theorist believes audiences do regard these
works as ideologically transparent. But what would we say of someone who
claimed that these realist representations appeared to lack an ideological
viewpoint? I conjecture that either such a person really does not understand
what these works are about or that, more likely, such a person embraces,
before encountering these respective representations, the ideological posi-
tions found in these works. That is, such a person probably takes the
ideological messages in question to be true rather than ideological—on the
basis of antecedent beliefs—and he regards the works as illustrations of
what, for reasons external to the work, he was already predisposed to accept
as an ideologically untainted fact. And certainly no one who antecedently
rejects the ideological position inscribed in the previous examples will be
prone to accept such representations as ideologically neutral.
If we want an account of why some people believe ideologically skewed
representations, we would, I think, be better advised to attend to the
antecedent beliefs such people have rather than to some special power—
like transparency—that realist representation is thought to automatically
acquire by effacing enunciation. I am not denying that representation,
including film, plays a part in the educational process through which people
develop beliefs. Rather I am suggesting that it will be the examination of
concrete cases of educational processes over time which will deliver ac-
counts of why certain ideological messages are not regarded as such by
ordinary spectators. That is, the history of the inculcation of certain beliefs
instead of the investigation of the structure of representation—in terms of a
quasi-formal effect like transparency—will be the most promising avenue
of research.21
The preceding argument advanced by means of examples of representa-
tions whose ideological commitments are quite overt. But the contempo-
rary film theorist might claim that it is not those sorts of works that notions
of transparency and the effacement of enunciation are meant to explicate.
Rather, we should think of more quotidian and covert examples such as TV
police shows, advertisements, and honor films.
First, contra many contemporary theorists, I should say that I am not
persuaded that every popular representation conveys a message that is
ideological, partly because it is unclear that every one of them contains a
message and partly because a suitably constrained concept of ideology may
not be applicable to all of those remaining representations which can be
said to have messages. Furthermore, I think, it is just false to suppose that
158 NARRATION

ordinary viewers are lulled by some transparency effect into believing that
advertisements arc not motivated by interests which the viewer easilv
identifies. Nor do I think that viewers are automatically blinded to the
ideological function of the police shows and doctors' shows that reassure us
that we are in the care of public servants who are incredibly committed and
tremendously talented.
But if viewers fail to analyze these messages it may be attributable to a
combination of the sort of complacency engendered by being bombarded
by a surfeit of such messages and by the fact that viewers bracket such
exercises under the rubric of entertainment, a category of things not to be
taken seriously. Psychological and sociological study of both the causes and
effects of complacency and of prevailing notions of entertainment, then,
seem better starting points for investigation here than the examination of
the structure of realist representation and narration as such—that is, if we
want to study the ideological effectiveness of popular culture.
Of course, it may be the case that many viewers do not recognize the
ideologically skewed representations found in popular narratives. But again
this may have less to do with some transparency effect or effaced enuncia-
tion than with the fact that viewers already accept the ideology embodied in
such a narrative. For example, in current horror films, like Halloween,
sexually active women are often punished by gruesome executions. If
audiences fail to recognize the ideological significance of this, it is probably
because they accept these executions as somehow morally appropriate.
I am not denying that a spectator may acquire an ideological belief in the
course of viewing a film narrative. However, this appears to occur most
often when an ideological conviction is presented by a character whom the
spectator already admires or approves of because of that character's not
necessarily ideological virtues—strength, cunning, cool, beauty, courage,
fortitude, wit, expertise, honesty, knowledge, civility, etc. 22 But this is not
an effect of realist narration as such. It is a persuasive technique available to
high-school football coaches and filmmakers alike. And there is no reason
to believe that this rhetorical strategy must be placed in the service of
ideology in every realist narrative. Nor need this rhetorical strategy be
limited to cases of the historical mode of narration.
Whereas contemporary film theorists, by means of notions such as the
effacement of enunciation, wish to insist that narration is inherently ideo-
logical, I want to stress that narration as such is ideologically neutral. This
by no means denies that there are certain narrative films that are ideological
nor that there are certain narrative devices that are often used to promote
ideology. For example, that women are dependent, that they want to be
NARRATION 159

raped and dominated by men, is a theme of Lina Wertmuller's Swept Away.

The film seems to propound this view as an eternal law about the nature of
women. But the sense that this is an eternal truth in this film is not projected
by the effacement of enunciation but is secured by the deserted island
device, a trope which at least since Robinson Crusoe has been used to portray
how people, supposedly, really are. The thinking behind this device is
crudely experimental; it assumes, pace Aristode, that human nature as it
really is will only be revealed once people are removed from the network of
social relations and restricted in a lonely laboratory. A similar device for
abstracting human nature is the atomic war which destroys society leaving
us to, well, live naturally. Pulp novels like Day ofthe Triffids, The Stand, and
Lucifer's Hammer, and films like Night of the Living Dead, Parasite, and the
Mad Max series are examples of this. We can analyze these works profitably
in terms of the way in which the-end-of-the-world device presents some
politically significant feature of human life as a fact of human nature. But
this device is not an example of narrative as such but rather is one type of
plot, which, by the way, could have a first person narrator as does Road
Warrior. That is, rather than supposing that all realist narration has some
ideological effect as a result of its putative effacement of enunciation, we
can note that ideology is often projected, and not always successfully, in
specific films through the use of certain recurring rhetorical strategies and
plot devices. These strategies need not always be employed to organize
ideological material, but they may be so deployed frequently enough that it
would be useful for film theorists to attempt to isolate and enumerate them.
Admitting this, however, does not commit us to agreeing that realist
narration as such is ideological.
Contemporary film theorists also believe that the purported effacement
of enunciation in realist narrative as well has the effect of instilling in
spectators the belief that they are unified subjects. The film effaces enuncia-
tion in such a way that the spectator identifies himself as the enunciator,
which gives the film a sense of completeness or harmonious wholeness
which positions the spectator as a unified subject. We shall take up the
notion that the narration positions the film spectator in the next section.
However, before concluding this section, it is useful to ask whether the first
step in this account seems plausible. Namely, do I, when watching a film in
which enunciation has been effaced, regard myself as the enunciator?
The problems with an affirmative answer to this question parallel certain
of the problems that arose with respect to camera identification in the last
chapter. How, if I regarded myself as the enunciator of the film could I
square this with the fact that many of our typical responses to things like
160 NARRATION

suspense films, mystery films, and horror films depend on the fact that I
don't know what is going to happen. Moreover, with these experiences, I
know that I don't know what the next reel will bring. It is my awareness
that I don't know whether the heroine will go over the waterfall that primes
my feeling of suspense. Also, I am surprised by what happens in films, and
sometimes shocked. Clearly, characteristic responses to films in terms of
suspense, mystery, shock, surprise, and so on presuppose that I do not have
the knowledge that the enunciator has, and that I am often aware of this. So
the problem is this: how can I be aware that I do not have all the knowledge
regarding the story that the enunciator has, and at the same time think I am
the enunciator?
Perhaps at this point the notion of disavowal will be bruited about. But
how will this work? Will I disavow my lack of knowledge about the
denouement of a suspense scene? Is it even logically possible? It is at least
hard to know how one could do this, if I really don't know the outcome of
the scene, and, as well, if one did, would the scene remain suspenseful?
Furthermore, whatever is involved in the idea of my identification with the
enunciator is very vague. I don't leave a screening of The Paradine Case,
saying "Hi, I'm Alfred Hitchcock," nor do I berate myself during or after
seeing a boring or ridiculous film. Perhaps the idea that spectators mis-
recognize themselves as the enunciators of realist films is an idea whose time
has passed.

T H E I N T E R N A L S T R U C T U R E AND F U N C T I O N
OF NARRATIVE
For contemporary film theorists, the narrative element in what are
thought of as realist films is a crucial element in the positioning of specta-
tors as unified subjects, in the Althusserian-Lacanian sense. Narrative is said
to be able to perform this function because it moves from a state of
equilibrium to one of disequilibrium to one of equilibrium, which final
state promotes the feeling of homogeneity which is used as the basis for a
sense of subject unity. Leading contemporary film theorist Stephen Heath
writes: "Narrativization is scene and movement, balance (with genres as
specific instances of equilibrium)—for homogeneity, containment." 23
Here, it is probably best if we have in the forefront of our thinking a movie,
that is, a popular mass entertainment film of the type that dominates the
international, commercial cinema. Such narrative movies, according to
contemporary film theorists like Heath, begin with a state of equilibrium
which is disrupted—E.T. and his friends are interrupted by earthlings in
NARRATION 161

such a way that E.T. is left stranded—and which original state is finally
restored after the plot complications have unraveled—E.T. rejoins his own
kind; E.T. goes home. The plot of a typical narrative movie is a set of
transformations which operate on a series of disequilibriums with the net
result that some sort of narrative equilibrium emerges. Heath writes:
A narrative action is a series of elements held in a relation of transformation
such that their consecution—the movement of the transformation from the
ones to the others—determines a state S' different to an initial state S. Clearly
the action includes S and S' that it specifies as such—beginning and end are
grasped from this action, within the relations it sustains; the fiction of the film
is its "unity," that of the narrative. A beginning, therefore, is always a
violence, the interruption of the homogeneity of S (once again, the homoge-
neity—S itself—being recognized in retrospect from that violence, that
interruption); in Touch ofEvil this is literal: the explosion of a bomb-planted
car, killing two passengers. The task of the narrative—the point of the
transformation—is to resolve the violence, to replace it in a new homogene-
ity. "Replace" here, it must be noted, has a double edge: on the one hand, the
narrative produces something new, replaces S with S'; on the other, this
production is the return of the same, S' replaces S is the reinvestment of its
elements. Hence, the constraint of the need for exhaustion, the requirement
of practicability: every element must be used up in the resolution, the disper-
sion provoked by the violence must be turned into a reconvergence. Ideally, a
narrative would be the perfect symmetry of this movement: the kiss the
explosion postpones is resumed in the kiss of the close of the film as Susan is
reunited with Vargas—the same kiss but delayed, set into a narrative.24

One of Heath's presiding metaphors in describing the work of movie

narrative is "getting things back into place." In the beginning of Touch of
Evil, the leading female character is morally (that is, sexually) untarnished.
The "disrupting" complications of the plot throw her reputation into
question. The conclusion reinstates her virtue. Heath does not claim that
every detail presented in a plot does in fact play into this transformation-
substitution nexus, but only that a realistically narrated plot gives the
impression (illusion) of this sort of homogeneity, and that pleasure in
narratives is grounded in the tension between equilibrium and disequilibri-
um (homogeneity and heterogeneity) in which homogeneity ultimately
wins out. The transformation-substitution process (which governs closure)
plus the narrative's consistency (presumably its logical consistency and its
internal cause-and-effect plausibility), plus its economy (the gun that ap-
pears in the first scene goes off in the last scene) give the narrative movie its
unity, a Althusserian-Lacanian homogeneity in which the moving picture
presents itself in the totalized form of a narrative image, "a kind of static
162 NARRATION

portrait in which it comes together."25 The repetitions inherent in the

transformation-substitution process (and the logic of other aspects of nar-
rative structure) imbue the movie as a whole with the appearance of
homeostasis, the illusion of a static (unchanging) equilibrium in the service
of an impression of homogeneity, wholeness, and plenitude.

In its films, classic cinema is a certain balance of repetition: a movement of

difference and the achievement in that movement of recurrent images—for
example, the woman as "the same," a unity constandv refound. Narrativiza-
tion, the process of the production of the film as narrative, is the operation of
the balance, tying up the multiple elements—the whole festival of potential
affects, rhythms, intensities, times, differences—into a line of coherence
(advances and recall), a finality for the repetition.2*

The narrative of a movie involves a play of repetition and difference, of

equilibrium and disequilibrium. Pleasure with the film is generated in this
play. But ultimately the economy of the movie narrative is rooted in
equilibrium and that which has been disrupted or thrown off balance is
finally restored. This returns the film to a state of balance or equipoise
which confirms the spectator's illusory position in relation to the address as
a unified, centered subject which ideology requires.
Film, of course, does not center or position the spectator as a unified
subject merely by means of narrative. As we learned in the previous chapter,
perspective also is a crucial element in this process. Perspective and narra-
tive work in concert to position or center the spectator as a unified subject.
Moreover, the nature of the relationship of perspective and narrative in film
is quite special; film narrative can be said to stabilize cinematic perspective
in a way that assures that the subject positioning brought about by perspec-
tive will not be undone by the movement of film.
That is, Heath and his followers believe that the operation of pictorial
perspective in film faces certain formal problems that do not exist with
respect to the functioning of perspective in painting. These problems are a
result of the fact that there is movement of character and camera in film.
This movement putatively provokes a crisis within the perspective system,
since, for example, with every cut the subject's position is altered. How
does the dominant, realist cinema camouflage this so that the spectator does
not become apprised of the disunity of subject positions from shot to shot?
The answer is to be found in narrative.
Each perspectively correct film image positions the spectator. But the
ideological effect of perspective in film is threatened by cinematic move-
ment. For Heath and his followers, character movement, camera move-
NARRATION 163

ment, and editing potentially imperil the subject position constructed by

the film image through perspective. This assumed, the question, then, is
how in the dominant, realist cinema, whose ideological task is to position
subjects, the filmic system is able to sustain the perspective effect, despite
the potentially conflicting disruptions of cinematic movement? Here narra-
tive is invoked. The coherence o f a narrative positions the subject and
maintains the perspective illusion by centering the subject through diegetic
intelligibility, equilibrium, and closure. The narrative is able to do this
because, like perspective, narrative is a system that places and centers
subjects. Perspective and narrative are not only coordinated systems but
also systems that independendy have the same effect. If and when the
perspective system begins to flag, the same effect is imposed on the specta-
tor by the story. We might think of perspective and narrative as two pumps,
calibrated to bail out a set volume of water. When the perspective pump
sputters, the narrative pump makes up the difference, though the less
reliable perspective pump is still an irreplaceable asset to the system insofar
as the given volume o f water could not be evacuated without it. Heath
writes:

Narrative contains the mobility that could threaten the clarity of vision in a
constant renewal of perspective; space becomes place—narrative as the tak-
ing place of film—in a movement which is no more than the fulfillment o f the
Renaissance impetus. . . . what enters cinema is a logic of movement [human
action] and it is this logic that, in other words, is constructed as narrative
space. 27

Once films are films o f people, the spectator is centered by the logic o f
human action, thus compensating for the distractions o f human movement
in the perspectival array. For Heath, it seems that the presence of people in
a film image is a sufficient condition for considering the image narrative
(though this does seem an overly broad notion of narrative). Moreover, the
narrative of such an image stabilizes the subject position. As shots are added
to shots through editing or as the camera moves, the subject center is
sustained by the coherence of the story as an intelligible representation o f
human action which tends toward equilibrating and balancing disruptions
whether perceptual or diegetic.
So far, then, we have isolated at least these three hypotheses about the
internal structure and function of film narrative: that film narrative is an
equilibrating system, based on the process o f transformation and substitu-
tion; that this equilibrating system stabilizes the potential disruption of the
centering work of perspective; and that the narrative system, by means of its
164 NARRATION

tendency toward balance, along with the perspective system, functions to

center or to position the spectator as a unified subject in the Althusserian-
Lacanian sense of that term. How plausible are these hypotheses?
Movie narration gives the film an aura of wholeness or homogeneity,
which aura is relevant to the spectator's misapprehension of himself as a
unified subject. The impression of unity of the film, in virtue of its narrative
structure, gets translated into subject unity. Every sort of narrative unity-
making device is credited with the same effect as the replacement-transfor-
mation-repetition-equilibrium scheme. The strongest case for the unity or
homogeneity instilling capacity of classical realist narratives is made by the
notion that movie narratives are machines for replacing disruptions of an
initial state of affairs by repetitions of the initial state of affairs—E.T. with
his ship; E.T. separated from his ship; E.T. back with his ship. However,
the connection between this kind of plot structure and a feeling of whole-
ness, homogeneity, and unity on the part of the spectator may be based
merely on the blurring of important distinctions.
It might appear that the citation of this sort of plot repetition warrants
the postulation of a response in terms of wholeness, homogeneity, and
unity because this kind of repetition is often called "circular" by literary
critics, and in the same (rather free) associative mood we might call it
"static," which would tempt us to "unchanging," which might seem to be
enough to argue that the causal conditions are in place for us to be infected
by a fantasy of homogeneity, an impression of undifferentiated wholeness
and plenitude.
However, such an explication is easily stopped by pointing out that
spectators who are entertained by circular structures do not literally believe
they have witnessed an unchanging or static image, nor do circular narra-
tives have a special potential for evoking impressions of wholeness or
plenitude. Rather, circular, static, and unchanging are really critical meta-
phors that describe certain patterns of literal change in traditional realist
narratives.
Let us, however, suppose that there is an argument for a connection
between narrative repetition and the spectator's purported impression of
homogeneity and subject unity. It must then be pointed out that the type of
repetitive device Heath has in mind is only one type of narrative ending. A
narrative may also end in the reversal of an original state of affairs, for
example, Bringing Up Baby or the remake of Invasion of the Body Snatchers.
Or a narrative may end in an event which stands in merely a consequent
(and neither a symmetrical or anti-symmetrical) relation to earlier narrative
states, for example, Nashville or Kiss Me Deadly. Nor need narratives be
comprised of scenes that vary earlier scenes either as repetitions of motifs or
NARRATION 165

as subplots. Thus, even if narrative repetition could be connected with the

appropriate form of homogeneity, it would account for only one type of
narrative device and an optional one at that.
Moreover, there are dimensions of narrative unity in movies that are not
dependent upon the repetition or recapitulation of scenes. A film's narrative
logic (the structure that organizes what follows what), its narrative econ-
omy, the internal plausibility of the film's causal relations, and the film's
consistency are each unity-making features, though none of them requires
repetitions or variations of earlier scenes in order to operate. Narrative
economy does demand repetition of certain objects and characters, but not,
however, the repetition of scenes. Heath treats all the unity making features
of movie narrative as if they were the same, and he speaks as if they all
produced the response in terms of wholeness, homogeneity, and unity that
he attributes to narrative repetitions. But this is only the result of his
confusing all of narrative's different structural unities with the type of unity
he associates with the transformation-replacement model. Where Heath
should be differentiating strata of effect, he is congealing them. Further-
more, Heath cannot argue that other increments of narrative coherence, for
example, consistency, can be explicated by the repetition effect on the
grounds that these other devices are always subservient to the ideal of
narrative symmetry. Narrative logic, economy, plausibility, and consistency
can and do operate in films that do not have scene repetitions and symme-
tries. These unity-making features are quite independent and distinct from
the symmetries of the transformation-replacement schema and should be
analyzed individually.
Contemporary film theorists have attempted to use the equilibrium
model as a general model for analyzing the dominant film narrative of
realist movies, but it is a blunt instrument. 28 It is not the case that all movie
narratives begin with a state of affairs that is restored at the end. A great
many films begin with an untoward state of affairs—the goodly townfolk
savaged by the cattle ranchers—and end with its reversal—the cowboys
vanquished by the lone gunslinger turned avenging angel. If there is any-
thing that we could profitably call an initial equilibrium here, it preexists
the film either as a golden age or as an ideal of law and order. Nor need a
film begin with an untoward state of affairs that the film proceeds to
correct: a character may just be unmarried, for example, only to be wed by
the ending. Calling this a transformation-repetition is simply wrong, as is
saying that the film must begin with some equilibrium. It begins. It begins
with some state of affairs or some event which may be restored (if only
symbolically), reversed, or merely forgotten.
Nor, where the film begins with a state of affairs, is it accurate to say it is
166 NARRATION

always disturbed or disrupted by violence.29 This may be a dramatic mode

of speech, but films do not require disturbances of equilibriums for the
narrative to proceed; they require changes which may be neither disruptive
nor violent: the male romantic lead notices that the girl next door is
suddenly "all-growed-up," or a billionaire hears so and so sing and decides
to bankroll a show. Also, a film may end without restoring either initial or
ideal states of affairs: e.g., King Kong or You Only Live Once.
Of course, a researcher may take it to be definitional that a film narrative
in the dominant realist tradition of movie ends in a balanced state or
equilibrium and then go on to take any type of closure as a case of equi-
librium. But then equilibrium no longer has any explanatory power; it
becomes just another word for the end (and, in this theory, for the begin-
ning). Butch Cassidy and the Sundance Kid achieves closure, but not as a
state of equilibrium generated by repetition. If we call its ending an equi-
librium then that term is merely synonymous with closure, the very thing
the equilibrium model supposedly explicates.30
The transformation-substitution model of realist film narration, as
stated, allows a certain ambiguity on the issue of whether the balance
achieved in closure is formal (the repetition of an initial scene) or moral (the
restoration of a preferred social order). In Heath's example of Touch of Evil,
the logic of narrative and the law of patriarchy converge in their effect; both
determine the restoration of the woman's virtue. And certainly the struc-
ture of the narrative can reinforce the moral of a film in this way.31 But
these two elements of a narrative film are very distinguishable and may not
be synchronized to achieve such harmonious effects.
A film, for example, like The Circus, can begin and end with moral
disequilibrium while formally repeating the tramp's initial condition in the
conclusion; or a movie such as Dead of Night or Invaders from Mars (either
version) can begin in moral order and proceed to moral disorder while the
beginning formally mirrors the end; and, of course, there is the very
common case of movement from moral disorder to order without a formal
structure of repetition. This last case can become especially tricky for
researchers committed to the equilibrium model because they are tempted
to takefilmsthat restore some moral order (rather than some specific scene)
as confirmations of the theory when literally they are not.
Heath also writes as if the transformation-replacement-repetition sche-
ma were the same as plot resolution when the former is merely one form of
the latter. This obfuscation of class and subclass may prompt some theorists
to look always to the end of the film in order to explain its structure, finding
there some repetition of or similarity to the beginning. With enough
cleverness, and with broad enough notions of repetition and similarity.
NARRATION 167

success is almost assured. But this enterprise is likely to miss the point of the
actual plot resolution for the sake of stressing some repetition. You could
say that Annie opens and closes with the titular juvenile singing, but this
misses the thrust of the narrative climax: she is now rich and has a daddy
whereas she started poor and an orphan. If you respond that the film has
reproduced the ideologically predictable moral order of things, you are
probably right, but you have switched from a discussion of formal order
based on repetition to a discussion of an ideal ethical equilibrium.
The transformation-replacement-repetition schema of the internal struc-
ture of movie narrative is not refined enough to deal with the data. To make
it appear to work, one must either inflate its central concepts so that they
will fit anything, or bend the evidence to fit the preordained scheme, or a
litde bit of both. From the viewpoint of advancing our understanding of
the narrative operation of film, the equilibrium model, in short, is thor-
oughly inadequate.
The repetition schema of movie narrative is thought to participate in the
positioning of subjects in the Althusserian-Lacanian sense of the unified
subject. Such narratives are, for example, thought of as "centering" sub-
jects. However, one must ask here whether or not the ostensible intel-
ligibility of the theory is based on much more than punning. Something
that is balanced or in equilibrium, like a teeter-totter, often has a literal
center; so one might think that a balanced film has a center in virtue of
which it positions or centers subjects. But even if balanced plots have
centers, it is not because of their centers that plots are equilibrated. That is,
in afilmwith twenty scenes, scene number ten may be the literal center, but
that need have nothing to do with the fact that the plot is balanced; the
main character, for example, wanted a wife, overcame obstacles, and got
married; there is no point that is the center of this process in the sense
Althusserian-Lacanians require. That a plot reaches equilibrium does not
provide the special connection with centering that film theorists presup-
pose. There is no special relation between plot equilibrium and centering,
though talk of "equilibrium" and "balance" may confusedly suggest one.
However, the theory discussed earlier, that narrative functions to stabil-
ize the centering of the subject brought about by perspective, may appear
to supply the requisite connection between the narrative and a center.
Supposedly, movement in cinema threatens the centering of the subject
secured by perspective, but this prospect is counterbalanced by narrative;
the intelligibility of narrative masks the disruption of the perspectival center
of the cinematic image. But this theory rests on the dubious presupposition
that the perspective effect is somehow endangered by cinematic movement.
There is no evidence that movement in film causes a disruption of the
168 NARRATION

perceptual work done by perspective, however we characterize that work.

There is no evidence, for example, that editing, at standard rates, causes anv
perceptual confusion or unclarity. Nor is this lack of evidence due to the
fact that most people watch narratively stabilized films. With regard to
editing, for example, the final row of replacement images in Frampton's
Zorns Lemma is easily comprehended, in terms of its visual content, by
normal viewers despite the renewal of the station point from shot to shot.
Film viewers will sometimes report confusion or disturbance if certain
kinds of shots are edited together too rapidly. The effect may be to animate
the contours in these shots and to leave the impression of tumult in the
images. Yet, this is not caused by the renewal of station points but by the
appearance of motion generated by stroboscopic or beta phenomena.
Moreover, this type of visual disturbance can occur whether a film has a
narrative or not or whether its images are perspectival. So this type of
disturbance does not provide the requisite evidence. Insofar as there is no
evidence that perspective is imperiled by the types of cinematic movement
described earlier, the theory is mistaken. One cannot suppose that the
narrative is solving the problem of the disturbance of perspective by means
of compensating for the viewer's compromised center. For where there is
no problem, there is no solution. That is, the theory is predicated on
imputing a function to narrative in relation to perspective. But the motive
for attributing this function to narrative and, therefore, the function itself
are groundless.
One might attempt to meet this objection by saying that I am miscon-
ceiving the problem as a perceptual one, and that this misleads me to look
for perceptual evidence of the putative disturbances of perspective. A
contemporary film theorist might prefer to say that the postulated distur-
bance is of the subject's position as a psychic center and not any disturbance
detectable by the naked eye in the visual field. But shouldn't something as
conditioned by vision as the subject center of perspective yield some cor-
relative disturbance in perception when it is disrupted? That is, how can
this subject center of vision be dislodged by occurrences like the renewal of
station points without any trace of that eruption showing up in the normal
viewing experiences of, for example, a succession of perspectivally correct,
nonnarrative images (such as Bruce Conner's Valse Triste)? The theory of
the narrative stabilization of perspective constructs a tier of spectator illu-
sions which originate in perception—the perception of perspective and its
disturbance—but which are not supported by the data of film perception.
We are given subject centers that are both visual and metapsychologi-
cal/metaphysical entities, and we are given their putative interrelations, but
NARRATION 169

at the ground level of vision we have no evidence for the initiating distur-
bance which purportedly sets the illusion machine in motion. And as we
will see in the next chapter in our discussion of suture, this is a recurring,
though intellectually befuddled, theoretical strategy in contemporary film
analysis—to erect a psychological theory where there are absolutely no data
to motivate the postulation of specific psychological mechanisms nor any
data to constrain the elaboration of the details of the mechanism.
Apart from the question of whether film narratives stabilize perspective,
contemporary film theory makes another substantive claim, one holding
that perspective and narrative are systems that serve the same purpose by
producing the same effect, namely centering. But this basically proceeds by
describing the effects of perspective and the various effects of narrative by
means of the same terms; both are said to place, position, and center
spectators in such a way that spectators acquire the illusion that they are
unified subject centers.32 But if perspective imparts a central position
illusion to spectators, it must be granted that this is putatively in some sense
a spatial illusion—an illusion of occupying a literal place in regard to which
the geographic layout and structure of the image is perspicuous. But when
we read that the narrative places, positions, and centers subjects by means
of narrative logic, coherence, and meaning, we note that though the terms
are spatial, the illusions they purportedly describe are those of intel-
ligibility, homogeneity, and unity, and not necessarily of space. So there is
no reason to suppose that perspective and narration perform equivalent or
even comparable functions. Moreover, when spectator-subjects are de-
scribed as centered or positioned, these terms have yet further senses, ones
connected to our feelings of being unified and stable. The putative causal
connection between the centered image and the centered narrative, and
then the centered subject, is really nothing more than a function of ambigu-
ously describing the three different types of phenomena by the same words
and presuming that this shows them to be causally interrelated. But this is
patendy a well-known fallacy of informal reasoning. 33
Of course, the contemporary film theorist may want to argue that narra-
tive does its work in producing unified subjects through its impression of
coherence which triggers a Lacanian type of psychic mechanism into con-
juring up a sense of wholeness and subject unity. The effect of the realist
film is said to be wholeness and that impression of wholeness both appro-
priates and reinforces the sense of subject unity. Here, one supposes, that
the coherence of the narrative calls into play at least the psychic mechanism
of the Imaginary. But this raises questions of a sort we have already
encountered. Namely, how can the Imaginary's purported response to its
170 NARRATION

coherent image in the looking glass be extrapolated as the model of its

response to the coherence of a narrative? The causes are radically different;
they would not seem likely to produce the same effect. Indeed, the onlv
thing they have in common is that they are described by the same word,
viz., "coherent," though, in truth, that word has different criteria of ap-
plication when used with reference to narratives and bodies. So even if
there is some Imaginary sense of wholeness, rooted in our supposed mirror
stage, there is no reason to attribute that as an effect of narrative, save by
way of equivocation.

AN A L T E R N A T I V E A C C O U N T O F M O V I E NARRATION
Contemporary film theory is concerned with explaining many effects of
film narration—such as the stabilization of narration and the transparency
effect—which I have denied are plausible effects to attribute to film narra-
tion. However, the repetition-transformation model is a legitimate, if
failed, theoretical endeavor. That is, it certainly is theoretically fair game to
attempt to describe the general structure of film narration. Thus, in order to
persuasively discard that model, I will attempt to offer a competing model
whose handling of the data is more comprehensive and precise.
However, before setting out my competing viewpoint, I must stress that
my theory is not one of film narration but more specifically of movie
narration. I wish to characterize the narrative structure of popular movies
and TV shows rather than the narrative structures of modernist master-
pieces like Last Tear At Marienbad or Duras's films. I do think that a theory
of movie narration may have interesting consequences for what we say of
other varieties offilmnarration. But my theory pertains, first and foremost,
to movie narration, to the process by which popular films and (narrative)
TV shows render sequences of events coherent.
The approach to movie narration that I advocate was suggested by
Pudovkin.34 It has not been developed further byfilmscholars. Perhaps the
reason for this is that Pudovkin's tone is prescriptive; he is telling prospec-
tive filmmakers how they ought to construct films; and such overt polem-
icism is often shunned by scholars. However, it is important to remember
that Pudovkin, like Kuleshov,35 was involved in distilling and conceptualiz-
ing central elements of what they observed in American cinema, e.g.,
Kuleshov's points about the importance for rapid editing of succinct,
uncluttered set design. Kuleshov and Pudovkin, of course, were motivated
by practical concerns. They studied American films, contrasting them with
Russian films, in order to discern what made the American films of the
NARRATION 171

twenties more effective with popular audiences than comparable Russian

films. And they were interested in this matter because they wished to forge a
Soviet cinema that would be a powerful mass art. In their investigation of
American films, the very paradigms of what I'm calling movies, Pudovkin
and Kuleshov identified, clarified, and articulated certain latent, stylistic
principles that seemed to determine and were exhibited by American film
practice. As recommendations about the way in which cinema should be,
Kuleshov's and Pudovkin's theories are surely open to question. Neverthe-
less, when examined for their insightful and accurate crystallizations of the
tendencies inherent in existing film practices, their remarks, reconceived as
observations, are quite impressive and useful. Specifically, I believe that
Pudovkin's analysis of the approach to editing together or connecting
scenes supplies us with a starting point for outlining the structure of one of
the basic, if not the most basic, linear narrative forms in the history of the
movies.36
A movie will portray a sequence of scenes or events, some appearing
earlier, some later. A practical problem that confronts the filmmaker is the
way in which these scenes are to be connected, i.e., what sort of relation do
the earlier scenes bear to the later ones. Pudovkin suggests that the relation
of earlier scenes and events in a film narrative to later scenes and events can
be generally understood on the model of the relation of a question to an
answer. One can grasp this by recalling primitive, two-shot narratives. In
the first shot, a child might be kidnapped. This raises the question: "Will
the child be saved or not?" The next and last shot answers the question; the
police apprehend a racially stereotyped Eastern or Southern European, and
the child is rescued. The basic narrative connective—the rhetorical bond
between the two scenes—is the question/answer.
The importance of this relation, which I call "erotetic," can be seen in
films more complex than two-shot narratives. If a giant shark appears
offshore, unbeknownst to the local authorities, and begins to ravage lonely
swimmers, this scene or series of scenes (or this event or series of events)
raises the question of whether the shark will ever be discovered. This
question is likely to be answered in some later scene when someone figures
out why all those swimmers are missing. At that point, when it is learned
that the shark is very, very powerful and nasty to boot, the question arises
about whether it can be destroyed or driven away. The ensuing events in
the film serve to answer that question. Or, if some atomic bombs are
skyjacked in the opening scenes, this generates questions about who stole
them and for what purposes. Once the generally nefarious purposes of the
hijacking are established, the question arises whether these treacherous
172 NARRATION

intents can be thwarted. Or, for a slightly more complicated scenario,

shortly after a jumbo jet takes off we learn that the entire crew has just died
from food poisoning while also learning that the couple in first class is
estranged. These scenes raise the questions of whether the plane will crash
and whether the couple in first class will be reconciled by their common
ordeal. Maybe we also ask if the alcoholic priest in coach will find God
again. It is the function of the later scenes in the film to answer these
questions.
Since most film narratives involve a series of actions, it may seem natural
to think that causation is the major connective between scenes in movies.
However, it is implausible to suggest that scenes follow each other in most
film narratives via a chain of causal entailments. I would guess that most
succeeding narrative scenes are causally underdetermined by what precedes
them. Rather the connection is weaker than a causal one. Earlier narrative
scenes raise or intimate questions, issues, or possibilities that are answered
or actualized by later scenes. A character robs a bank; this raises two well-
structured possibilities: he will be caught/he will not be caught. In the next
scene, the police, hitherto unseen, grab him as he exits the back door of the
bank. The later scene is not causally implied by the earlier scene. Instead the
earlier scene raised a structured set of possibilities, one of which the later
scene realized.37
Using the idea of a question to capture the idea of raising narrative
possibilities seems appropriate since the most convenient way in ordinary
language to state such possibilities is "Will x happen or not?" The concept
of the question, as well, enables us to explain one of the most apparent
audience responses toward linear movie narratives: expectation. That is, the
spectator expects answers to the questions the film raises or intimates about
its fictional world.
Some readers may balk at the preceding account on the grounds that it
does not seem plausible to characterize the spectators of movies as engaged
in a constant process o f question formation. Such spectators are not intro-
spectively aware of framing questions nor are they moving their lips—
silently speaking said questions—as scenes flicker by. So in what respect is
it accurate to say that such spectators are possessed of the kind of questions
so far discussed?
Clearly I must say that such spectators frame narrative questions tacitly,
and that they subconsciously (rather than unconsciously) expect answers to
them. The notion of a subconscious expectation—one we are unaware of
until it is perhaps short circuited—should cause no difficulty. When we are
told the plumbers will turn off the water for an hour, we still surprise
NARRATION 173

ourselves by expectantly walking over to the sink. But maybe it is thought

that some special problem arises when our subconscious expectation takes
the form of a question (which awaits an answer) rather than being based on
a reason or a belief. True, reasons and beliefs are best represented by
assertions rather than by questions. But then, o f course, if anything hinges
on this grammatical point, we might think to recast our narrative questions
as assertions—e.g., as predictions taking a disjunctive form such as "either x
will happen or y will happen or x will not happen. . . But, in fact, to hold
that a thought cannot be subconscious, depending on the grammatical
format o f its representation, is a highly unlikely hypothesis.
Needless to say, objectors to my characterization o f spectators as ques-
tion-formers may have in mind another issue—viz., that spectators are not
involved in explicit acts o f questioning when watching films. But here the
error is to confuse having a question—which may be an implicit and tacit
matter—with performing a self-conscious operation. Not all mental pro-
cesses can be equated with consciously performed processes; nor are all
mental states—such as having a question—to be equated with performing
a mental action such as that o f internal question-posing.
When following a narrative film, I want to say, a spectator internalizes
the whole structure o f interests depicted in the drama, and this structure
includes alternative outcomes to various lines o f action which the spectator
must keep track o f in some sense before one alternative is actualized in
order for the film to be received as intelligible. I postulate that the spectator
does this by tacitly projecting the range o f outcomes as subconscious
expectations which we can represent as questions. One argument in favor
of the tacit question model is that it explains how spectators are able to
regard films as intelligible. Another reason is supplied by the result o f
subverting the postulated expectations. If we stop the film midway, the
tacit questions soon surface: "Well, did he marry the princess, or did she fly
around the world?" greet us when the projector hum dies down.
At first glance, it may appear that the question/answer model is ill-
conceived to handle flashback scenes; however, the purpose o f most flash-
backs is to answer (or to offer information in the direction o f an answer to)
questions about why characters are behaving as they do, or why they are as
they are, in antecedent scenes. Though further qualifications are necessary,
my central hypothesis is that the major connective or logical relation in one
of the most basic form of linear movie narratives is erotetic. You can turn on
your T V any night o f the week and find several movies and weekly pro-
grams whose basic plot structure can be almost completely explained on the
interrogatory model.
174 NARRATION

The ways in which a question is made salient by a scene or group of

scenes is diverse, and will occupy much of the next chapter. A great deal of
the work is done in the writing, not only the dialogue and/or the intertitles,
but also in the choice of subject and the dramatic focus of given scenes—
i.e., the dramatic organization of the scene will make it clear that the major
issue is, for example, "will x propose to y ?" or "will z draw his gun?" This is
not to say that nonverbal factors like gesture, framing, character, and
camera position are not major components in leading an audience to regard
a certain set of characters and their intentions as primary nodes of interest.
Obviously, a whole ensemble of stylistic choices, often redundant ones,
prompt the audience to identify this or that issue as central in a scene.
Suppose a telephone is off the hook in the distant, blurry background of a
shot while a character is begging for a loan from a rotund banker, center-
frame foreground, in a well-lit, large medium close-shot. We know the
question the scene raises is whether the petitioner will receive the loan (or,
more broadly, will he get the money he needs?) and not whether the phone
will be hung up. And we expect ensuing scenes or events to answer that
question.
Though the question/answer structure is fundamental to certain linear
movie narratives, such narratives are not comprised baldly of simple ques-
tions and answers. Not every scene or event in the narrative can be de-
scribed as a simple question or answer. Most linear movie narratives have
scenes with more complicated functions than providing a simple question
or answer. The following is an inductive characterization of the scenes in an
idealized, erotetic, linear movie narrative: an event or scene in an erotetic
narrative is:
1. an establishing scene—an event or a series of events or a state of
affairs that introduces characters, locales, etc., or that establishes important
attributes of a character, locale, etc., and that, perhaps, but not necessarily,
raises or intimates a question. An establishing scene often initiates a movie
but one can come at any point in the film when the story involves the
addition of new characters, locales, etc.
2. a questioning scene. (More than one question may be introduced or
intimated in such a scene.)
3. an answering scene. (More than one question may be answered in
such a scene.)38
4. a sustaining scene. A scene may continue and intensify an earlier
question. The question, "will x escape," is intensified by a subsequent scene
in which we learn that, unbeknownst to x, he is surrounded. A scene that
NARRATION 175

begins to answer a narrative question but then frustrates the answer—e.g.,

a detective follows up the wrong clue—is also a sustaining scene.
5. an incomplete answering scene. A partial answer may be given to a
preceding question, e.g., "who killed Jones" is partly but not completely
answered when we learn that the killer is left-handed. 39
6. an answering/questioning scene. A preceding question may be an-
swered by a succeeding scene which also immediately introduces a new
question. A man and a woman meet in such a way that in scene # 1 the
question arises whether or not they will become a couple. Scene # 2 opens
on a confetti-strewn bed, answering the question of scene # 1 . But as the
man stumbles into the kitchen for breakfast, he is surprised to meet a child,
his new spouse's heretofore unmentioned daughter. Suddenly, the ques-
tion arises as to whether or not this new variable will endanger the new
relationship. Several ensuing scenes or an entire film could be devoted to
answering this question.
By using the question/answer model as the core concept of this categori-
zation of linear narrative scenes and events, I am not suggesting that it is a
competitor with taxonomies based on temporal relations—e.g., parallel
scenes, flashbacks, flashforwards, etc. The interrogative—will x be executed
or not?—can be articulated by two alternating questioning scenes of paral-
lel narration, as in Intolerance. The idea of parallel narration describes a
temporal relation in the fictional world of the movie while the ques-
tion/answer format describes the rhetorical-logical relation of scenes in the
film's structure.
These six functions (plus the fulfilling scene discussed in note 38) give us
a picture of the basic skeleton of a great many narrative movies. Whether a
scene or an event is part of the core plot of a linear narrative movie depends
on whether it is one of these types of scenes, i.e., whether it is part of the
circuit of questions and answers that powers the film. A scene that is not an
establishing scene is a digression if it lies outside the network of questions
and answers. A digression, of course, need not necessarily be something
bad; digressions may enrich the movie as a whole, as well as detract from it.
But a scene in a linear narrative movie will be a digression, for good or ill, if
it does not perform one of the core functions on our hst. 40
I hasten to add that I am not saying that all film narratives are or should
be erotetic linear narratives. Especially in art films, in contrast to popular
movies, there are episodic narrative structures, such as one finds in The Tree
of Wooden Clogs or Amarcord in which scenes are generally linked, for
realistic effect, by principles of rough temporal contiguity and often geo-
176 NARRATION

graphical propinquity, rather than in terms of questions and answers. This

type of narration often has as its aim the desire to impart a holistic sense of a
given milieu by itemizing or layering details concerning life in a certain
culture or subculture at a given time. As in literary ventures, such as Pictures
from an Institution, the importance of linear progression is deferred in favor
of provoking an elaborate sense of the texture and tempo of the "world"
depicted in the fiction. The film does not rush us forward along an arc of
expectations, but is said to invite us to "live in," to appreciate the rhythms
of life of, to savor (and thereby understand) the milieu that it represents.
Films can forego a linear structure for all sorts of reasons, especially for
expressive effects. The scenes in Satyricon do not follow a question/answer
logic, but that, of course, is exactly what engenders the alien, mysterious
quality that Fellini sought when he created what he called this "science
fiction" film of the past. Likewise Welles' The Trial abrupdy shifts scenes to
instill a sense of arbitrariness. Fantasy films—whether supernatural or
psychological—at times have scenes that could not plausibly be mapped
onto the question/answer model; the apparitions of Death 'mAll That Jazz
could not plausibly answer any questions any spectator could have as the
film proceeds; they are there to signal the egocentric view Bob Fosse has of
himself as a special someone in touch with an eruptive, exclusive, transcen-
dent reality. Modernist exercises like Last Tear atMarienbad and India Song
literally defy the erotetic model—they are all questions with no answers.
Consequendy, because of these and many other types of examples, the
question/answer model does not apply to all narrative films, nor is it an
evaluative grid with which we can measure the worth of every narrative
film. But it is at least a description of the core structure of a great many
linear narratives, particularly of the type found in mass market movies.
For my present purposes, I would like to maintain that the erotetic
narrative represents the core structure of movie narration. One might also
speculate (and I stress "speculate" here) that the erotetic structure of narra-
tion may represent something deeper. It may be a model of what we might
call the basic film narrative. It may be the erotetic narrative that most of us
have in mind when we hear the phrase "narrative movie." Two consider-
ations count in favor of this hypothesis. On the one hand, we do perceive a
difference between a mere chronicle film—my home mo.vie of my summer
vacation in which each event follows the next simply because it was what
happened next—and a film like The Lone/UUe Operator. Clearly, the differ-
ence between these different representations of human action is one of
structure. The question/answer model provides us with one general struc-
tural differentia that we can use to distinguish the chronicle from some-
NARRATION 177

thing that we might consider a minimal narrative film. However, such

considerations only lead us to regard the erotetic model as a candidate for
the tide basic narrative.
The question/answer model gives us a means of differentiating one very
simple narrative film type from a chronicle. But why should this type be
considered more basic than the episodic film narrative or the other narrative
variations alluded to previously? One reason is that we understand these
alternative modes of film narration by comparing these modes to a more
basic linear structure of the sort described by the erotetic model. The
disjunctiveness of Satyricon and the attendant qualities we associate with it
involve an implicit contrast or deviance from more standard forms of
conjunction—which, in turn, may be related to a propensity to form
certain cognitive expectations (viz., questions to be answered). The lack,
omission, or foregoing of a structure that evokes expectation and/or its
fulfillment is a pertinent stylistic element or choice in a film because it is a
contrast to a more basic, "normal" type in which certain connectives are
expected. Even with the case of the episodic structure, as it developed as a
major vehicle of film realism, we note that it was able to do so in lieu of its
divergence from the linear forms of classical narrative cinema. That is, part
of the reason why an episodic structure is held to have a special affinity with
realism—i.e., it is said to project the quality of realism—is because it is said
to be looser ("more inclusive" and, thereby, "truer to reality") than the
historically dominant, alternative mode of cinematic narration, viz., the
linear narrator film which is based on the erotetic model.41 My point here is
not to draw an absolutely clean demarcation between erotetic narrative
films and other sorts. Some films will mix elements of different narrative
types in different proportions. For example, a realistic film like The Tree of
Wooden Clogs, though predominantly episodic, employs the question /an-
swer structure at crucial points; indeed, one of the most pressing issues in
the film hinges on the question of whether the father will be caught and
punished for cutting down "the tree of wooden clogs." Moreover, even if I
am wrong that the erotetic model is some sort of basic film narrative, I may
still be correct in asserting that it gives us an account of the core narrative
structure of mass movie story telling that is far superior to the equilibrium
model.
Before leaving the topic of the question/answer model of film narration,
a distinction between two types of narrative questions should be drawn. I
have been emphasizing the question/answer model as a means of linking
scenes and events in movies. But questions are also a means for organizing
whole narratives, as we saw earlier. Thus, it is worth drawing a distinction
178 NARRATION

between macro-questions and micro-questions in movie narration. In

Buster Keaton's The General, there are three macro-questions—will
Johnny Gray win his true love, will he recover his train " T h e General," and
will he eventually succeed in enlisting in the Confederate Army. In The
General these three questions are interrelated, o f course. Gradually they
dovetail with each other. When they are all answered, the film is effectively
over. We don't worry about whether or not the happy couple will have
three children because that is not a question raised in the film. 4 2 We say that
a film is complete and that we feel a sense o f closure when all the macro-
questions in the film have been answered. 4 3
The General has three macro-questions but it also has a large number o f
micro-questions which connect scene to scene and fictional event to fic-
tional event. For example, in one scene the Union hijackers scatter debris
on the railroad track in order to frustrate Johnny's pursuit. This is undoubt-
edly related to the macro-question o f whether Johnny will recover his
engine—one might call it an instantiation o f the macro-question—but at
this point the answer to the macro-question is momentarily dependent on
the answer to a micro-question—will Johnny be able to handle these
obstacles and avoid derailment—a question that following scenes or events
answer.
Most movies are animated by macro-questions which organize the large
movements o f the bulk o f significant action in the film and the micro-
questions are generally hierarchically subordinate to the macro-questions.
In Wargames, at a certain point, most o f the action is devoted to answering
the question o f whether nuclear destruction can be averted. O f course,
movies often have more than one macro-question. Into the Night asks both
whether the romantic leads can escape the Middle Eastern villains and
whether this couple will become lovers. Both macro-questions are an-
swered by means o f roughly the same sequences of action and the micro-
questions and answers that structure those sequences tend, finally, to
dovetail with the answers to these presiding macro-questions. What is
called closure in movies can be explained as that moment when all the
saliently posed and sustained questions that the movie has raised have been
answered.
Though most movies are organized in terms o f what have just been
referred to as presiding macro-questions, certain films, notably serials, like
Fantomas, and films composed o f serial-like material, like Spies, often do not
seem to have overarching or dominant macro-questions. They often seem
to be a string o f suspense episodes, escapes, and entrapments, that are only
very broadly connected under a vague question like "will g o o d triumph or
NARRATION 179

will evil?" Such films structure individual sequences by means of macro-

questions but the whole story is not subordinated to a handful of organiz-
ing macro-questions from episode to episode. In all probability, the use of
presiding macro-questions to organize films was an innovation which cor-
responded to the transition to the dominance of feature filmmaking in
contrast to the period when serial filmmaking and featuresfilmmakingwere
of roughly the same importance.44 It is the lack of presiding macro-ques-
tions that may make the tying together of the episodes in such serials seem
like a string of firecrackers'going off as well as promoting the sense of
temporal shapelessness that such films may occasion. But, in any case, the
vast majority of movies for quite some time have not only employed the
erotetic narrative structure but they have organized their questions by
means of presiding macro-questions.
So far, it has been argued that movies, the products of what might be
thought of as Hollywood International, narrate by generating questions
that are internal to the film that subsequent scenes answer. Admittedly, this
is not a form unique to film or movies, for it is also exploited in mystery
novels, adventure stories, Harlequin romances, Marvel comics, and so on.
Nevertheless, it is the most characteristic narrative approach in the movies.
How can this be proved? Again, the best suggestion one can make here is
to embrace the question/answer model of movie narration—the erotetic
model—and then turn on your TV: watch old movies and new ones, TV
adventure series and romances, domestic movies and foreign mass market
films. Ask yourself why the later scenes in the films make sense in the
context of the earlier scenes—ask why the later scenes are there. My predic-
tion is that you will be surprised by the extent to which later scenes are
answering questions raised earlier, or are at least providing information
that will contribute to such answers. In adopting the hypothesis that the
narrative structure of a randomly selected movie is fundamentally a system
of internally generated questions that the movie goes on to answer, you will
find that you have hold of a relationship that enables you to explain what
makes certain scenes especially key: they either raise questions or answer
them, or perform related functions including sustaining questions already
raised or incompletely answering a previous question, or answering one
question but then introducing a new one.
Apart from the confirmation of the hypothesis afforded by this con-
frontation with the data, further support for the question/answer model
might be gained by using it, not to analyze, but to develop movie scenarios.
For when certain complexities and qualifications are added to the model of
the erotetic narrative, it is a very serviceable guide for producing stories that
180 NARRATION

strike one as typically "movieish," especially in their economy. Partial

confirmation of the question/answer model, then, would be its capacity to
direct the simulation of movie scenarios.
As a theory of movie narration, the model of erotetic narration seems
more comprehensive than the equilibrium model as well as more precise in
its elaboration of the structure of movie narration. At the same time, it may
also be especially useful in supplying an explanation of an important issue
for contemporary film theorists. We have encountered Baudry's interest in
explaining the powerful impression that film imparts. Now it is not clear
that film, as such, has this powerful impression, though movies do. Thus,
rather than search for the source of this powerful impression in a generic
feature of film as film such as projection, we might search for it in a
characteristic feature of movies, such as erotetic narration. The model of
erotetic narration, rather than psychoanalysis, might identify the source of
the powerful impression that concerned Baudry.
A successful erotetic narrative tells you, literally, everything you want to
know about the action being depicted, i.e., it answers every question, or
virtually every question that it has chosen to pose saliently. I say "virtually"
in order to accommodate endings such as that in the original version of
Invasion of the Body Snatchers, where the audience is left with one last
pregnant question. But even countenancing these cases, an erotetic narra-
tive has an extraordinary degree of neatness and intellectually compelling
compactness. It answers all the questions that it assertively presents to the
audience, and the largest portion of its actions is organized by a small
number of macro-questions, with litde remainder. The flow of action
approaches an ideal of uncluttered clarity. This clarity contrasts vividly with
the quality of fragments of actions and events we typically observe in
everyday life.
Unlike those of real life, the actions observed in movies have a level of
intelligibility, due to the role they play in the erotetic system of questions
and answers. Because of the question/answer structure, the audience is left
with the impression that it has learned everything important to know
concerning the action depicted. How is this achieved? By assertively intro-
ducing a selected set of pressing questions and then answering them—bv
controlling expectation by the manner in which questions are posed. The
powerful impression whose source Baudry sought is nothing but the excep-
tional perspicuousness, economy, and clarity of the action in movies which
is due to erotetic narration. The action in movies, that is, unlike most of the
action we encounter outside our cinemas (and fictions), is imbued with
heightened sense, direction, and intelligibility. Thus, it is not, as some
NARRATION 181

might argue, the realism of movies that compels us; rather erotetic narra-
tion, which departs from a "realist" recounting or chronicling of events,
attracts us. By virtue of erotetic narration movies are not only larger than
life but more legible too. Erotetic narration endows the movie with an aura
of clarity while also affording an intense satisfaction of our human propen-
sity for intelligibility.
Moreover, if the model of erotetic narration explains why whole movies
attract us, in respect of their "unrealistic" intelligibility of action, the model
also explains how it is we stayrivetedin our seats. Given the erotetic model,
we can say what it is that audiences expect: they expect answers to the
questions that earlier events have made salient—will the shark be de-
stroyed, will the jumbo jet crash, will Johnny Gray win Annabelle? If it is a
general feature of our cognitive makeup that, all things being equal, we not
only want but expect answers to questions that have assertively been put
before us, this explains our intense engagement with movies. For even if the
question is as insignificant to us as whether the suburban adolescent in
Risky Business will be found out by his parents, our curiosity keeps us glued
to the screen until it is satisfied.
 5

CINEMATIC
NARRATION

A major concern of the last chapter was accounting for what might be
thought of as the large-scale, or overarching coherence, of films, especially
of the category of mass market movies. That is, we examined competing
theories of the ways in which the narrative as a whole, so to speak, is
organized. The equilibrium model proposes a process of transformation-
substitution-repetition as the means by which the narrative of the movie
secures intelligibility, whereas the erotetic model holds that the major
connective between scenes and scenes, events and events in movies is an
internal process of questioning and answering.
However, these structures are merely narrative skeletons or armatures
which need to be filled in. That is, cinematic scenes, sequences, and
events—the things to then be related by processes of repetition or internal
questioning/answering—need to be built up and organized. Thus, in order
to round out one's theory of narration, one needs a small-scale narrative
theory, a theory that accounts for the principles and processes by which film
scenes, events, and sequences are coherently composed. This involves us in
explaining the ways in which shots arc coherently connected with other
shots and how the continually shifting views of cinema, as a result of such
elements as film editing, can be assimilated by spectators who derive intelli-
gible scenes, sequences, events, and, ultimately, whole stories from the
alternating views of the cinematic image flow characteristic of most narra-
tive films and certainly of most mass market movies.
Typically, a narrative film or movie confronts its audience with a flow of
images or views which are often fragmentary as well as spatially discon-
CINEMATIC NARRATION 183

tinuous. Shots are compounded with shots taken from different camera
positions. These views are built up into actions and events, scenes and
sequences which, in turn, are worked into whole stories. This raises ques-
tions about the principles of film editing, but, as well, is also related to
camera movement, for, as Hitchcock dramatically illustrates in Rope, narra-
tive camera movement, in large measure, apes the principles offilmediting.
The question of cinematic narration is the question of the small-scale
narrative coherence of films; the question of the ways in which events,
scenes, and sequences are constructed from shots and alternating views.
Verbal language plays an obvious role in this process. But though one
would not dare to say that the role of dialogue, intertides, and commentary
in cinematic narration requires no theoretical investigation, an account of
the visual elements in cinematic narration seems perhaps more pressing
since it is easy, at least speculatively, to imagine that the shifting views of
typicalfilmsand movies could result in confusion rather than coherence. So
this chapter will be primarily, though not exclusively, concerned with
cinematic narration as visual narration, and with the ways in which visual
narration constructs coherent units from shifting views.

SUTURE
Though the term "suture" is often treated as a generic term for subject
positioning, conceived of as a "binding" or stitching" of the subject into a
film, the term in film studies arose with reference to answering certain
questions about small-scale cinematic narration in relation to the coherence
of such editing structures as the shot/reverse shotfigureand point-of-view
editing. The idea of suture appears to have been introduced into contempo-
rary film theory by Jean-Pierre Oudart, primarily through a discussion of
the shot/reverse shotfigure.1The shot/reverse shotfiguretypically involves
a pair of shots in which each shot represents, "from a more or less oblique
angle, one endpoint of an imaginary 180-degree line running through the
scenographic space;" as well, film theorists also seem to include, as a
variation of the figure, pairs of successive shots that are (counter-field)
reverse images of each other, i.e., that cross the aforesaid imaginary 180-
degree line. 2 This sccond variation often occurs in the context of certain
types of point-of-view editing, though Oudart, but not his followers, is not
centrally concerned with such "subjective" figures.
According to Oudart, when a shot appears on the screen, it is greeted
with jubilation. Of course, this very description invokes the mirror stage
experience of the wholeness of Lacanian psychoanalysis and the coherence
184 CINEMATIC NARRATION

o f the infant's body postulated therein. As contemporary film theorists are

quick to point out, with the first o f the pair o f shots in the shot/reverse shot
figure:

The viewer of the cinematic spectacle experiences shot 1 as an imaginary

plenitude, unbounded by any gaze, and unmarked by difference. Shot 1 is
thus the site ofjouissance akin to that of the mirror stage prior to the child's
discovery of its separation from the ideal image which it has discovered in the
reflecting glass.3

However, Oudart believes that shot one also implies an absence, namely
the absence of the off-screen space behind the camera, a "fourth side, a field
o f pure absence." 4 So the initial sense o f plenitude gives way to a sense o f
limitation and Lacanian absence. Daniel Dayan describes this phase thus:

When the viewer discovers the frame—the first step in reading the film—the
triumph of his former possession of the image fades out. The viewer discovers
the camera is hiding things, and therefore distrusts it and the frame itself
which he now understands to be arbitrary. He wonders why theframeitself is
what it is. This radically transforms his mode of participation—the unreal
space between characters and/or objects is no longer perceived to be pleasur-
able. It is now the space which separates the camera from the characters. The
latter have lost their quality of presence. The spectator discovers that his
possession of space was only partial, illusory. He feels dispossessed of what he
is prevented from seeing. He discovers that he is only authorized to see what
happens to be in the axis of the gaze of another spectator who is ghosdy or
absent.5

That is, the initial sense o f plenitude o f the shot gives way to a sense o f
absence, an awareness o f not only what has been left out by framing but also
o f the reverse field o f the shot, the field in which the viewer o f that shot,
figuratively speaking, would have been stationed. The viewer in this absent
reverse field is called "the absent one." The classical narrative film, however,
moves to efface this absence. It does so by manifesting the offscreen reverse
field o f the absent one o f shot one in shot two.
Oudart identifies the absent one with the narrator of the film. It is the
narrator's position in regard to shot one that is absent from the image. By
supplying that absent perspective in shot two, "the appearance o f a lack
perceived as a someone (the Absent One) is followed by its abolition bv
someone (or something) placed within the same field."6 Thus, the specta-
tor, identifying with the field o f view from shot two, fills in for the absent
one, thereby suturing the absence or gap created by shot one. David
Bordwell writes, summarizing Oudart's position, that:
CINEMATIC NARRATION 185

Shot/reverse-shot cutting plays down narration by creating the sense that no

scenographic space remains unaccounted for. If shot 2 shows that something
is "on the other side" of shot 1, there is no place for the narrator to hide.7

Indeed, some contemporary film theorists think this suturing operation

effaces narration not only by leaving the narrator no place to hide but by
inducing the spectator to fill in for the absent one in an act of imaginary
identification, which would help account for the transparency o f such
films—if this is thought to lead the spectator to identify herself as the
enunciator. The film generates gaps and absences engendering the specta-
tor's processes of identification to fill them in or suture them. The specta-
tor's sense of identity as a unified subject is appropriated by the film,
indeed, stitched or bound into it.
It is important to note that in this account o f suture the potential source
of incoherence and disruption in the visual flow of narration is the gap
provoked by shot one in terms of the implication of the absent one.
Coherence and the understanding o f the flow of imagery is secured by the
operation of the spectator's filling-in activity. The visual narration, in
consequence, proceeds seamlessly and without apparent disjunctiveness.
Certain followers of Oudart, in contrast to their mentor, apply the
notion of suture specifically to point-of-view editing. There are a variety of
types of point-of-view structures, of course. However, for the purposes o f
illustrating suture, theorists often cite one in which shot one presents a view
of an object or character in a way that prompts the audience's eventual,
uneasy wonderment about whose view this might be. Shot two then shows
us a character who possesses this viewpoint, and with whom the spectator is
said to identify. Daniel Dayan appears to be responsible for this application
of suture theory to point-of-view editing, and it has been embraced by
many contemporary film theorists. O f the effects of the suturing process in
regard to point-of-view editing, Annette Kuhn claims:
the look of the spectator seems to come from a place identical to the source of
the look of character A: this appearance is brought about by constructing,
through editing, an "eyeline match." The spectator therefore stands in for A,
who is of course absent from this shot. When this is followed by . . . (a shot),
in which A is once again present to the image, the shot-reverse shot figure is
closed and the gap in the spectator's relation with the film is for the moment
"sewn up," sutured. Such moments of absence, presence and suture are
repeated throughout the film, thus ensuring an ongoing process of subject
position in cinematic enunciation.8

So, whereas in the account of suture in regard to the shot/reverse shot,

the spectator appears to identify with the view from the reverse field,
186 CINEMATIC NARRATION

thereby effacing narration, in the theory o f suture in relation to point-of-

view editing, the identification is with the look o f the character, the net
effect o f which is also the effacement o f narration insofar as the absent one's
look has been masked by the conflation o f the view o f the spectator and the
character. However, as in Oudart's account, the problem for the unity and
coherence o f visual narration that suture solves is the danger that the view
o f the first shot in the chain is not assigned to a viewer. As well, in terms o f
abetting the cinematic transparency effect, when the putative subject o f the
view o f the first shot is sutured into the text, it is a misleading one, for it is
not the view o f the narrator that is signaled, but the view o f a misrecognized
subject o f vision.
As this already suggests, the contemporary film theorist, given the Al-
thusserian-Lacanian framework, is likely to see ideology at work in the
process o f suturing. Dayan writes:

What happens in systematic terms is this: the absent one of shot one is an
element of the code that is attracted into the message by means of shot two.
When shot two replaces shot one, the absent one is transferred from the level
of enunciation to the level of fiction [i.e., to a fictional character, the possessor
of the point of view]. As a result of this, the code effectively disappears and
the ideological effect of the film is thereby secured. The code, which produces
an imaginary, ideological effect, is hidden by the message. Unable to see the
workings of the code, the spectator is at its mercy. His imaginary is sealed
into the film: the spectator thus absorbs an ideological effect without being
aware of it. 9

Here the ideological significance o f suture appears to be at least twofold.

On the one hand, it masks the operation o f the narration, thereby enhanc-
ing the putative transparency o f the film as well as leading to the absorption
o f the ideological effect which involves the constitution o f unified subjects
who in this case are sutured and positioned in the film.
I f suture theory is restricted to the phenomenon o f shot/reverse shot, it
will not be able to offer a general account o f the intelligibility o f visual
narration in cinema, for the best available estimates indicate that since the
twenties, only 30 to 4 0 percent o f the average film is composed o f reverse
angle cuts. 1 0 However, one may argue, as do Stephen Heath and his
followers, that though the shot/reverse shot figure and point-of-view edit-
ing are particularly illustrative instances o f suture, suture is in operation
throughout the cinematic narration or the discursive organization o f classi-
cal film. 11 Framing, the nature o f the cinematic image as a simulacrum,
narrative ellipses, camera movement, the use o f offscreen space and charac-
ter movement, each entail absences which must be sutured.
CINEMATIC NARRATION 187

This broadened notion o f suture, moreover, corresponds with the way in

which that notion is used by Lacanian psychoanalysts such as Jacques-Alain
Miller, since he thinks o f suture as an operation that makes all signification
possible. 12 That is, discursive structures, though posing as coherent unities
are always marked by difference and require spectators to complete them.
Film, as one form o f discourse, presents the spectator with structures
marked by absences which the spectator fills in, momentarily completing
them and, thereby, deriving a sense that the film is coherent, unified, and
homogeneous, which qualities, in turn, are appropriated by the subject
himself. The film is sutured by binding or positioning the subject in the
process o f the production o f meaning by the film-as-discourse.
Heath writes:
In its movement, its framings, its cuts, its intermittances, thefilmceaselessly
poses an absence, a lack, which is ceaselessly recaptured for—one needs to say
"form"—the film, that process binding the spectator as subject in the realiza-
tion of the film's space.
In psychoanalysis, "suture" refers to the relation of the individual as subject
to the chain of its discourse where it figures missing in the guise of a stand-in;
the subject is an effect of the signifier in which it is represented, stood in for,
taken place (the signifier is the narration of the subject). Ideological represen-
tation turns on—supports itself from—this "initial" production of the sub-
ject in the symbolic order . . . directs it as a set of images andfixedpositions,
metonymy stopped into fictions of coherence. What must be emphasized,
however, is that stopping—the functioning of suture in image,frame,narra-
tive etc.—is exactly a process: it counters a productivity, an excess, that it states
and restates in the very moment of containing in the interests of coherence—
thus the film frame, for example, exceeded from within by the outside it
delimits and poses and has ceaselessly to recapture.13
Suturing is an element in all discourse. For discourse to be intelligible
suturing must occur. In film, discursive structures, including not only
shot/reverse shots and point-of-view editing, but the use of offscreen space,
ellipses, camera movement, character movement, perspectival images, and
representation must be contained by suturing if they are to be coherent.
Suturing is supposed to explain how the alternating views of the cinematic
images hang together intelligibly and coherendy, which phenomenon it
attempts to explicate by means o f the positioning o f the subject. Moreover,
suturing must be continual as each discursive structure opens a gap or
absence. Each absence appropriates a subject as a condition o f its intel-
ligibility and coherence, while each suturing confirms the unified subject
position. The coherence o f cinematic narration hinges on suturing which
reinforces the ideological effect of the subject's positioning.
188 CINEMATIC NARRATION

What strikes one immediately about suture theories, whether Oudart's,

Dayan's, or Heath's, is that they are long on explanation and short on
cogent evidence. This can be illustrated by reference to Oudart's account.
The first shot appears on the screen and is greeted by jubilation, but this
feeling gives way to some unease which is assuaged by the appearance of the
reverse angle shot. Here at least three different states are being attributed to
the spectator. But what behavioral evidence, verbal or otherwise, or even
what introspective evidence, is there that the spectator is in any of these
states, or that the spectator moves from one of these states to another. The
problem here is not that the data have been contaminated by some theory.
Rather, the problem is that there are no data; there is merely theory. Yes,
there are shot changes. But why suppose that these shot changes corre-
spond to changes in the internal states of the spectator? Before one goes on
to theorize about the nature of internal changes in spectators, one should
produce some evidence that the spectators are undergoing some sort of
transition. That there are changes in the stimulus does not indicate a
corresponding change in the spectator, and certainly not such dramatic
changes as Oudart advances.
A rational constraint on explanation is that when explanations postulate
unobservable changes, these should be correlated with something about
the data. Without such constraint, there would be no rational way to decide
the difference between explanations that postulate fifty changes in the
spectator of a shot/reverse shot exchange and ones that postulate three,
forty, a thousand, or none at all. The explanation should not be discon-
nected from data in such a way that it free wheels in glorious "ad hocery."
But that, it would appear, is the case with Oudart's formulation of suture. It
offers a scenario of subject stages unconnected to any observable data about
spectators. Nor have I been able even to introspect anything like the
dialectic of delight and dejection Oudart hypothesizes in relation to
shot/reverse shots. And though their accounts at points differ from
Oudart's, Dayan's and Heath's are also singularly uninformative concern-
ing the data that ground the complex psychic gyrations they postulate.
Another way to get at this objection is to note that an explanation is
unpersuasive insofar as its explanatory machinery—here, changes in the
internal states of subjects—totally outweighs the evidence the explanation
purportedly organizes. If a formulation is all explanation and proposes little
or no evidence, then we have good reason to reject it. But the suturing
theories we have reviewed are exacdy of this sort; they are complex running
stories, perhaps better called myths, of successive changes in the intrapsy-
chic states of the spectator, but these changes have nothing to do with anv
evidence, external or even introspective, about concrete spectators.
CINEMATIC NARRATION 189

When linguists theorize about the grammaticality of sentences, they do

so on the basis of reports of competant native speakers. But there is no
corresponding restraint on cinematic suture theories. One might think that
the evidence might be easy to secure. One could legitimately speculate that
if we showed normal spectators something like Warhol's Empire, an eight-
hour "shot" with no counter shot, the average audience would become
uneasy. Probably so. But this would not give us the kind of evidence that
Oudart needs unless we could eliminate such countervailing hypotheses as
that it was the duration and/or the paucity of psychologically significant
content, and not the lack of a counter-shot, that causes the uneasiness with
regard to Empire. Ex hypothesi, I see no reason to think that one cannot
arrange a row of shots without reverse shots to which audiences will
respond without indications of unease, aggravation, boredom, or whatever
other behavioral evidence the suture theorist would like to correlate with
his theory.
This particular proposed experiment, if it can be sustained, of course,
would only cut against Oudart's and probably Dayan's versions of suture
theory, but not Heath's. For Heath contends that suturing is going on even
without counter-shots and the character's point-of-view shot. But this is no
theoretical advantage because it only makes it harder to imagine how
evidence could be adduced to support the theory.
This problem, of the lack of connection between theory and evidence, is
not merely a feature of suture theory but a disturbing proclivity of many
facets of contemporary film theory. Consider this account of the child's
entry into language which Colin MacCabe derives from Luce Irigaray and
which is obviously related to suturing:
In the development of the child there is a moment when the infant {tnfans:
unable to speak) enters language. In this process of entry, he/she becomes
aware of certain places which he/she as subject can occupy—these are the
points of insertion into language. Crucially this involves the learning of
pronouns: the realization that the "you" with which the child is addressed by
the father or mother can be permutated with the "I" in a situation from which
it is excluded—when the parents speak to cach other. This realisation is the
understanding that the "you" with which he or she is addressed can be
permutated with a "he" or "she," which is the possibility that the proper name
is articulated in a set of differences—and that the child is only a signifier
constantly defined and redefined by a set of substitution relations. The binary
I/you is transformed from two terms into a relational structure by the passage
through the empty place of the "he" or "she" and it is through the experience
of this empty place that the child enters language. 14

Two things are apparent immediately about this account. First, it is

190 CINEMATIC NARRATION

almost exclusively metaphorical; grammatical functions get redescribed

spatially, so that entering language, which might be more literally called
acquiring pronominal reference, is "explained" as if it were a matter of
crawling into an empty space, which is something MacCabe somehow
knows is the way infants regard pronouns. This mode of explanation by
means of a pastiche of puns and equivocations is, of course, the hallmark of
contemporary film theory.
However, MacCabe's account, like the suture theories above, also shows
a remarkable obliviousness to any canons of evidence. Basically his explana-
tion is a long, rambling story, with lots of asides about the intrapsychic
changes the infant undergoes, which include little or no reference to data or
changes in the data which would correspond at the observation level as
grounds for postulating the various sequences within the event-series of the
child's maturation as charted by the explanation. The explanation is a story
or narrative, but a narrative whose changes of events are not really corre-
lated to any reported observations of changes in the child's behavior, save
the acquisition of pronominal reference. But the mere acquisition of pro-
nominal reference hardly compels acceptance of the details of the involuted
intrapsychic saga MacCabe wants to tell. Contrast MacCabe's way of ex-
plaining child development with the rich citation of observation and detail
one finds in Piaget's theories. MacCabe's theory adduces no behavioral
evidence to ground the details of the abstract operations and goings-on that
comprise his story. Of course, similar problems beset the discussion of the
mirror stage. And since the mirror stage and the entry into language are
both prototypical of suturing, it should be no surprise that suture theory
flies untethered by evidence.15
Returning directly to the problems with suture theory, one is struck by
the fact that though suture theory is supposed to tell us how it is that
cinematic narration and its discursive organizations are coherent it savs
nothing explicit about the ways in which the discursive structures of cinema
address cognition. Suture theory is about the way the organism deals with
absences. But surely there must be more to assimilating a cinematic message
than an intimation of absence followed by a feeling of wholeness (and so on
and so on). That is, for example, by what means does the spectator under-
stand that this close-up of a piece of fruit represents the piece of fruit in the
medium shot of the table that follows it. Even if it is true that the close-up
of the piece of fruit engenders an intimation of absence, which is sutured by
the shot of the table, how is the understanding of the reference of shots and
their interrelation—the understanding of their coherence (if you will)—
explained by suture.
CINEMATIC NARRATION 191

Presumably, the Lacanian will want to say something like the operation
of the faculty of the Symbolic does the cognitive work about which I am
asking. This faculty, however, operates in such a way that along with
deciphering languages, in the broad sense of "language," it also inevitably
causes an intimation of absence which must be sutured. So a consequence
of the cognitive processing of the images would be the potential disruption
of understanding by absence, and suturing explains the way which this
disruption is buffered. But note we still have not been told how the image-
flow is understood, but rather how the process of understanding—con-
ceived as cognitive processing—has not auto-destructed. That is, we have
not learned the means by which the images have been rendered coherent,
but rather only how the work of cognition does not emotionally befuddle
itself by raising the specter of absence.
We may, of course, doubt that the play of absence and wholeness in fact
subtends every reception of discourse, including cinematic discourse. This
would return us to the question of evidence. What data motivates us to
postulate the feeling of absence and its consequent suturing? Moreover,
might it not be the case that if the cognitive process of understanding were
specified by contemporary film theorists, there would be no reason to
hypothesize further operations like suturing?
In criticizing Oudart, David Bordwell interprets Oudart's theory of
suture as an attempt to characterize "the viewing activities that the specta-
tor often engages in—anticipation, recollection, and recognition of the
spaces which narration presents."16 However, Bordwell argues that there
are perfectly plausible, alternative hypotheses to the invocation of the
unconscious Absent One by which we may organize the relevant spectator
activities. Bordwell himself favors an account which presupposes the exis-
tence of spatial, temporal, and logical schemata which a spectator antece-
dently possesses and uses to test the visual array for intelligibility.
Whether or not Bordwell's positive account is correct, several things
about it arc significant. First, unlike the suture theorist, Bordwell proceeds
with the understanding that a theory is advanced in competition with other
theories. He can defend the relative superiority of his theory over suture
theory on several counts: it does not postulate the existence of exotic
entities like the absent one and it respects an important methodological
constraint on psychoanalytic theorizing, viz., it attempts to find a rational
account of the image processing before resorting to an irrationalist hypoth-
esis. Bordwell does have to postulate schemata, though that hypothesis can
be defended by its theoretical usefulness for contemporary cognitive psy-
chologists. And, of course, by using a hypothesis shared with cognitive
192 CINEMATIC NARRATION

psychology, Bordwell brings his theory of film cognition closer to that of

ordinary cognition which would appear to represent a gain in theoretical
simplicity. Indeed, in terms of simplicity, since most contemporary film
theorists, given their semiotic orientation, probably already accept the
existence of schemata, along with suture, Bordwell's theory is more elegant
since he need only hypothesize schemata. Bordwell can explain with sche-
mata what contemporary film theorists try to explain with reference to
schemata and the absent one. Moreover, if one can adequately explain the
phenomenon with reference to schemata, or some other cognitive process,
there seems to be little reason to add further theoretical entities to the story,
since the story has already been told. Or, recalling Wittgenstein, a wheel
that turns no other parts is not part of the mechanism.
The neglect of cognitive-perceptual psychology in contemporary suture
theory is extremely troublesome. 17 Part of the aim of suture theory is to
explain the way in which spectators make sense out of films. An obvious
source for explanation or explanatory frameworks of the way that subjects
comprehend stimuli is cognitive psychology. But most contemporary film
theorists (with exceptions like Bill Nichols) attempt to do the whole work
of explaining the spectator's apprehension of coherence within the frame-
work of psychoanalysis. Why, generally, are no cognitive or perceptual
structures explicidy included in the suture model of film reception when it
seems so painfully clear that some such mechanisms must come into play
when audiences recognize a given film to be coherent? If contemporary
theorists believe that the type of structures hypothesized by cognitive and
perceptual psychology are inadequate to the task at hand, they should at the
very least say why.
There is, moreover, an important consideration of method here. Scien-
tific theories vindicate themselves through competition with other scientific
theories, and the way to argue for a new scientific theory is to show that it is
superior to all other competing explanations. Freud, fully realizing this,
introduced The Interpretation of Dreams with an exhaustive refutation of
previous dream theories. One accepts Freud's theory and one allows him to
postulate certain theoretical, unobservable processes because he has dem-
onstrated that his theory explains the data better than competing theories.
Contemporary film theorists do not confront competing cognitive-psycho-
logical explanations of the way film spectators recognize screen stimuli to
be coherent. The suture theorist does not scientifically demonstrate that the
operations he postulates are warranted since he has not shown that we
cannot get better explanations from research in cognitive psychology.
It is also helpful to remark that with psychoanalysis it does not seem
CINEMATIC NARRATION 193

advisable to postulate the operation of an unconscious mechanism, as in the

case of parapraxis, unless there is a breakdown in a cognitive (or organic)
mechanism which cannot be accounted for except by postulating an irra-
tional mechanism. That is, something irregular has to compel the inference
of unconscious agency. This seems to me to be a methodologically sound
constraint if only because the irrational mechanism will perforce be less
familiar, more a matter of speculative extrapolation than a cognitive pro-
cessing mechanism. And, of course, given what defines the object of psy-
choanalytic investigation, there is really nothing to explain by means of the
unconscious mechanism, by reference to the irrational, until rationality has
broken down. However, it seems peculiar to think that rationality has
broken down when one recognizes the coherence of a chain of shots in a
movie. Recognizing coherence in such a case is the epitome of regular,
normal, rational behavior.
The indeterminateness, contradictoriness, unpredictability, and ar-
bitrariness imputed to irrational structures supply at least one group of
considerations that make them epistemologically hazardous processes to
postulate. One should not rush to postulate unconscious processes. One
should attempt to exhaust other possible, logically determinate explana-
tions first. I am not saying that the unconscious should be foresworn as an
explanatory device, but rather that we should hypothesize unconscious
operations only after we have exhausted other avenues. Suture theorists,
however, do not wait to see whether cognitive models might explain the
kinds of coherences that interest them. This is as intemperate as starting to
explain how I know to answer "twelve" when you ask me "How much is six
plus six" by reference to my psychosexual development. Similarly, the
apprehension or recognition of the coherence of a string of shots can be, it
seems to me, explained thoroughly as a piece of rational behavior without
reference to unconscious operations.
Moreover, there are deep reasons for the type of methodological con-
straints advanced above. Psychoanalytic theory was designed to explicate
the irrational. As pointed out in an earlier chapter, the general paresis and
epileptic fits due to injury to Broca's area in the brain are wmrational and
obviously not appropriate subjects for psychoanalytic inquiry. Similarly, I
think that when an agent does something that is rational, we have no prima
facie reason to investigate further into the causes of his actions. That is, a
methodological constraint on psychoanalytic explanation, one which de-
fines the domain of psychoanalysis, is that it not be mobilized until there is
an identifiable disruption of rationality. Not all beliefs and behaviors, not
all social, aesthetic, emotional, and cognitive responses are candidates for
194 CINEMATIC NARRATION

psychoanalytic investigation. Insofar as psychoanalysis is designed to con-

ceptualize irrational behavior, which is only identifiable as a deviation from
some norm of rational behavior, there is no work for it to do where the
behavior in question is of an unmistakably rational sort. Furthermore, it
does not seem to me that understanding movies is either irrational or
involved in the engagement of irrational processes.
The identification of the irrational is logically dependent upon the ra-
tional. Consequently, it is necessary to establish a deviation of the rational
before resorting to the irrational. Thus, psychoanalysis has no role to play
except in such cases as inexplicable deviations from the norm of rationality
occur which are not attributable to organic causes. The restriction of
psychoanalysis to irrational phenomena follows from the logically depen-
dent or parisitic relation that the irrational bears to the rational.
Moreover, the basic concepts of psychoanalysis are metaphoric exten-
sions of the concepts of the rational—e.g., motive, intention, wish, drive,
need, and so on. That is, the concepts are all purposive, or ends' seeking.
The difference is that these forces are conscious and perform roles in
deliberation in rationalist psychology, but they are metaphorically ex-
tended to unconscious forces in psychoanalysis where intrapsychic entities
perform them. Psychoanalysis, by examining unconscious intentions and
repressed operations, explains actions, purposively characterized, that can-
not be explained by conscious or merely tacit intentions, beliefs, and rea-
sonings. Psychoanalysis, so to speak, takes over when purposes, intentions,
beliefs, and so on go underground, but it nevertheless continues to explain
behavior in terms of unconscious beliefs and intentions that are understood
in terms of an organism in pursuit of ends and goals. Thus, when we have
an adequate explanation available in virtue of rational goals and ends, there
is no reason to postulate further unconscious goals and ends.
Of course, another way to make this point would be to note that psycho-
analysis is not an explanation of all human behavior, but only of certain
aspects of human behavior. There is no reason to suppose that the kinds of
questions pursued by cognitive psychologists will be found by psycho-
analysis. The suture theorist should not presume that an explanation of the
ways in which a spectator apprehends the coherence of a shot chain is a
matter about which psychoanalysis has anything useful to say.
It may be that the suture theorist believes not that cognitive models for
understanding shot chains and movies are incorrect but that they will
ultimately complement the Lacanian model. Thus, the suture theorist does
not explore cognitive-perceptual frameworks for explaining the com-
prehension of the flow of cinematic images because he believes that these
CINEMATIC NARRATION 195

models will only enrich the fundamental suture description that he gives
with more (rather than competing) details. But this approach raises an
epistemological problem. For how does the suture theorist know that
cognitive mechanisms alone cannot fully account for the way spectators
come to see cinematic spectacles as coherent? That is, how can the suture
theorist rest assured that, once we have an account of cinematic coherence
in terms of cognitive mechanism, talk of suture will be extraneous or beside
the point? What suture theory gropes to explain abstractly, cognitive psy-
chology might explain concretely, without reference to almost mystical
concepts like the "absent one."
Previously it was noted that though suture theory was initially intro-
duced to conceptualize the spectator's assimilation of the shot/reverse shot
figure and point-of-view editing, the application of the notion of suture
was extended to every device and convention of film. Once suturing is
regarded as an attribute of everything that is coherent in film, however, the
theory offers little insight at the level of structure and pragmatics into the
workings and differentiation of specific cinematic devices and conventions.
Suture theory seems less concerned with exactly how—logically, rhetori-
cally, and structurally—film devices make (or appear to make) sense (com-
municate meaning), and more concerned with the fact that they do make
sense. That a particular device sutures is the conclusion of each explanation.
Since suture is a general account of what makes coherence possible, it
merely tells us how every film device and practice is the same—they all
suture; they appear to make sense by masking an absence and binding a
subject.
Some specificity seems required in suture explanations when the analyst
names the particular absences incurred by a given device, and when the
analyst outlines the way in which the device poses its appearance of com-
pleteness, wholeness, and homogeneity. But this is not a really compelling
degree of specificity, because (1) the absences denominated are fairly rou-
tine; and (2) the structural descriptions of the device's coherence involve
the same key words—center, position, bind, and the rest of the Lacanian
litany—for whatever device is under examination. Thus, suture theory
gives almost the same account for every cinematic device and practice.
Some limited ingenuity may be required to blend Lacanian terms into these
descriptions, but once this is done, the explanation converges on the same
Lacanian subject-in-language myth. Film theory seems to dissolve into
showing how every question about cinema and cinematic devices can be
given the same (ultimate) answer which, as well, is the same answer given
by Lacanian theory in its explorations of linguistics and culture at large.
196 CINEMATIC NARRATION

However, the generality and abstractness o f such suture explanations leave

unanswered the specific film theoretical questions that we might ask about,
for example, the different potential uses o f the close-up.
The suture theorist may say that the above objection is grounded in a
misconception o f what film theory should be. Such a theorist might say that
what has been gained by suture theory is a bridge between film theory and
the theory o f the subject and o f ideology in general. We should put aside
the remaining questions o f film theory as unimportant and be happy that
we now have a formula—the suture scenario—that tells us about the inner
spring that powers everything. But the remaining questions o f film the-
ory—such as what are the different types o f camera movement and what
underlying principles interrelate them?—will rtot disappear when we are
told that camera movements suture subjects. People who make, study,
teach, and otherwise strive to understand film will still require answers to
the mundane questions o f film theory, despite their acquisition, via suture
theory, o f the ontology o f the subject.
Indeed, oddly enough, the suture theory o f film, when extended to every
cinematic coherence-making device, seems to require the existence o f ordi-
nary film theory in order to isolate the coherence-making structures o f film
and to supply the basic account o f those devices upon which the suture
account is then grafted. That is, the suture theorist explains the way in
which a device works—say offscreen sound—in the manner an ordinary
film theorist might and then goes on to conclude that in virtue o f the
coherence just described the device sutures. One must ask whether suture
theory, in this regard, is an addition to or a parasite upon the structural and
rhetorical analyses o f ordinary film theory? Or, to put the matter slightly
differently, does suture theory add anything to ordinary film theory or,
again, is suture a wheel that turns nothing?
If suture theory is threadbare as film theory, it is also impoverished as a
putative scientific theory. Scientific theories are aimed at explaining specific
variations in phenomena. But expanded versions o f suture theory, in claim-
ing that all discourse, including all film discourse, is to be explained in terms
o f suture, is rather like the theory that God makes everything happen. I ask
why the flower died, the brakes jammed, and the sun rose, and I am told
that in each case "God made it happen." I soon see that this kind o f answer
is going to get me nowhere in understanding the phenomenon at issue, and
I search for answers in terms o f the more restricted fields of biology, auto-
mechanics, and astronomy. Similarly, if I ask what makes a simple declara-
tive sentence coherent, an offscreen sound intelligible, and a structuralist
materialist film comprehensible, and I am told in each case that suture made
CINEMATIC NARRATION 197

it happen, then I begin to suspect that the answer is more general than the
question with which I am concerned.
In order to avoid vacuity, a theory must not only explain why x is the case
but also under what circumstances x would not be the case. If I attempted
to explain both why a certain flower would live and why it would die by
saying that "God wills it," then my explanation would be vacuous. A
scientific theory must not only explain how such and such a state of affairs
came about but also how things might have been otherwise had the relevant
conditions been otherwise. However, it is not clear that expanded versions
of suture theory meet this requirement. For example, Stephen Heath ap-
plies the concept of suturing not only to Hollywood films but to more
subversive works such as News from Home as well as to structuralist-mate-
rialist films. Hollywood films suture but so do attempts to transgress the
forms and conventions of narrative film. The structuralist-materialist, av-
ant-garde filmmaker's attempt to subvert suture is said to result in a special
intensity of meaning. There seems to be nothing an avant-garde filmmaker
can do to achieve a condition of nonsuture. Every cinematic articulation is
explained by means of the idea of suture. But a theory that explains every-
thing, explains nothing.
So far we have raised theoretical questions concerning the very notion of
suture, construed both narrowly and broadly. We have not dealt with
certain of the ideological consequences that are thought to follow from
suture by contemporary film theorists, viz., that suture abets the illusion of
transparency of films and that it confirms the sense of subject unity that the
Althusserian-Lacanian finds so suspect. Of course, if the very notion of
suture is questionable, it is doubtful whether these ideological conse-
quences can flow from it. However, we have also considered, in previous
chapters, reasons for disbelieving whether these two putative ideological
effects of film are warranted.
We have pointed out that the notion that a movie appears transparent by
persuading the spectator that she is its enunciator is implausible. Via suture
theory, this is supposedly explained in terms of the spectator, in various
ways, identifying with the absent one. But we have seen that there may be
no theoretical justification for postulating the absent one, and, therefore,
no reason to suspect that anyone identifies with it. Similarly, we have
previously shown that there is no ground for thinking that spectators
believe they are the enunciator of the film, so we need not hypothesize
suture as the means by which this is accomplished.
Presumably, suture abets the ideological operation of transparency by
encouraging the spectator's apprehension of herself as the enunciator of the
198 CINEMATIC NARRATION

movie. Thus, the spectator would, I suppose, accept whatever ideological

views the film proposes as her own. Or, if the spectator has identified with a
character rather than the enunciator, the narrator is still thought to be
hidden in such a way that the spectator is unaware that the film's representa-
tion of events is controlled by the ideological interests of the enunciator.
However, we have seen that there is not only little reason to think that
these various identifications are really occurring, but, as well, litde reason to
think that spectators believe that films lack narrators. If a spectator agrees
with the ideological position of a film, he may not attend to the way in
which the enunciator has distorted the subject material. But this has
nothing to do with the purported, quasi-formalist effects of enunciation
and suturing but rather is connected to the social conditioning of the
spectator. And, of course, where the spectator does not agree with the
ideological position of the film, he or she can be quite aware of the
controlling distortions of the filmmaker, despite the supposed facts that the
operation of suturing and masked enunciation should render this improb-
able.
The second ideological repercussion of suturing is, of course, the way in
which it reinforces the spectator's faith in subject unity. We have had
occasion to question both the Althusserian connection between subject
unity and ideology, and the Lacanian account of imaginary identification.
If those objections carry the day, then the proposed political consequences
of suture for subject unity can be rejected. At the same time, we may be
skeptical as to whether the connection between identification in suturing
with respect to the flow of cinematic images and the process of imaginary
identification relevant to Althusserian-Lacanian subject unity is anything
more than verbal. According to some suture theorists, a perspectivally
organized film image might be said to suture me by inducing an identifica-
tion with the camera such that I am positioned as a unified subject, mis-
recognizing myself as the enunciator of the shot. Thus, I am positioned as a
unified subject in the Althusserian-Lacanian sense. But, as has been pointed
out more than once already, the way in which I am positioned by such
shots, if it is appropriate to speak of positioning here, is spatial, whereas the
kind of positioning relevant to the Althusserian-Lacanian brief is a meta-
phor that pertains to a social position. However there is no reason, save
equivocation, to think that a spatial position and a social position have
anything to do with each other.
Indeed, the whole vocabulary of suture theory seems based upon meta-
phorical confusions. Suture binds gaps, stitches ruptures, closes absences.
In virtue of this it is said to render the cinematic array coherent. But
CINEMATIC NARRATION 199

"coherence" here must mean something like without gaps or a sense of

wholeness. However, this is not the sense of "coherence" one has in mind
when thinking about intelligibility. Being intelligible or coherent to the
understanding has nothing to do with whether or not a communication has
gaps; I can easily understand an enthymeme, i.e., discern that it makes
sense, though something is missing. What the suture theorist appears to
have done is to move from a process of finding something coherent, in the
sense of "whole," to a claim that this explains the way in which the spectator
finds the cinematic array coherent in the sense of being intelligible. Thus,
suturing is proposed as a means by which one makes sense out of cinematic
arrays when in fact the coherence secured is rather that of a sense of
wholeness. Thus, even if there is such a process as suturing, which is a rather
large i f , we might be skeptical about whether it tells us anything about how
cinematic arrays make sense. For the connection between coherence as
wholeness and coherence as intelligibility appears to ride merely upon an
ambiguity in terminology.

AN A L T E R N A T I V E A C C O U N T O F
CINEMATIC NARRATION
The flow of imagery in narrative films and popular movies is charac-
teristically fragmentary, a series of alternating views from varying camera
positions. But in the case of movies—as opposed to art films and experi-
mental films—ordinary spectators typically have no problem following the
narrative embodied in the fragmentary stream of images. Though one
might think that the discontinuities in the image track would make it
incoherent, ordinary viewers standardly have no difficulty in finding visual
movie narration coherent and intelligible. One proposed way to under-
stand this is suture theory. But, as we have seen, this is a highly implausible
approach. So an alternative account of how the spectator follows the flow
of visual imagery in the cinematic narration of movies is called for.
Though the fragmentary flow of imagery in movies might make it seem
that movies would be difficult to follow, movies are in fact quite easy to
follow. In fact, movies, the popular films of Hollywood International, are,
it would seem, easier to follow than plays as they are standardly presented,
even though the scenes within plays are typically continuous in terms of
their space, scale, and the angle from which we view their action. Thus, one
way to approach the question of the coherence of the movie image track is
to contrast it with the theatrical image in order to isolate the standard
features of movies that differentiate them from the standard presentation of
200 CINEMATIC NARRATION

plays in such a way that typical movies arc more accessible than typical
theatrical performances. Our guiding hypothesis here is that due to certain
cinematic devices, which structure the cinematic image track and which
were developed early in the evolution of movies, the typical movie narrative
is more accessible than the typical play. That is, a movie narrative appears
easier to follow than a play. Moreover, if we can explain why this is so, we
should be on our way to isolating the features of the cinematic image track
that facilitate the spectator's reception of the flow of movie imagery as a
coherent story.
We have claimed that the typical movie, all things being equal, is more
accessible or easier to follow than the typical play, i.e., theatrical perfor-
mances as they are commonly encountered. The caveat, "commonly," is
added, of course, because there is no reason to believe that theatrical devices
that would be functionally equivalent to the movie devices about to be
discussed could not be invented, thus changing the relative accessibility and
easy coherence of typical movies versus typical plays.
We have asserted that movies are easier to follow than plays. What is it
that is distinctive about the way in which spectators follow movies? With
the typical movie, given certain of its characteristic devices, notably variable
framing, the movie viewer is generally in a position where he or she is
attending to exacdy what is significant in the action-array or spectacle on
screen. Another way of getting at this is to say that the filmmaker in popular
movies has far more potential control over the spectator's attention than
does the theatrical director. A crucial consequence of this is that the movie
spectator is always looking where he or she should be looking, always
attending to the right details and thereby comprehending, nearly effortless-
ly, the ongoing action precisely in the way it is meant to be understood.
Various cinematic devices, particularly variable framing, make movies
easier to follow and, therefore, more accessible than theatrical productions
because movie narratives are more perspicuous cognitively. And, as we
shall see, the cognitive clarity of movies may not only help account for their
accessibility but also for the special intensity of engagement and response
that movies elicit.
Of course, movies and standard theatrical productions share many of the
same devices for directing the audience's attention. Both in the medium-
long shot in movies and the proscenium stage in theater, the audience's
attention can be guided by: the central positioning of an important charac-
ter; movement in stasis; stasis in movement; characters' eyelines; light
colors on dark fields; dark colors on light fields; sound, notably dialogue;
spotlighting and variable illumination of the array; placement of important
CINEMATIC NARRATION 201

objects or characters along arresting diagonals; economy of set details;

makeup and costume; commentary; gestures; and so on. But movies appear
to have further devices and perhaps more effective devices for directing
attention than does theater as it is standardly practiced.
The variability of focus in film, for example, is generally a more reliable
means of making sure that the audience is looking where the spectator
"ought" to be looking than is theatrical lighting. But even more important
is the use in movies of variable framing. Through cutting and camera
movement, the filmmaker can rest assured that the spectator is perceiving
exacdy what she should be perceiving at the precise moment she should be
perceiving it. When a camera comes in for a close-up, for example, there is
no possibility that the spectator can be distracted by some detail stage left.
Everything extraneous to the story at that point is deleted. Nor does the
spectator have to find the significant detail; it is delivered to her. The viewer
also gets as close or as far-off a view of the significant objects of the story—
be they heroines, butcher knives, mobs, fortresses, or planets—as is useful
for her to have a concrete sense of what is going on.
Whereas in a theater the eye constantly tracks the action—often at a felt
distance, often amidst a vaulting space—in movies much of that work is
done by shifting camera positions, which, at the same time, also assure that
the average viewer has not got lost in the space but is looking precisely at
that which she is supposed to see. Movies are therefore easier to follow than
typical stage productions, because the shifting camera positions make it
practically impossible for the movie viewer not to be attending where she is
meant to attend. And, of course, the virtually unavoidable clarity of the
movie narrative in respect to the potentially impeccable direction of atten-
tion is a major constituent of the coherence of movies for spectators.
Variable framing in film is achieved by moving the camera closer or
farther away from the objects being filmed. Cutting and camera movement
are the two major processes for shifting the frame: in the former, the actual
process of the camera's change of position is not included in the shot; we
jump from medium-range views, to close views, to far-off views with the
traversal of the space between excised. In camera movement, as the name
suggests, the passage of the camera from a long view to a close view, or vice
versa, is recorded within the shot. Reframing can also be achieved optically
through such devices as zooming-in and changing lenses.
These mechanical means for changing the framing of an onscreen object
or event give rise to three formal devices for directing the movie audience's
attention: indexing, bracketing, and scaling. Indexing occurs when a cam-
era is moved toward an object. The motion toward the object functions
202 CINEMATIC NARRATION

ostcnsivcly, like the gesture of pointing. It indicates that the viewer ought
to be looking in the direction toward which the camera is moving, if the
camera's movement is being recorded, or in the direction toward which the
camera is aimed or pointing, if we have been presented with the shot via a
cut.
When a camera is moved toward an array, it screens out everything
beyond the frame. To move a camera toward an object, either by cutting or
camera movement generally, has the force of indicating that what is impor-
tant at this moment is what is on screen, what is in the perimeter of the
frame. That which is not inside the frame has been bracketed, excluded. It
should not, and in fact it literally cannot, at the moment it is bracketed, be
attended to. At the same time, bracketing has an inclusionary dimension,
indicating that what is inside the frame or bracket is important. A standard
camera position will mobilize both the exclusionary and inclusionary di-
mensions of the bracket to control attention, though the relative degree
may vary as to whether a given bracket is more important for what it
excludes, rather than what it includes, and vice versa.
There is also a standard deviation from this use of bracketing. Often the
important element of a scene is placed outside the frame so that it is not
visible onscreen, e.g., the child killer in the early part of Fritz Lang's M.
Such scenes derive a great deal of their expressive power just because they
subvert the standard function of bracketing.
As the camera is moved forward, it not only indexes and places brackets
around the objects in front of it, it also changes their scale. Whether by
cutting or camera movement, as the camera nears the gun on the table, the
gun simultaneously appears larger and occupies more screen space. When
the camera is pulled away from the table the gun occupies less screen space.
This capacity to change the scale of objects through camera positioning—a
process we call scaling—can be exploited for expressive or magical effects.
But scaling is also a lever for the direction of audience attention. Enlarging
the screen size of an object generally has the force of indicating that this
object, or gestalt of objects, is the important item to attend to at this
moment in the movie.
Scaling, bracketing, and indexing are three different ways of directing the
movie spectator's attention through camera positioning. In general, a stan-
dard camera positioning, whether executed by cutting or camera move-
ment, will employ all three of these means. But one can easily think of
scenes in which the bracket is reoriented, but the scaling stays effectively the
same—for example, a lateral pan as a character walks toward the edge of the
frame. Likewise, a camera movement might be important for what it
indexes rather than for whatever changes occur in the bracketing or the
CINEMATIC NARRATION 203

scaling; there are moving shots in the early ItalianfilmCabiria, for example,
where the camera nudges a few feet forward in a spectacle scene in order to
point the viewer's eye in a certain direction, though neither the bracket nor
the scale of the objects in the scene are changed appreciably. Both the
swamp scene and the trolley-car scene in Sunrise are artistically important
for the way in which they call attention to the bracket, rather than for their
scaling or indexing.
However, bracketing, scaling, and indexing can be employed in tandem,
and when they are, they afford very powerful means by which the movie-
maker controls the audience's attention. We suddenly see a medium shot in
which the gun is being pointed at the heroine by the villain, telling us, in
effect, that now the important thing about the gun is its role within this
newly framed context or gestalt. The constant reframing of the action that
is endemic to movies enables the spectator to follow the action perfectly,
and, so to say, automatically.
It may seem strange to some readers that I am treating the variable
framing due to editing and that due to camera movement in much the same
way. Of course, I readily admit that editing and camera movement are
different in many respects. However, in their use of the frame for narrative
purposes they appear to exploit the same formal potentials of reframing.
Perhaps this claim can be rendered somewhat more convincing by a histor-
ical conjecture. The perfecting and wide dissemination of narrative camera
movement, in such films as Regeneration, The Vagabond, and finally The
Last Laugh, appears to develop after the popularization offilmediting with
its routine alternation of close views, medium views, and far views. One
suspects it develops in imitation of editing. Of course, there has been
camera movement since the earliest days of film. Lumiere's cinematogra-
phers mounted their cameras on boats floating down the Nile, and, as well,
early on, cameras were stationed in fast moving cars to keep track of
hurtling action scenes—trains speeding ahead, cowboys on horseback rid-
ing to the rescue, and slapstick comics and Keystone Cops racing in pro-
gressively deteriorating jalopies. But the use of camera movement to pick
out and focus, and to recontextualize narrative detail emerges, albeit quite
rapidly, after editing, receiving its inspiration, one may speculate, from the
clarification of the means for controlling attention via film editing. That is,
indexing, scaling, and bracketing seem to have been mastered first in film
editing which opened the way for narrative reframing via camera move-
ment. Thus, if this account is correct, it should come as no surprise that in
terms of the use of the frame for directing narrative attention, editing and
camera movement exploit the same functions.
Needless to say, our account of variable framing is only meant as an
204 CINEMATIC NARRATION

account of its operation with respect to its direction of the spectator's

attention to the details of the ongoing story. This is not an account of everv
aspect of variable framing. Variable framing also can perform a range of
expressive functions. The speed of editing or of camera movement, or the
distances traversed from view to view, can invest the scene portrayed with
expressive or aesthetic qualities. For example, a large, slow camera move-
ment may strike us as stately, while a brace of short shots of different,
spatially quite disparate, locales may project feelings of nervousness or
excitement. Thus, by emphasizing the service variable framing performs in
the direction of the spectator's attention to the narration, we in no way
deny that variable framing is involved in further functions. Nor, moreover,
do we claim that variable framing is the only means for the direction of
cinematic attention, but only that it is very central.
Returning to our contrast between movies and theatrical performances,
we can note that adaptations of stage technology might be available that
would provide theatrical means that would be functionally equivalent to
the scaling, bracketing, and indexing functions of variable framing in
movies. Magnifying mirrors might be used to enlarge stage details at
appropriate moments; the leg curtains could be motorized to constantly
reframe the action; and indexing might be approximated by the use of
revolving stages that rotate the important characters and actions toward the
audience. If these devices were not too distracting in and of themselves,
they might supply the theater director with attentional levers that are
functionally equivalent to scaling, bracketing, and indexing. However,
these devices are not customary in theater as we know it, and the putatively
greater, cognitive accessibility of movies in contrast to theater seems expli-
cable in terms of the potentials that variable framing affords for the direc-
tion of attention.
Of course, films can be made without variable framing; but movies,
popular mass market narratives, rely on variable framing to automatize the
spectator's attention. Also, variable framing is not unique to the movies;
other types of film—including art films and experimental films—employ it.
Yet variable framing is key to why movies are accessible—to why movies
appear to be almost automatically coherent and intelligible. The variable
framing insures that the spectator is always attending to the details and
configurations that, for the purposes of the story, are appropriate; variable
framing virtually guarantees that the spectator is attending where and when
she should.
So far we have concentrated upon the way in which variable framing—
one of the central visual devices of cinematic narration—serves to render
the ongoing story in a movie intelligible, almost automatically. But perhaps
CINEMATIC NARRATION 205

the consideration of variable framing can also contribute to our under-

standing of the incredible impact that mass market movies as a form appear
to have on audiences.
As already noted, through variable framing, the movie director assures
that the spectator is attending when and where she should. The action and
its details unfold in such a way that every element that is relevant is
displayed at a distance that makes it eminendy recognizable and in a
sequence that is intelligible. Ideally, variable framing allows us to see just
what we need to see at changing distances and at cadences that render the
action perspicuous. The action is broken down into its most salient ele-
ments, distilled, that is, in a way that makes it extremely legible.
This kind of clarity, which is bequeathed to the audience automatically
by variable framing, contrasts strongly with the depiction of action in
theatrical representations. There, depiction is not analytic but a matter of
physical enactment, generally occurring in something approximating real
time, and presented at afixeddistance to each viewer. Of course, theatrical
action is abstracted, simplified, for the sake of legibility, often employing
emblematic gestures. It is clearer, that is, than the actions we encounter in
everyday life. But theatrical action is not as clear and analytically distinct as
movie action as portrayed by variable framing.
Movie action, given the way it can be organized through camera posi-
tioning, is also far more intelligible than the unstaged events we witness in
everyday life. This is an important feature of movies which helps to account
for the way in which movies grip us.
Our experiences of actions and events in movies differ radically from our
normal experiences; movie actions and events are so organized, so automat-
ically intelligible, and so clear. The arresting thing about movies, contra
various realist theories, is not that they create the illusion or impression of
reality, but that they reorganize and construct, through variable framing,
actions and events with an economy, legibility, and coherence that are not
only automatically available, but which surpass, in terms of their imme-
diately perceptible basic structure, naturally encountered actions and
events. Movie actions evince a visible order to a degree not found in
everyday experience. This quality of extraordinary, uncluttered clarity grat-
ifies the mind's quest for order, thereby intensifying our engagement with
the screen. That is, along with partially explaining the way in which cine-
matic narration is coherent, variable framing can also contribute to an
account of the powerful impression that movies impart to spectators, which
impression is of interest to contemporary film theorists such as Baudry.
We have discussed the visual devices of movies in terms of the type of
clarity they afford the audience—in terms of how they enable the audience
206 CINEMATIC NARRATION

to sec all that it is relevant for them to see at the appropriate distance and in
the appropriate sequence. At the same time, in the chapter prior to this one,
another sort of clarity has been attributed to the type of large-scale narrative
structure that we identified as central to movies, i.e., to the eroteric narra-
tive. How, one might wonder, do these two "clarities"—of variable fram-
ing and erotectic narration—relate to each other? Well, generally, in
movies, devices such as scaling, bracketing, and indexing will be employed
so that the first item or thefirstgestalt ofitems that the audience is led to attend
to in a given shot is the item orgestalt that is most relevant to the progress of the
narrative—i.e., to the posing, sustaining, or answering of those questions the
movie elects to answer. The importance of variable framing for movies is the
potential it affords for assuring that the audience attends to everything that
is relevant, and that it does so automatically, so to speak. "Relevance" is here
determined by the narrative, or, more specifically, by the questions and
answers that drive the narrative, which in turn are saliently posed and
answered in important ways by means of variable framing.
Variable framing, along with the sorts of devices enumerated earlier for
directing attention in the single, medium shot, are the filmmaker's means of
visual narration. They enable her to raise questions visually: the question
"Will Jones be shot?" can be "posed" by focusing on a close-up of a gun. At
the same time, the visual depiction of an action can either sustain or answer
a question. "Will Eli Wallach die by hanging?" can be sustained by showing
him teetering on a chair with a noose around his neck, or answered by
showing us Clint Eastwood severing the rope in an act of superhuman
marksmanship. Of course, many of the pressing questions that drive movies
forward are not primarily set forth visually but are stated explicitly in the
dialogue, or are already implied in the scripting of the action. Nevertheless,
the devices of visual narration, if not the original source of the questions,
help make those questions salient.
The large-scale narrative coherence of movies derives from the erotetic
structure. The questions the film internally pose structure, broadly speak-
ing, the audience's expectations, roughly circumscribing their horizon of
alternative possibilities in a way that prepares the audience for what will
happen. The small-scale visual narration in movies resides in devices such
as, most importandy, variable framing as well as various strategies for
directing attention internal to the single shot (e.g., placing the most impor-
tant element in the center of the image). Moreover, these systems of large-
scale movie narration and small-scale visual narration are coordinated by
the principle that the first thing the audience sees in a shot is that which is
most relevant to posing, sustaining, or answering the questions of the
ongoing story.
CINEMATIC NARRATION 207

While in general it is true that processes o f visual narration such as

variable framing are coordinated with the narrative for the purpose o f
emphasizing the first item, or gestalt o f items, seen by the audience, there
are standard deviations from this principle. These deviations are often
employed in thrillers for shock effect: the important subject, say the killer, is
hidden in the shot in such a way that the audience only comes to see him
belatedly (but unavoidably). In terms o f our account, these deviations are
not destructive counter-examples, for they still illustrate not only how the
flow o f narration is kept in strict control but also how important the
coordination principle is, since this convention (i.e., cultural invention) can
be used to hide elements in order to provoke surprises.
The narrative intelligibility o f a movie is, in large measure, a function o f
the coordination o f the large-scale, erotetic structure with processes o f
visual narration such as variable framing. The erotetic structure puts in
place a range o f audience expectations and the variable framing saliently
poses, sustains, and answers the questions o f the erotetic structure, gener-
ally by reframing events in such a way that what is most relevant to the
presiding questions o f the ongoing story is brought to the spectator's
attention first. The apparendy fragmentary flow o f imagery in the popular
movie is in fact structured by variable framing in a way that is subordinated
to the questions o f the large-scale, erotetic narration.
Erotetic narration, in coordination with variable framing and the other
visual devices for controlling the spectator's attention, gives the events and
actions portrayed in movies an unaccustomed intelligibility and coherence
when contrasted with the events and actions we generally encounter in
everyday life. Events are organized by presiding questions and broken
down by reframing in such a way that we attend, without distraction (or
virtually without distraction) to everything that is appropriate in the array
to the question at hand. In most cases, o f course, we are unaware o f the
questions, or, perhaps more accurately, the problematics that structure
most o f the events we encounter outside our movie theaters, and, o f course,
they are not framed for us in a way that neatly and economically organizes
them in order to make those features of said events which are relevant to
their problematic salient for observation. Movies, due to the coordination
o f erotetic narration and visual narration, have a hyper-clarity and cognitive
perspicuousness that pleases the mind through its exceptional intel-
ligibility.
This account of cinematic narration does not depend on any specific,
academically established psychological theory. It stresses the way in which
cinematic narration engages cognition, but it does not offer particular
hypotheses about the nature o f the specific cognitive-perceptual processing
208 CINEMATIC NARRATION

mechanisms that cincmatic narration exploits. I do not doubt that such

cognitive mechanisms or processes subtend the reception of the cinematic
narration in movies, nor do I think that it is beside the point to attempt to
identify them, perhaps by means of experimentation. However, I leave the
exact specification of those mechanisms to practicing, cognitive psycholo-
gists. My hope is that the characterization I have offered of cinematic
narration in movies may provide them with useful starting points for the
formation of further hypotheses.
I have attempted to characterize cinematic narration in movies without
resorting to psychoanalytic categories. I have tried to show the way in
which coherence is made in movies by means of the coordination of the
system of erotetic narration especially with variable framing. I contend that
this results in a portrayal of events notable for an especially high degree of
intelligibility, which intelligibility I regard as one of the features of movies
which make them so compelling for mass audiences. However, though I
emphasize the importance of intelligibility—of, indeed, a kind of hyper-
bolic intelligibility of movies in contrast to everyday experience—in ac-
counting for the powerful impression of movies, it should be clear that I see
no connection between the kind of intelligibility of represented events I am
discussing and the triggering of some sense of psychological subject-co-
herence in spectators. For me, the spectator recognizes the coherence of the
portrayed event and derives pleasure from that alone. Similarly, I do not wish
to claim that recognizing the intelligibility of a represented event in any
way predisposes the spectator to regard the portrayal as transparent, or to
accept the ideological insinuations that may attend an intelligible piece of
narration. The acceptance of the ideological message of a movie is not a
function of its use of forms such as erotetic narration and variable framing
but primarily a consequence of the beliefs the spectator embraces before he
or she enters the movie theater.

THE POWER OF MOVIES

In the preceding three chapters, we have seen that contemporary film
theorists have attempted to explain the ideological effect of film by means of
their analyses of the cinematic image, large-scale film narration, and pro-
cesses of visual narration such as the shot/reverse shot figure, point-of-view
editing, and so on. We have steadfastly denied that these structures are
inherently ideological, maintaining that ideology in film is a matter of the
content of specific films and their rhetorical organization, and of the gener-
ally antecedent belief systems spectators bring to films. We have also at-
CINEMATIC NARRATION 209

tempted to develop alternative accounts to those of contemporary theory

regarding the cinematic image, narrative, and the coherence of the image
track. Needless to say, these alternative accounts do not explain the way in
which these formal structures promote ideology, since we strenuously
challenge the validity of that view. However, interestingly, these alternative
approaches can be connected in a way which answers another question that
concerns some contemporary film theorists and which we encountered in
the first chapter, viz., what accounts for the powerful impression that film
imparts?
Baudry attempted to answer this question by examining the projection
apparatus of film from a psychoanalytic viewpoint. We, in turn, attacked
the plausibility of his theory of the apparatus. Furthermore, we are now in a
position as a result of the last three chapters to offer a countervailing view of
the power of cinema in terms of our findings concerning the cinematic
image, narrative, and the organization of the visual track in popular movies.
In one sense, of course, we cannot answer Baudry's question, for Baudry
wanted to account for the reason why film is powerful, which is why he
turned to an analysis of generic features of the medium such as projection.
Baudry, that is, offers a theory of why the medium of cinema, irrespective
of particular styles, genres, or films therein, imparts a powerful impression.
But we are really in a position to say only why movies—a quite particular
adaptation or use of the medium—are powerful. But nevertheless, I suspect
that upon reflection contemporary film theorists should recognize that
their question is really misconceived by Baudry, for the powerful impres-
sion they want explained is not really a function of cinema as such, but
rather an effect of one kind of cinema, namely movies. That is, a theory of
why movies are powerful is actually what theorists are searching for, not a
theory of why the medium itself is powerful.
The argument for the reorientation of our theoretical question from why
the cinematic medium is powerful to why movies are powerful is simple:
cinema as such is not powerful. It has not been surgical documentaries,
ballistics tests, or even modernist masterpieces that have made motion
pictures (including TV) the dominant mass art of the twentieth century,
one that is so widespread, internationally pervasive, accessible across
boundaries of class and culture, and that is intensely engaging. Rather it is
the adaptation of the medium for the purposes of movies, for the purposes
of Hollywood International, which has gripped such widespread audiences
so intensely. Not every type of cinema is powerful in the sense that it elicits
intense responses from global audiences. Thus, I believe, that when people
speak of accounting for the power of the cinematic medium, they really are
210 CINEMATIC NARRATION

talking about the power of one style of cinema, namely movies (including
narrative TV). For it is the movies, not experimental films or anthropologi-
cal records, that have captivated the twentieth-century mass imagination. It
is the power of movies about which researchers are really curious.
T o speak of movies rather than film or cinema deliberately eschews
essentialism. Posing the problem in an essentialist idiom—i.e., what makes
the medium of cinema powerful—perverts our question. And furthermore,
this has direct repercussions for research. For if, like Baudry, one thinks
that one is speaking of the power of cinema as such, one will be prone to
look for one's answer at a fairly generic level of investigation, such as the
projection apparatus, since that apparatus has some claim to being a com-
mon feature of all films. But once one realizes that since not all films impart
the powerful impression that make motion pictures such an extensively and
intensively engaging art, then one sees that the explanation of the powerful
impression that concerns us will not be a feature common to all films. That
is, plumbing the essence of the medium or some common feature of all
films, if there are such things, will not explain the phenomenon under
examination.
The powerful impression that theorists attribute (wrongly) to cinema as
such seems to comprise two factors: widespread engagement and intense
engagement. But not every style of film found in the medium is widely
accessible to nor intensely engaging of mass, popular audiences. Movies
are. Moreover, we are in a position now to identify those features of the
movie system which make them highly accessible to widespread audiences
as well as intensely engaging.
In our discussion of the cinematic image, as it is employed in standard
movies, we argued that such images could be recognized by untutored
spectators without the necessity of training in processes of symbol reading,
decoding, deciphering, or inference. Pictures, including motion pictures,
we contend, are the sorts of symbols whose capacity to be recognized is
acquired by percipients with the ability to recognize the objects or kinds of
objects the pictures represent. Pictures, then, are eminently transmissable
between cultures and their transmissability is not hindered by illiteracy.
Clearly, this suggests an important feature of movies that accounts for their
widespread accessibility across cultures and classes.
Another feature of movies that accounts for their accessibility is that
movies tend to be narrative, concerned primarily with depictions of human
actions. For narrative is, in all probability, our most pervasive and familiar
means of explaining human activity. If you ask me why George is watering
the tulips, I may answer that George intends to have, or wants, a beautiful
CINEMATIC NARRATION 211

garden, and that he believes that he can't have a beautiful garden unless he
waters the tulips. So I say he undertakes watering the tulips. You might ask
me how he formed the desire to have a beautiful garden. I may refer to
either his belief that this is a means of being a good citizen or to his guilt
about never caring for his father's garden, or both, if his action is overdeter-
mined. If you ask, where did he get the notion that the garden would not be
beautiful unless he watered it, I say he read it in a book called Beautiful
Gardens on May 17, 1953. Now if we tried to sum up this somewhat banal
explanation of George's action, a narrative would probably be the likeliest,
though not the only, means of organizing our information. George, racked
with guilt feelings about his father's tulips and convinced that a beautiful
garden is the means to the coveted ideal of good citizenship, decided to
have a beautiful garden; and when he read, on May 17, 1953, that such
gardens could not be had without watering the tulips, he went out and
watered the tulips (on May 18). We might add that he continued to do so
happily ever after. Insofar as this sort of narrative is one of the most
common forms of human explanation, and insofar as much movie narration
belongs to this category, movies will be familiar and accessible.
Moreover, the explanatory quality of such narration will also contribute
to the clarity of movies which, we have argued in this chapter and the
previous one, is relevant to the special intensity of our engagement with
movies. Indeed, in a nutshell, we might express our thesis about the power
of movies by saying that it resides in their easily graspable clarity for mass
audiences.
Erotetic narration and the use of visual devices such as variable framing
contribute to the special clarity of movies—to their heightened intel-
ligibility when compared to the typical series of events we encounter in
everyday life. Furthermore, this clarity, I submit, is the basis of our intense
response to and engagement with movies. Movies appeal to our cognitive
faculties by virtue of their forms. They answer questions that they vividly
pose and they do this by means of potentially very economical devices for
making relevant details salient.
We conceptualize the question of the power of movies as one concerning
the ways in which movies have engaged the widespread, intense response of
untutored audiences throughout the century. We have dealt with the issue
of the widespread response to movies by pointing to those features that
make movies particularly accessible. We have dealt with our intense en-
gagement with movies in terms of the impression of coherence they impart,
i.e., their easily grasped, indeed, their almost unavoidable, clarity. The
accessibility of movies is at least attributable to their use of pictorial repre-
212 CINEMATIC NARRATION

sentation, variable framing, and narrative, the latter being the most perva-
sive form of explaining human actions. Their clarity is at least a function of
variable framing in coordination with the erotetic narrative, especially
where erotetic narration and variable framing are coordinated by the princi-
ple that the first item or gestalt of items the audience apprehends be that
which, out of alternative framings, is most important to the narration. In
short, our thesis holds that the power of movies—their capacity to evoke
unrivaled widespread and intense response—is first and foremost, at least, a
result of their deployment of pictorial representation, erotetic narration,
and variable framing.
It will undoubtedly be noted that in this attempt to account for the
power of movies, we have restricted our purview to features in movies
which address the cognitive faculties of the audience. For only by focusing
on cognitive capacities, especially ones as deeply embedded as pictorial
representation and the drive to get answers to our questions, will we be in
the best position to find the features of movies that account for their
phenomenally widespread effectiveness; since cognitive capacities, at the
level discussed, seem the most plausible candidates for what mass-movie
audiences have in common. That is, the question of the power of movies
involves explaining how peoples of different cultures, societies, nations,
races, creeds, educational backgrounds, age groups, political affiliations,
and sexes can find movies easily accessible and gripping. Thus, the power of
movies must be connected to some fairly generic features of human organ-
isms to account for their power across class, cultural, and educational
boundaries. The structures of perception and cognition are primary exam-
ples of fairly generic features of humans. Consequently, it seems that if we
can suggest the ways in which movies are designed to engage and excite
cognitive and perceptual structures, we will have our best initial approx-
imation of their generic power.
Some qualifications are in order. First, we are not claiming that people
do not respond intensely to forms other than movies; indeed, some people
respond more intensely to other art forms than they do to movies. There are
opera buffs and balletomanes, after all. But this is compatible with the claim
that we are examining: that there is something special about the widespread
and intense, though not necessarily universal, response that movies have
been observed to command.
Next, we are not denying that there may be levers beyond those we have
discussed that also figure in the account of the power of movies. Marketing
structures, including advertising, are important elements as well as such
factors as the transportability and reproducibility of movies. Research in
CINEMATIC NARRATION 213

these areas should not be abandoned. However, considerations along these

lines do not obviate the present sort of speculation, since there still must be
something about the product, so marketed, that sustains interest.
Pictorial representation, variable framing, erotetic narration, and the
interrelation of these elements in the ways proposed will, at the very least,
be constituents of any account of the power of movies. I shall not pretend
to have offered a complete account of why movies are powerful—my mod-
esty signaled by the hedge "at the very least." Perhaps movies employ other
clarifying features, such as music (which will be examined in the next
section). Furthermore, apart from the question of why movies are power-
ful, we may wish to pursue different, but related, questions about why
certain movies or groups of movies are powerful for certain groups of
people; how do movies, or at least certain types of movies, engage particu-
lar classes, nations, generations, genders, and so on. Theoretical interest in
these questions would undoubtedly lead to a focus on elements of structure
and content that have not been addressed here, since we have just been
concerned with the generic power of movies, not the power of movies for
specific times, locales, sexes, age and interest groups. However, nothing we
have said suggests an objection in principle to these more specific ques-
tions, which questions, of course, will, in all probability, lead to speculation
about aspects of audiences over and above their cognitive faculties. Social
conditioning and affective psychology, appropriately constrained, might be
introduced to explain the power of given movies or types of movies for
target groups. Sociology, anthropology, and certain forms of psycho-
analysis are likely to be useful in such investigations. We can therefore
continue to examine the power of movies by asking about the power of
certain movies for historically specific audiences. However, if we wish to
explain the power of movies for the world community, then pictorial
representation, variable framing, and the erotetic narrative will be key
elements in our account because of the ways in which they address common
cognitive and perceptual capacities.

A C O N T R I B U T I O N TO T H E THEORY OF MOVIE MUSIC

This section differs from previous ones in several respects. Earlier we
were concerned with the visual elements of movies which contributed to
their narrative intelligibility. Here we shall be preoccupied with certain
aural, rather than visual, aspects of movies, which are not, stricdy speaking,
parts of the cinematic narration {qua representation), but which are inti-
mately related to it. This change of focus, however, appears useful both
214 CINEMATIC N A R R A T I O N

because it corrects the tendency within film theory to be overly preoccupied

with the visual elements offilmand enables us to illustrate the way in which
one might extend the investigation of the power of movies discussed in the
preceding section.
Another difference between this section and earlier chapters is that
previously our method of exposition was to outline the views of contempo-
rary film theorists on such matters as the cinematic image and narrative, to
subject those views to very detailed criticism, and to propose alternative
hypotheses. With movie music, however, we will forgo the intricate exposi-
tion and demolition of contemporary film scholars and plunge, almost
immediately, into our own theory. This change in our approach is dictated
by the fact that contemporary film theorists, though they have suggested
approaches to the musical track, have not really integrated it into their
theories in a manner explicit enough to sustain detailed criticism. Neverthe-
less, even though movie music is not a central topic of contemporary film
theory, exploring this topic will prove instructive about the way in which
our method of theorizing about movie music can serve as a contrast to the
general approach to theory endorsed by the reigning film establishment.
One contemporary film theorist who has commented on movie music is
Philip Rosen. 18 Rosen's project, however, is not to construct a theory of
movie music, but to explicate the theory of film music expounded by
Theodor Adorno and Hanns Eisler in the book Composing for the Films.19
But in expounding the Adorno/Eisler theory of film music, Rosen makes a
number of asides which suggest, very broadly, the way in which their
theory might be segued with the Althusserian-Lacanian paradigm, specifi-
cally in terms of suture theory. Rosen believes that the Adorno/Eisler
theory of film music "has important implications, for if the analysis has
correct elements, current studies of the suturing of the spectating subject
'into' the film may require greater attention to the musical track."20 Thus,
Rosen's reading of Adorno and Eisler suggests that movie music may have
a role to play in the positioning and construction of identity of the subject
in the Althusserian-Lacanian sense.21 Rosen also believes that Adorno and
Eisler might show us the way in which film music contributes to the illusion
of reality that cinema is said to promote, making the film image seem
natural and, presumably, transparent.
Adorno and Eisler contend:
Since their beginning, motion pictures have been accompanied by music.
The pure cinema must have had a ghosdy effect like that of the shadow play—
shadows and ghosts have always been associated. The magic function of
music . . . consisted in appeasing the evil spirits unconsciously dreaded. Mu-
sic was introduced as a kind of antidote against the picture. The need was felt
CINEMATIC NARRATION 215

to spare the spectator the unpleasantness involved in seeing effigies of living,

acting, and even speaking persons who were at the same time silent. The fart
they were living and nonliving at the same time is what constitutes their
ghosdy character, and music was introduced not only to supply them with
the life they lacked—this became its aim only in the era of total ideological
planning—but to exorcise fear or help the spectator absorb the shock.. . .
The sound pictures have changed this original function of music less than
might be imagined. For the talking picture, too, is mute. The characters in it are
not speaking people but speaking effigies, endowed with all the features of
the pictorial, the photographic two-dimensionality, the lack of spatial depth.
Their bodiless mouths utter words in a way that must seem disquieting to
anyone uninformed. Although the sound of these words is sufficiendy dif-
ferent from the sound of natural words, they are far from providing "images
of voice" in the same sense that photography provides us with images of
people.22

The ghostliness of the cinematic image, including the speaking, cine-

matic image, is counteracted by film music which "intervenes, supplying
muscular energy, a sense of corporeity, as it were." 23 The way in which this
is brought about is a bit obscure, but it appears to rely in part on the fact
that the music is experienced as in the same three-dimensional space as the
audience, lending it an air of corporeity in contradistinction to the two-
dimensional, ghosdy images; speech in film putatively does not have a
corresponding effect because it issues from ghostly mouths. 24
For Rosen, this account suggests the way in which film music may suture
an absence (and a purported contradiction[?!] between ear and eye)
thereby contributing the illusion of reality and of naturalness while, in the
process, positioning a subject. We have, of course, challenged the viability
and coherence of these imputed ideological effects in other contexts, so
there is no need to repeat previous criticisms of them at this point. Also,
Rosen introduces these speculations hypothetically, holding these effects
might occur if Adomo and Eisler are right. So the shortest way to deal with
Rosen's suggestions is to deny that Adorno and Eisler are right.
As in our encounter with suture theory, we initially note with respect to
Adorno and Eisler that their theory seems based on prettv flimsy evidence.
They claim that film spectators feel discomfort viewing cinematic images
because we find them ghosdy (though these ghosts are different from the
Absent One). Informal evidence for this might be that audiences are often
restless during silent films, unaccompanied by music, and frequently com-
plain about the silence. But perhaps we should take spectators at their
word. Perhaps it is the silence that strikes them and not some putative fear of
ghosts.
216 CINEMATIC NARRATION

Moreover, pace Adomo and Eisler, one does not encounter comparable
complaints with sound films. There is, to my knowledge, no evidence for
believing that spectators respond to sound films in the way that Adomo
and Eisler say they do. Like suture theorists, Adorno and Eisler base their
analysis on postulating a state in the spectator that has no basis in the data
of film viewing. One could, of course, test to see if spectators were uneasy
when confronted with talking films (without music), however, such a test
would have to take pains to assure that, if spectators showed signs of
discomfort or complained, this was not caused by the particular kind of
sound recording system employed. My own suspicion is that spectators will
not be bothered by talking films without music, if the dialogue is suffi-
ciendy interesting to them; and if they are disturbed that may be because
the dialogue is boring. But, in any case, whether I am right is less important
than the point that the burden of proof rests with Adorno, Eisler, and their
contemporary followers (if there are any). Without evidence, we have no
reason to accept the function of film music they postulate, nor any incor-
poration of that putative function into theories about the subject and
ideology. The way, that is, for an alternative theory of movie music is open.
Like Adorno and Eisler, I shall stress a functional relationship between
music and movies though, of course, that functional relationship will be
quite different than the one they propose. My position is closer to that
articulated by Schopenhauer when he writes in the Third Book of The
World as Will and Idea that "suitable music played to any scene, action,
event or surrounding seems to disclose to us its most secret meaning, and
appears as the most accurate and distinct commentary upon it."
There is not one and only one function that music can perform in relation
to movies. Aaron Copland suggested five broad functions: creating atmo-
sphere, underlining the psychological states of characters, providing back-
ground filler, building a sense of continuity, sustaining tension and then
rounding it off with a sense of closure.25 These do not seem to be neces-
sarily exclusive categories, nor do they exhaust the range of functions that
music can perform in movies. This is not said in order to criticize Copland,
for, in fact, I intend to follow his example. I will analyze a function of movie
music, freely admitting that there are others, and, moreover, I will not deny
that this function may also be yoked together with the performance of
other functions, such as those Copland enumerates.
The type of music I will discuss is quite central in popular, mass-market
movies; it is a basic use of music, if not the most basic. I call this use of
music modifying music, and in what follows I will attempt to describe its
structure, to explain how it works and how it fits into the system of popular
expression called the movies.
CINEMATIC NARRATION 217

To begin to approach modifying music, it is helpful to consider some

examples. In Gunga Din, there is an early scene where the British, led by
Cary Grant, Douglas Fairbanks Jr., and Victor McLaglen, enter a seem-
ingly deserted village in search of foul doings. Indeed, the village has been
raided by the nefarious Thugs, and those dastardly followers of Kali are
lying in wait for the British. We have been somewhat alerted to this insofar
as the scene is initiated by the use of an oboe in imitation of the sort of
double-reed instrument associated with snake charmers, thereby signaling
the presence of the Thugs in the deserted village. There is an ambush.
During the ensuing battle, there is a recurring theme that is associated with
the efforts of Grant, McLaglen, and Fairbanks. Earlier, we had heard the
same theme accompanying their drunken brawl over a phony treasure map.
In the ambush scene, an interlude of strings will be followed by horns at a
scherzo-like tempo. Often this theme comes in when our soldiers of fortune
gain the upper hand, but not always. The horns are bouncy, light, and
playful. The batde scene, full of death and danger, could be the object of
high anxiety. But the use of the horns in this theme colors the scene in such
a way that we come to view it as a lark, as cheery, as a game, as comic rather
than potentially tragic. This, of course, corresponds to one of the views of
war and manhood that the film promotes—i.e., war as an oudet for boyish,
beamish energy.26 Of course, from our point of view, what is important
about the scene is the way in which the scherzo-like refrain directs the
audience to view the mayhem as jaunty—almost comic—good fun.
In Rebel Without a Cause, we find a wholly different feeling associated
with the onscreen violence. Underlying the confrontation and the fight,
called the blade-game, which occurs after the visit to the planetarium, is
atonal music, marked by odd time signatures and dissonant blaring brass.
The use of the timpani and horns, along with the timing, gives the music a
Stravinsky-like flavor. As well, the music is sometimes recorded low, and,
then, abrupdy, the recording level is raised. The dissonance imparts a
brooding feeling to the scene, a sense of latent, almost muscular violence
that flashes out when the brass blares or the recording level shoots up. The
uneasy, unstable quality of the music serves to characterize the psychologi-
cal turmoil—the play of repression and explosive release—with which the
scene, and the movie, is concerned.
For an example not involved with violence, consider the opening of The
Yearling. The camera displays views of the Everglades as Gregory Peck,
playing a Civil War veteran, recalls how he came to make his home there.
The score is dominated by strings which have strong connotations of
richness and lushness, reflecting, of course, the way in which we are to take
this place. What Peck's voice and the visuals may fail to make you realize
218 CINEMATIC NARRATION

about the landscape, the music enables you to grasp. Also, the strings have a
slightly haunting flavor and a sense o f pastness which coincides with the
appearance this film suggests o f being swathed in memory. When we are
introduced to the juvenile lead Jodv (played by Claude Jarmon Jr.), the
music sounds somewhat pentatonic, like an elongated country melody,
conveying a feeling that is both lazy and dreamy. This not only corresponds
to what we immediately see o f J o d y — h e is playing lisdessly with a toy
windmill—but to what we learn o f Jody throughout the film, viz., that he is
a dreamer. In terms o f the subject matter o f the movie, a major source of
tension between Peck and his wife, played by Jane Wyman, develops
because Peck believes that youth should be a time when the imagination is
given its head, before the hardships and responsibilities of practical life
force one to turn to sterner things. Wyman resists this, and the battle
between youth and imagination, on the one hand, versus adulthood and
practicality, on the other, is staged over Flag, the yearling from whom the
film derives its tide. Throughout The Yearling, the use of the strings
repeatedly stresses the theme o f imagination by underscoring and charac-
terizing the various spoken reveries and gambolings o f characters in terms
o f an undeniable, albeit very nineteen-fortyish, feeling of dreaminess.
These examples are not alike in every respect. The theme from Gun/fa
Din functions narratively as a leitmotif, whereas the example from Rebel
Without a Cause does not. However, the three examples share a very basic
function, one which in fact enables the theme from Gunga Din to d o its
more specialized work s o well. Namely, in each o f these examples the music
characterizes the scene, i.e., imbues the scene with certain expressive prop-
erties. This may be a matter o f enhancing qualities that are already sug-
gested in the imagery, but it need not be; the music may attribute to the
visuals an otherwise unavailable quality. Nor does the expressive quality in
question have to be grounded in the psychology o f a character; in the
Gunga Din example the jauntiness of the music appears to attach first and
foremost to the action rather than to internal states o f characters. And,
lastly, the expressive qualities projected in these examples are in the music.
We d o not suddenly become dreamy when we hear the strings o f The
Yearling. Rather the dreaminess o f the music characterizes Jody as dreamy
to us. If we are pro-dreaminess, the way Gregory Peck and the film are, then
we are apt to feel sympathetic (rather than dreamy) in regard to Jody. That
is, by speaking o f the projection o f expressive qualities, we are not claiming
that the music arouses in spectators the selfsame feeling qualities that it
projects.
We can call this use o f movie music modifying music. The music modifies
CINEMATIC NARRATION 219

the movie. The music possesses certain expressive qualities which are intro-
duced to modify or to characterize onscreen persons and objects, actions
and events, scenes and sequences. To use a crude analogy, one which must
eventually be abandoned, the visual track is to a noun as the music is to an
adjective, or, alternatively, the visuals are to a verb as the music is to an
adverb. Just as adjectives and adverbs characterize, modify, and enrich the
nouns and verbs to which they are attached, modifying music serves to add
further characterization to the scenes it embellishes. This is a very pervasive
use of movie music. Let us now turn to a discussion of its internal dynamics.
Movie music involves coordinating two different symbol systems: music
and movies, the latter including not only visuals but recorded sounds, both
natural and dialogic. In the case of modifying music, these two symbol
systems are placed in a complementary relationship; each system supplies
something that the other system standardly lacks, or, at least, does not
possess with the same degree of effectiveness that the other system pos-
sesses.
Music, for example, is a highly expressive symbol system. This is not to
say that all music is expressive or that it should be expressive, but only that
much music is expressive. For example, that the Prelude to Tristan and
Isolde is expressive of yearning or that the "Great Gate at Kiev" from Pictures
at an Exhibition is expressive of majesty are part of the incontestable data of
aesthetic theorizing. To say that music is expressive is to say that it projects
qualities describable in anthropomorphic, emotive terms. The symbol sys-
tem of music is also sometimes thought to have more direct access to the
emotive realm than any other symbol system. Nietzsche called music "the
immediate language of the will."27
At the same time, it is often noted that nonvocal music—orchestral
music28—though quite effective in expressing a broad palette of emotions,
is not the ideal means for particularizing the feelings it projects. That is, a
piece of nonvocal orchestral music may strike us as sorrowful or even more
broadly as "down," but we generally cannot specify much further the kind
of dolors or dumps the music projects. Is it melancholic, neurasthenic,
suicidal, adolescent, etc. ? That is, nonvocal music standardly lacks what the
philosopher of music Peter Kivy calls emotive explicitness.29
This lack of emotive explicitness has figured in numerous debates in the
history of music. Some, like Johann Adam Hiller, took it as a limitation to
be overcome, urging that if music is to become intelligible, i.e., emotively
explicit, it must be combined with speech.30 A similar view was espoused
by James Beattie, who held that "the expression of music without poetry is
vague and ambiguous."31 Peter Kivy has brilliantly demonstrated that the
220 CINEMATIC NARRATION

development of the expressive arsenal of Western orchestral music, as we

know it, was the result of solving the perceived problem of music's emotive
inexplicitness through text setting. 32 In a different mood, Eduard Hanslick
argued against the expression of emotions as a goal of music because he
believed that music cannot express definite emotions, while Nietzsche,
staking out an altogether different position, sees the emotive inexplicitness
of music as the path to some coveted form of universality: "whoever gives
himself up entirely to the impression of a symphony, seems to see all
possible events of life and the world take place in himself." 33
The vicissitudes of the preceding positions are less important to us than
their recurring assumption, which we shall state weakly as follows: typ-
ically, nonvocal music is expressive of emotive qualities but ones that are
inexplicit, ambiguous, and broad. A theoretical explanation of why this
should be is also readily available. Emotions are directed, directed at per-
sons, objects, states of affairs, and events. Indeed, it is in virtue of the
intentional objects to which emotions are directed that we individuate
emotions. 34 1 am afraid ofbeing run over by a train; you are in love with Bob;
we are angered by apartheid. For an emotion to be fully explicit and
particularized, it must be aimed at some object. The object may be real, like
South Africa, or fantasized, e.g., you may be terrified of the Green Slime.
To become explicit, that is, the emotion must be referred to something. To
say whether the joy in the music is hysterical or Utopian, we would have to
know toward what the joy was directed. And, of course, it is this sort of
reference that is most commonly absent from music, that is, nonvocal
music. 35 Insofar as representation is not a primary function of standard
orchestral music, most music of this sort will lack the logical machinery to
secure emotive particularity. This is not to say that orchestral music cannot
be representational: e.g., Wellington's Victory, Honegger's Pacific 231, and
the use of percussion to refer to King Kong's offscreen footsteps in the film
of the same name. 36 And where the music is representational, a measure of
emotive explicitness may be achievable. However, as I have said, as a matter
of fact, most nonvocal music lacks the logical machinery which emotive
explicitness requires.
So far I have claimed that orchestral music of the sort often employed in
movies is a symbol system that makes a powerful yet broad and inexplicit
emotive address. And this inexplicitness, in turn, is a result of the fact that
generally such music is non-referring. Movies, on the other hand, are
symbol systems with numerous overlapping referential dimensions, includ-
ing the cinematographic image, dialogue, narrative, and synchronized
sound. Wedding the musical system to the movie system, then, supplies the
CINEMATIC NARRATION 221

kind of rcfcrcncc required to particularize the broad expressivity of the

musical system. The dreaminess of the strings in The Yearling is specified as
Jody's dreaminess, as the dreaminess of a young boy prior to the hard
lessons of life.
The relation between the music and the movie in the case of modifying
music is reciprocal. The movie—the visuals, the narrative, the dialogue,
and the synchronized sound—serve as indicators. At one level, these ele-
ments establish what the scene is about. They indicate the reference of the
scene. The music then modifies or characterizes what the scene is about in
terms of some expressive quality. In a manner of speaking, the music tells us
something, of an emotive significance, about what the scene is about; the
music supplies us with, so to say, a description (or, better, a presentation)
of the emotive properties the film attaches to the referents of the scene.
In our Gunga Din example, the movie establishes the subject, the battle,
and the music imbues it with a feeling, that of jauntiness.37 The musical
element, which I call a modifier,fillsin the subject matter in terms of the
feeling the filmmaker finds appropriate to the scene. However, at the same
time, the movie elements, what I have called the indicators, stand in an
important relation of influence to the musical component. The music on its
own is bouncy, light, and comic. When conjoined with the movie elements
those feelings become further particularized as manly, dare-devil bravado.
The musical system, so to speak, carves out a broad range or spectrum of
feeling, in this case, one that is positive, lively, and energetic. The movie
elements, the indicators, then narrow down or focus more precisely the
qualities in that range or spectrum that are relevant to the action. The music
no longer signals mere energy but more precisely bravado. This focusing
operation of the movie-as-indicator, in turn, enables the music-as-modifier
to fill in the action as a highly particularized feeling.
It might be initially helpful to think of the relation of the movie-as-
indicator to the music-as-modifier on the model of the subject-predicate
relation: the music says " . . . is jaunt)'" and the movie specifies the blank
with "the battle." However, though suggestive, this analogy cannot be
taken too seriously because the movie elements perform functions other
than referring and focusing, and because the linguistic notion of predica-
tion seems to be strictly inapplicable to motion cinema (i.e., pictures lack
discriminate subject-predicate elements and show objects with their prop-
erties, all at once, so to speak). Thus, though modifying music resembles
linguistic predication loosely, it should not be taken as a literal example
of it.
Another possible avenue of misunderstanding modifying music would
222 CINEMATIC NARRATION

be an oversimplification that regards music as exclusively expressive and the

movie components as exclusively representational. As was earlier remarked,
music can be used representationallv. Similarly, movie elements have mvr-
iad means of expression—not only through acting, but through lighting,
camera movement, camera angulation, cutting, etc. Indeed, the generally
referential soundtrack can be "musically" arranged in order to aspire to
musical expressivity, e.g., the natural sounds at the opening of Street Scene
and the dialogue in Force of Evil. Thus, it is not the case that the movie is
pure representation to be supplemented by means of musical expression.
However, in reaching out for music, the movie is seeking to incorporate an
added, particularly powerful, augmented means of expression along with
the visual, narrative, and dramatic means already at its disposal. The addi-
tion of music gives the filmmaker an especially direct and immediate means
for assuring that the audience is matching the correct expressive quality
with the action at hand. This is not to say that music is the film's only
expressive lever; rather it is a notably direct and reliable one. It enhances the
filmmaker's expressive control over the action.
If adding music to the movie enhances one's expressive control over the
action, it is also the case that the imagery intensifies the impact of the music
by particularizing its affective resonance. The unnerving, shrieking strings
in Psycho are cruel, painful, and murderous when matched with Norman
Bates' descending knife. Here, the reference afforded by the movie ele-
ments serves to individuate the emotive content of the music in the way that
narrative and pantomime do in ballet, and as words do in a popular song or
opera.
Modifying music is one of the major uses of music in popular movies. It
may be used to embellish individual scenes and sequences, or it may be
integrated into leitmotif systems. Structurally, modifying music involves
the use of movie elements—photography, narrative, dialogue, and syn-
chronized sound—as indicators that fix the reference of a shot, scene, or
sequence. The associated musical elements are modifiers which attribute
expressive qualities to the referent, thereby characterizing it emotively as,
for example, dreamy or jaunty. Functionally, the addition of musical modi-
fiers to the scene augments the expressivity of the scene, though this does
not preclude the possibility that the scene already possesses many nonmusi-
cal expressive devices. Nevertheless, music is a particularly privileged means
of direct, expressive augmentation. The musical modifiers function to fill-in
the scene expressively, to set the expressive tone the filmmaker takes to be
appropriate to the scene. The music "saturates" the scene expressively. At
the same time that the musical modifiers influence the reception of the
CINEMATIC NARRATION 223

movie, the movie indicators also reciprocally influence the reception of the
music. For music typically, sans referential machinery, projects a very broad
and inexplicit range of emotive qualities. Thus, in The African Queen, when
the boat is stuck in the channel, the slow, spaced-out drum beats project a
generic, plodding feeling while the movie elements specify that feeling as
Bogart's effort, an effort charged with all his hopes and commitments.
Thus, as the music fills in the movie, the moviefocusesthe emotive content
of the music, particularizing and intensifying its effect which, of course,
also abets the filling-in work that the musical modifier does.
I have attempted to explain the way in which modifying music operates.
Modifying music is not employed, of course, only in movies—it occurs in
other sorts of films, such as art films, as well as in other art forms, such as
ballet. As well, it is not the only use of music found in movies. Yet, though
the relation between modifying music and the movies is not unique in any
sense, there is a way in which modifying music serves the aims of the movie
system quite expeditiously. Thus, I will conclude this sketch of modifying
music by discussing the way that modifying music segues into the economy
of the movies.
Movies are a means of popular expression. That is, they are aimed at mass
audiences. They aspire for means of communication that can be grasped
almost immediately by untutored audiences. Another way of putting this is
to say that moviemakers seek devices that virtually guarantee that the
audience will follow the action in the way that thefilmmakerdeems appro-
priate. The movie close-up, for example, as we saw in the preceding section,
assures thefilmmakerthat the spectator is looking exactly where she should
be looking at the appropriate moment. Similarly, modifying music, given
the almost direct expressive impact of music, assures that the untutored
spectators of the mass movie audience will have access to the desired
expressive quality and, in turn, will see the given scene under its aegis.38
Second, an important element accounting for the power of movies is the
clarity that movies bestow upon the events that they depict. In contrast to
our encounters in everyday life, movie events have an unaccustomed intel-
ligibility and lucidity; movies, that is, are so much more legible than life.
Modifying music contributes to the clarity of movies in several different
respects. The filling-in function of the music modifier keeps the expressive
quality of the scene constandy foregrounded, thereby supplying a contin-
uous channel of information about the emotional significance of the action.
Unlike our quotidian experience of events, the music constantly alerts us to
the feeling that goes with what we see. Whereas in life, the affect that goes
with an observation is so often unknown, in movies, we not only have some
224 CINEMATIC NARRATION

affect but also the appropriate affect tied to virtually everything we see,
through modifying music. The movie world is emotionally perspicuous
through and through.
Reciprocally, the focusing function of the movie indicators render the
emotive content of the music more and more explicit, again enhancing their
clarity in yet another way. The concerted interplay of the music and the
movie yields images replete with highly clarified, virtually directly accessi-
ble, expressive qualities. Thus, though modifying music is not a unique
feature of movies, its capacity for promoting immediately accessible, ex-
plicit, and continuous emotive characterizations of the ongoing, onscreen
action makes it so suitable to the presiding commitments of mass movie
communication that it would be a mystery had movies failed to exploit it.
The suture theorist, hearing this account of modifying music—or, for
that matter, our earlier accounts of erotetic narration and variable fram-
ing—might respond that all I have done is to isolate the coherence making
features of movies which enable suturing to occur. That is, a contemporary
film theorist could be tempted to say that my theories are merely prepara-
tory to suture theory. Suture theory can integrate these small-scale theo-
ries—of movie imaging, narrating, and music—in the larger theory of
suture. Once we isolate the nitty-gritty mechanics of the way in which these
processes operate, the suture theorist can step in and appropriate these
accounts of cinematic coherence, adding, of course, that with coherence (or
the appearance of coherence) comes subject positioning. Perhaps with
regard to modifying music, a suture theorist might say that it positions the
subject emotively, or that the music sutures the absence of affect from the
image track.
Of course, we can resist the imperialism of suture theory by demanding
to know why, once we have offered a small-scale explanation of the way in
which modifying music—or erotetic narration or variable framing—oper-
ates, we should go on to postulate subject positioning? And we can, for
example, ask what features of the data are left out of our explanations such
that we need to hypothesize a psychological state of suturing?
Whereas our theories, such as that of modifying music, are small-scale
theories, suture theory, in its expanded version, is a large-scale theory, one
designed to subsume theories of cinematic imaging, narration, and music
under its abstract vocabulary of subject positioning. But this may indicate
that suture theory, as well as other articulations of contemporary theory,
are not particularly useful frameworks for answering our questions about
film. For they are not likely to shed light on the way in which specific types
of cinematic configurations function but rather say of each of them: thev
CINEMATIC NARRATION 225

suture, or they position subjects. However, these conclusions follow theo-

retically for contemporary film analysts once one acknowledges that a
cinematic configuration is coherent. Thus, these explanations do not
account for the way in which specific, coherent cinematic patterns are
constructed. In the following "Conclusion," I will attempt to assess the
methodological advisability of the large-scale theorizing preferred by con-
temporary film theorists versus the kind of small-scale theorizing—such as
the theory of modifying music, erotetic narration, variable framing, and
even of the power o f movies—that I endorse.
 CONCLUSION
PROBLEMS AND PROSPECTS OF
CONTEMPORARY FILM THEORY

I n the seventies and eighties, in the United States, cinema studies, as an

academic discipline, burgeoned. As a rising area o f university study, film
scholarship sought to legitimatize itself. And, as well, the rapid expansion
o f cinema studies entailed the entry o f many young scholars into the field.
These young scholars, given the massive, unwieldy, and virtually disor-
ganized nature o f their data, searched for a ready-made framework with
which to regiment their materials. The result of these pressures was the
adoption o f what, in this book, we refer to as "contemporary film theory
an amalgam o f Anglo-French vintage, most often comprised, at least, o f
Althusserian-derived Marxism, Barthesian textual criticism, and, most im-
portandy, o f Lacanian psychoanalysis.
The theme o f this book has been that "contemporary film theory" has
been nothing short o f an intellectual disaster and that it should be dis-
carded. In concluding my brief against contemporary film theory, however,
I will not recapitulate the particular objections that were mounted in
previous chapters, but rather will attempt to offer some general summary
remarks about the liabilities o f contemporary film theory.
The most obvious, recurring problem with contemporary film theory is
that its central concepts are often systematically ambiguous, due in some
cases, but not all, to their essentially metaphorical nature. A consequence o f
this is that, in the main, the arguments and analyses o f contemporary film
theory turn out to be little more than extended exercises in equivocation.
For examples o f ambiguity, recall how we saw in our opening chaptcr
that both the immobility o f the eighteen-month-old child and the seated
CONCLUSION 227

position of thefilmviewer were described in terms of inhibited motoricity.

Also, under the rubric absence, Metz considers both the absence of the
referent of thefilmimage from the screening room and the "absence" of the
film viewer from the world of the film as cognate phenomena. But surely
these concepts are being stretched to the breaking point. Indeed, terms like
suture and mirror, in the language of contemporary film theory, can be
extended to virtually anything. This, of course, is not testimony to the
explanatory power of these concepts, but rather evidence that they are ill-
defined and ambiguous in ways that render them theoretically useless. That
is, for a term to be theoretically adequate we should have a clear sense of
when not to apply it as well as a sense of when to apply it; it should have
some rigorously stipulated parameters. For only then can we rest assured
that the concept has something relevant to tell us about whatever specific
phenomenon we are investigating.
The tendency toward ambiguity in contemporary film theory is abetted
by what might be called a penchant for Platonizing. All different sorts of
desire, such as a male viewer's sexual desire for a movie character and any
viewer's desire that the movie be intelligible, are slotted under the abstract
noun Desire, whose laws the Lacanian then charts. One, of course, wonders
whether, ontologically, there is such a thing as Desire per se, rather than
particularized desire* for this or that. Desires, that is, are individuated by
reference to their objects (e.g., the desire fir ice cream) not as instances of
some unified, univocally named force called Desire. Similarly, something
may be absent from a movie theater or absent from the world of the film
without being the occasion of some abstract Absence of which castration is
also an instance. By trading on these vague nominalizations, contemporary
film theorists obfuscate the phenomenon at hand rather than clarifying it.
As has been noted already, many of the putative explanations in contem-
porary film theory are couched in metaphors and analogies; "suture" and
"mirror," for example, were mentioned above. Now some may feel that the
problem here is that metaphors and analogies have no place in theoretical
or scientific discourse.1 However, such a view seems too draconian and I, at
least, would agree with those who see a legitimate role for metaphors and
analogies in theorizing.2 But although I think that metaphors can be
scientific, I do not believe that the use of metaphor and analogy in contem-
poraryfilmtheory generally accords with acceptable scientific or theoretical
practice.3
The function of a scientific metaphor or analogy is to give us knowledge
about some phenomenon that we know little about by reference to some-
thing we know more about. The metaphor of waves is particularly useful in
228 CONCLUSION

physics. But in order for a scientific metaphor to work, we must have a

distinct picture o f the known phenomenon that is being used to illuminate
the unknown phenomenon. However, in contemporary film theory, the
supposedly explanatory metaphors are often as vague or more vague than
the topic they are meant to elucidate. For example, we often hear, in
contemporary film theory, that the subject is turning or sliding during
subject positioning. These images—of turning and sliding—are intro-
duced in order to help us grasp obscure psychic processes. But forget the
psychic processes for the moment. Now concentrate. Soon you will ask—
just what are we supposed to envision as concrete examples of sliding and
turning. Is the sliding done in a groove or on a slippery surface? Is the
turning like that of a rock suspended by a string, like a planet on its axis, or
like the needle of a compass? That is, what concrete physical forces arc we to
have in mind? For if we have none in mind, how can we extrapolate from
them so as to understand the even more obscure psychic processes that
supposedly correlate to them?
The metaphors and analogies o f contemporary film theory are generally
so vague and abstract that they are not internally rich enough to supply us
with a picture of anything, let alone with a template with which to trace the
oudine o f obscure psychic processes. We may call this the fallacy of the
indigent metaphor, and it is rampant in contemporary film theory.
It might be said that such metaphors as "turning" and "sliding" do
perform some service. They tell us that the subject is slippery and in
motion. But this is not particularly informative unless we have a grasp of
what such subjects are slippery, and, in motion, in reference to. The meta-
phor will be viable in proportion to the clarity of its focusing term.
Moreover, a strong scientific metaphor should have some degree of
systematicity and fecundity. That is, it should enable us to expand upon it.
If A and B are similar in respect to property x, and if A, the focusing term of
our metaphor, also has properties y and z, then we test to see if properties
analogous to y and z can be found in B. But since the metaphors of
contemporary film theory are often too vague, we cannot use them to
articulate further the structures and forces within the psychic processes in
question. Also, when we have a firm grasp of the processes referred to by
the focusing term of a metaphor, we can use that to predict the next stage of
the process that is mysterious to us. This feature of scientific metaphors is
also related to teaching the metaphor—that is, the student can use the
metaphor to see how the mysterious process hangs together as a sequence
because the student has a clear picture of the process in the focusing term of
the metaphor. But with examples such as the hazy "slidings and turnings"
CONCLUSION 229

of the subject in contemporary film theory, it is anyone's guess as to what

precedes or follows them.
As we have seen throughout this book, the reliance on metaphors and
analogies in contemporary film theory has time and again led to arguments
and analyses that exemplify the fallacy of equivocation. One might indeed
say that contemporary film theory is little more than a tissue of puns. As a
general method of refutation, one should replace all the qualifications that
the contemporary film theorist deletes and the desired "emperor has no
clothes effect" will follow almost immediately. For example, in Questions of
Cinema, Heath wants to make the rather startling point that narrative is "a
decisive instance of framing in film."4 Thus does contemporary film theory
attempt to equate pictorial mimesis with diegesis. Heath supports this
discovery by citing the importance of narrative closure. He also feels that, in
some way, the narrative "encloses" the space of the film. He wants us to
think of narrative as a literal frame. But what does narrative closure have to
do with a frame? If one says the narrative encloses the space of thefilmone
uses an extremely loose metaphor that means the space of a narrativefilmis
coherent in terms of the narrative. A narrative does not, however, enclose a
visual expanse by surrounding it in the way a frame does. The argument—
that a narrative encloses and, therefore, frames—is, consequendy, not
sound. For once we add its missing premise—all frames enclose (in the
qualified sense of surround)—we see that the way narratives enclose is not
really covered by the relevant, qualified (though previously deleted) gener-
alization. By searching for equivocations, one can derail a surprising num-
ber of the analyses of contemporary film theory. Since there is an especially
high incidence of equivocation in contemporary film theory, a reader must
patiendy attend to every different meaning of words like center and frame
within a given analysis.
The equivocations in contemporary film theory are encouraged by the
fact that many of the main concepts in the system are associatively inter-
substitutable. Meaning, closure, coherence, position, binding, frame, center,
homogeneity, unity, balance, and so forth are either cognates for the contem-
porary film theorist, or so enmeshed that mention of one licenses the
invocation of any of the others (despite the apparent disjunctiveness of
many of these concepts—e.g., to have a meaning is not a necessary or
sufficient condition for being balanced, and vice versa). This allows the
contemporaryfilmtheorists to move rapidly from claims such as "narratives
have closure" to "narratives frame" to "narratives center." These associative
trains run over the differences within the phenomena under discussion. But
what really grounds these associations? In fact, they are not rigorously
230 CONCLUSION

dcñned and interrelated theoretical terms. Rather, they are garden-variety

image clusters—words grouped together in ordinary language because
they share rather broad connotative qualities. Center and unity go together
in the same way that gap and absence could be grouped with cold and gray,
or, to use a classic example, with ping (rather than pong). These words, in
short, seem interchangeable because they share affective resonances and not
because the contemporary film theorist has shown us that they represent
the same phenomena in every case, or even in many cases. The contempo-
rary film theorist acts as though his terms are interdefinable when they
really only belong to similar emotive meaning clusters. However, what is a
boon for poetry is a bane for theory.
The ambiguity of the central concepts of contemporary film theory is
perhaps connected to the other major liability of this approach, viz., its
ambition. Contemporary film theory is engaged in the totalizing attempt to
erect The Theory of Film, i.e., a theory that contrives to explicate in one
unified theoretical vocabulary queries into issues as diverse as the mecha-
nism of point-of-view editing, the nature of the avant-garde, the mecha-
nisms of movie advertising, the nature of the soundtrack, the nature of the
camera lens, the operation of ideology, etc. That a theoretical vocabulary
that answers so many different kinds of questions in the same terms (e.g.,
subject positioning) tends toward ambiguity is predictable (though still
inexcusable for the reasons given above). However, we should question not
only the ambiguous rhetoric of contemporary film theory but also the
advisability of attempting to construct a unified theory of film.
I have no argument to show that a unified theory of film is impossible.
But one can suggest the ways in which the attempted construction of such a
theory, in the manner in which contemporary film theorists pursue it, is
likely to go wrong. Contemporary film theory is "top down." From their
readings of such people as Lacan and Althusser, they derive a general
theory, such as that of subject positioning, and then they attempt to graft
that account onto specific phenomena, such as point-of-view editing or
movie advertisements, usually by means of equivocation or some other
exercise of ambiguity. That is, they apply the general theory to each aspect
of film which they want to explain. The problems this creates in their
reasoning has been explored, as well as the fact that the general theory they
employ is, to say the least, questionable. However, the "top down" ap-
proach is also suspect in a way that can be readily confirmed by a patient,
"nonaligned" reader—viz., everything you ever wanted to know about film
gets roughly the same answer, most often in terms of subject positioning.
Movie ads, genre films, avant-garde films, point-of-view shots, movie mu-
CONCLUSION 231

sic, stars, and so on are all explained by the same subject positioning
formula with its interplay of the Imaginary and the Symbolic, and of
Presence and Absence. Similarly, if the theory is employed in the analysis of
individual films, we find, surprisingly, that avant-gardists like Malcolm
LeGrice are indulging the Imaginary no less than a Sylvester Stallone. That
is, the explanatory bottom line turns out to be pretty much the same no
matter what film, and no matter what level of filmic phenomena, is being
discussed. This is not only downright monotonous, it is theoretically
threadbare, if not vacuous.
At this time, a "top down" theory of film seems inappropriate. For by
explaining every aspect of film by means of the same formulas it blurs our
understanding of specific aspects of film. The answers it supplies to our
questions are too general; specific films, specific film forms, and specific
filmic articulations are all painted with the same theoretical brush. But
surely we expect theories to tell us about the characteristic workings of
specific phenomena rather than about what everything has in common.
There are many things that we would like to know about film. Why do
mass market movies have such an intense appeal to so many different types
of audiences? What is the nature of film metaphor? What are the images of
women in American film? How does movie music work? What are the
characteristic structures of film editing? In what ways does film disseminate
ideology? Can documentary films be objective? What is the structure of the
movie market? Can avant-garde films save the world and should they be
expected to? How are film narratives processed cognitively? What is a
movie genre? What is the structure of the film rtoir? And so on. These
questions address different levels of abstraction and there is no reason to
expect that they can be adequately answered within one theoretical frame-
work or be reduced to a few formulas. Indeed, there is more reason to
suppose that if all these questions are answered in the same way, the theory
is likely to be, in fact, uninformative.
It is, of course, my claim that contemporary film theory is uninformative
and overly general in exacdy this way. By attempting to answer all our
questions of cinema from the overarching framework of Marxist-psychoan-
alytic theory it has given us a small package of slogans instead of explana-
tions. And even those who would reject my overall diagnosis here, I believe,
would admit that contemporary film theory, after a burst of energy in the
seventies, is languishing in the mid-eighties.
On my account, contemporary film theory is totalizing (i.e., pretending
to be The Theory of Film, as well as a lot else), and is "top down." My
suggested remedy for malaise of film theory is that it become piecemeal and
232 CONCLUSION

"bottom up." Instead of searching for the unified theory of film, we should
attempt to carve out clear and manageable questions about aspects of
film—such as " H o w does the Art Cinema work?"—and then go on to
answer them. This, at the very least, has great heuristic value. It is far less
daunting to ask and answer questions, for example, about the structures
and compositional potential of wide-screen cinematography, than it is to
determine what Film Theory, construed as a single, totalized system,
should comprise. Nor—and this is more to the point—is there any reason
to believe that all the theoretical interests we have in film will fall neady into
a single systematic framework. We may wish to know how Hollywood star
imagery affects American voting habits; and we may wish to know how
certain editing patterns suggest causation between spatially discontinuous
objects. Answers to these questions call for film theort«, i.e., piecemeal
theories about aspects of film. Indeed, looking at the diversity of questions
we have about film—which stand on different levels of abstraction and
which call upon the resources of very different disciplines—should lead us
to embrace at least an initial presumption that these questions are not likely
to be answered by some overarching Film Theory that systematically inter-
relates everything one wants to know of film. In fact, the desire for such a
Film Theory may stand—and I would contend that it has stood—in the
way of acquiring theoretical insight into the workings of cinema.
Currendy, then, we are in need of piecemeal theorizing, in need of
theorr« about film rather than Film Theory—be those theories sociologi-
cal, psychological, philosophical, narratological, or devoted to the study of
audience reception, or to the quasi-rhetorical study of specific film devices,
structures, genres, etc. Perhaps some of the fruits of our various piecemeal
theories will be organizable into larger, systematic and theoretical constella-
tions. In the chapter preceding this one, for example, I attempted to
coordinate an account of the workings of movie music with a broader view
of the function of movies. Just as there is no reason to believe that there is
an overarching Film Theory, there is no reason to preclude antecedently
that some of the answers to our piecemeal theories can not be coordinated
into larger theoretical complexes. But at this point, all we can do is to
generate small-scale theories, watching out of the corner of our eye to see if
their results can be gathered into larger theoretical constructions. This, of
course, is what is meant by urging that our approach to theory be "bottom
up," viz., that if more comprehensive and complex large-scale theories can
be derived, they should be derived from the comparison and scrutiny of
what we regard to be successful piecemeal theories.
In producing small-scale theories, our concern is that we frame our
CONCLUSION 233

questions explicitly and clearly and in a way that is manageable enough for
us to supply answers to our questions. The researcher must worry about
whether his or her questions are clear and worthwhile, and not about
whether they are really part of Film Theory. If a question is interesting and
its answer contributes some general knowledge about film—whether so-
ciological, philosophical, or formal—then it is a contribution to our
growing theoretical knowledge of film.5
Throughout this book I have attempted to supply examples of piecemeal
theorizing. Typically, after criticizing an account of this or that topic, as
explicated by contemporary film theorists, I have attempted to sketch what
I think is the correct account of the phenomena in question—e.g., cine-
matic representation, visual narration, and so. Part of my motive in de-
veloping alternative accounts was argumentative. That is, I not only wanted
to show that the approach in question was false but also that there was a
competing approach that can handle the facts more expeditiously. The
fielding of rival theories, of course, is a standard tool of scientific debate.
But, at the same time, my rival accounts are also specimens of piecemeal
theorizing—small-scale theories of movie narration, the structures of visual
narration, and of movie music. Suggestions are made about the way in
which these small-scale theories can be coordinated, but the result is not an
overarching theory of film because first of all my domain is restricted to
what are called movies and, of course, I do not pretend to have answered
every question, or even every important question, about movies. More-
over, the coordinated theory of movies that emerges from my debates with
contemporary film theorists is a "bottom up" theory.
My emphasis on piecemeal theorizing manifests my view that theoretical
enquiry into any area requires the framing of precise questions as a condi-
tion for meaningful progress. In order to know, we must know, to some
extent, what we want to know; one will not get very far by intendy
contemplating a body of data. You have to know what you are looking for
and the best way to get clear about what you are looking for is to develop
precise questions. Without precise questions, our theories will founder.
Moreover, at this point in film studies, the questions we are best equipped
both to propose and to pursue with any precision are piecemeal in nature.
By calling for piecemeal theorizing, I am not striking allegiance with the
emerging viewpoint that in place of the grand-scale theorizing of the
seventies and early eighties, film studies should now redirect its energies to
the interpretation of cinematic masterpieces, whose study, it is thought,
will provoke theorizing in respect to specificfilms.I do not regard the study
of individual films, in and of itself, as theoretical activity. At best it will
234 CONCLUSION

result in undirected ruminations about cinema. The film theorist of the

present must be more assertive, framing precise questions of general import
rather than awaiting "theoretical" perplexities to emerge from the examina-
tion of single films that may be idiosyncratically interesting but in theoret-
ically irrelevant ways. The totalizing theory of the seventies and early
eighties is to be rejected, but not at the cost of despairing altogether the
possibility of acquiring generalized knowledge of film.
For the last decade, film studies in America have been dominated by an
established theory, the psychoanalytic-Marxist theory described in this
book. And, moreover, this period of domination corresponds to a time in
which the academic study of film has grown dramatically. The establish-
ment theory has, as a result, become, effectively, the lingua franca of a new
academic field. It provides a common medium of discourse for an entire
generation of film scholars. The problem with this language, however, is
that it says virtually nothing. It has impeded research and reduced film
analysis to the repetition o f fashionable slogans and unexamined assump-
tions. New modes of theorizing are necessary. We must start again.
 NOTES

INTRODUCTION

1. For a discussion of classical film theory, see Noel Carroll, Philosophical Problems
of Classical Film Theory (Princeton: Princeton University Press, 1988).
2. Christian Metz, Essaissur la signfication au cinema (Paris: Klincksieck, 1968).
An English translation of this text, in its 1971 edition, was published by Oxford
University Press in 1974, under the title Film Language: A Semiotics of the Cinema.
3. Philip Rosen, Narrative, Apparatus, Ideology: A Film Theory Reader (New
York: Columbia University Press, 1986).
4. See especially Andre Bazin What Is Cinema? (Berkeley: University of Califor-
nia Press, 1971), vol. I.
5. See the second chapter of my Philosophical Problems ofClassical Film Theory, and
my "Concerning Photographic and Cinematic Representation," in Dialectics and
Humanism (1987), no. 2.
6. For further discussion of some of the problems with Metz's theory, see my
review of his Film Language in Film Comment, (Fall 1974).
7. Barthesian semiology, in such works as Mythologies and Critical Essays, ap-
pealed to the political ambitions of young film scholars; Barthes used semiology to
demvstify bourgeois ideology, and that appeared to establish a socially critical role
for the theory.

1. PSYCHOANALYSIS: M E T Z AND BAUDRY

1. See Roland Barthes, "Rhetoric of the Image," in Image-Music-Text (New

York: Hill and Wang, 1977), pp. 3 2 - 5 1 .
236 1. PSYCHOANALYSIS: M E T Z A N D BAUDRY

2. Jean-Louis Baudry, "The Apparatus," in Theresa Hak Kyung Cha, cd.. Appa-
ratus, (New York: Tan am Press, 1981), pp. 4 1 - 6 2 ; Christian Metz, "The Imagi-
nary Signifier" and "The Fiction Film and Its Spectator: A Metapsvchological
Study," in Metz's The Imaginary Signifier (Bloomington: Indiana University Press,
1982), pp. 1 - 8 7 and 9 9 - 1 4 7 respectively.
3. Jean-Louis Baudry, "Ideological Effects of the Basic Cinematographic Appa-
ratus," in Apparatus, pp. 2 5 - 4 0 .
4. Metz, "The Fiction Film and Its Spectator."
5. Hugo von Hofmannsthal, "Der Ersatz filer Traeume," in Neue Freie Presse,
March 27, 1921. This essay is analyzed in Siegfried Kracauer's Theory of Film (New
York: Oxford University Press, 1960), pp. 167-168. Another early theoretical
development of the film dream analogy can be found in Henri Diamant-Berger, Le
Cinema (Paris: Le Renaissance du livre, 1919). This text is discussed in Stuart
Liebman's doctoral dissertation: Jean Epstein's Early Film Theory (New York Uni-
versity, 1980). See especially the Chapter "Early French Film Theory." These texts
are not cited in order to establish thefirstfilm/dreamtheory, but only to show the
approach is quite old.
In The Struggle fin- the Film: Towards a socially responsible cinema (New York: St.
Martin's Press, 1986), Hans Richter analogizes film to dream in a way reminescent
of Hofmannstal, while in Running Away from Myself: A Dream Portrait of America
Drawn from the Füms of the Forties (New York: Grossman, 1969), Barabara Dem-
ming employs a dream/escape model for film.
The film/daydream analogy also has precedent: see Martha Wolfenstein and
Nathan Lettes, Movies: A Psychological Study (New York: Atheneum, 1970), pp. 12-
13.
6. Coincidentally, Stanley Cavell, who actually opposed thefilm/dreamanalogy,
also cites the invisibility/absence of the film spectator in regard to the world of the
film as a crucial element of the power of cinema. See Stanley Cavell, The World
Viewed, Enlarged Edition (Cambridge: Harvard University Press, 1979), esp. ch.
19.
7. Susanne K. Langer, Feeling and Form (New York: Scribner's, 1953),
pp. 4 1 1 - 4 1 5 .
8. Baudry, "The Apparatus," p. 57.
9. Baudry, "The Apparatus," p. 53.
10. Baudry, "The Apparatus," p. 54.
11. Baudry, "The Apparatus," p. 58.
12. Plato, The Republic, vii, 514.
13. Baudry, "The Apparatus," p. 44.
14. Baudry, "The Apparatus," p. 50.
15. Baudry, "The Apparatus," p. 50.
16. "Ordinarily" here is meant to acknowledge that there are certain tricks
which, as in Cinerama, can induce the impression of, say, plummeting. But these are
extraordinary moments of cinema, and not the sort of evidence to be adduced in an
account of the customary effects of film.
1. PSYCHOANALYSIS: METZ A N D B A U D R Y 237

17. In Metz, "The Fiction Film and its Spectator."

18. Ibid.
19. A film/dream analogy of Baudrys that I have not touched upon is the
assertion that, in film and dream, representation is taken as perception. I reject the
notion that film spectators mistake cinematic representations as perceptions and,
therefore, reject the basis of this analogy. However, since much of the following
argumentation of this book is aimed at discrediting this account of the reception of
cinematic representation, I have not found it convenient to replicate that debate at
this point. Thus, I ask the patient reader to wait for the refutation of this particular
analogy in what follows.
20. Moreover, if Baudrys analogies between film and dream are groundless,
then it is hard to sec how film could simulate dream, nor is it easy to see how film
could trigger the same sort of regression that dreams do.
21. Metz, T h e Imaginary Signifier," p. 36.
22. Metz, "The Imaginary Signifier," p. 44.
23. Metz, "The Imaginary Signifier," p. 50.
24. Metz, "The Imaginary Signifier," p. 46-47.
25. The discussion of voyeurism can be found on pp. 58—68 of Metz's "The
Imaginary Signifier."
26. The discussion of disavowal can be found on pp. 69-74 of Metz's "The
Imaginary Signifier."
27. Metz, "The Imaginary Signifier," p. 75.
28. One might also wish to challenge the viability of an essentialist approach in
art theory in general. The tradition of essentialist art theory is rejected in both my
"The Specificity of the Media in the Arts," in The Journal of Aesthetic Education
(1985), vol. 19, no. 4, and my "Medium-Specificity and the Self-Consciously
Invented Arts," in Millennium Film Journal (1984-1985), no. 14/15.
29. Metz, "The Imaginary Signifier," pp. 64-65.
30. Metz, "The Imaginary Signifier," p. 136.
31. Metz, "The Fiction Film and Its Spectator," pp. 106-107.
32. Metz, "The Fiction Film and Its Spectator," p. 107.
33. One might attempt to defer this objection by invoking Freud's "The Poet in
Relation to Daydreaming," in Philip RiefF, ed., Character and Culture (New York:
Collier Books, 1963). However, Freud's analogies there ^re more a matter of
themes than of narrative structure.
34. This hypothesis obviously bears some relation to Freud's conclusion in "The
Poet in Relation to Daydreaming," though Freud seems concerned primarily with
the content of popular stories in relation to daydreams. Freud, it seems to me, is also
susceptible to an observation that we will raise against Metz, viz., that the source of
daydreams may be popular narratives rather than the other way around.
35. See Jerome L. Singer, The Inner World ofDaydreaming (New York: Harper
and Row, 1975), especially ch. 6, "Childhood Origins of Daydreaming."
36. This is not meant to suggest that we understand only what we create, but
only that we have a particular advantage in such circumstances. Obviously, the
238 1. PSYCHOANALYSIS: M E T Z AND BAUDRY

natural sciences provide us with a great deal of knowledge about that which we have
not created. However, even here, knowledge is often acquired by envisioning the
way in which variables can be manipulated in order to produce the outcome in
question. And we test our hypotheses by constructing experiments.
37. This concluding objection is primarily directed at "The Apparatus" and "The
Imaginary Signifier" rather than "The Fiction Film and Its Spectator." For the
former essays speak of the apparatus as such and of cinematic representation as such
whereas "The Fiction Film and Its Spectator" restricts its compass to a kind of film,
namely the fictional. However, a similar line of attack could be developed in relation
to the latter essay by questioning whether fiction, irrespective of particular stories, is
a practice that warrants psychoanalyzing.
38. One motive for employing psychoanalysis in an account of movies, which
defenders of Metz and Baudry may feel I have overlooked, is that what are callcd
movies obviously exert a great deal of power over large masses of people. That is, the
intensity of the experience of vast numbers of people of movies is distinctive and
calls for explanation. The psychoanalytic projects of Metz and Baudry are war-
ranted, then, it might be argued, as a means of explaining the power of movies. This
is not an altogether implausible maneuver. However, given the methodological
constraints noted above, such a maneuver needs to show that there is no cognitivist
account of the power of movies with the same explanatory persuasiveness as the
psychoanalytic accounts. And I, at least, deny this. For an attempt at a cognitivist
account see my "The Power of Movies," in Daedalus (Fall 1985), 114 (4):79-104.
In later chapters in this book, I will further develop the view introduced in that
article. In this way, I will attempt to discredit the pretensions of psychoanalytic
theory by demonstrating the superiority of my own rival, cognitivist theory in
accounting for the power of movies.

2. M A R X I S M AND P S Y C H O A N A L Y S I S ; T H E
A L T H U S S E R I A N - L A C A N I A N PARADIGM

1. Herbert Marcuse, An Essay on Liberation (Harmondsworth: Penguin, 1969),

p. 24.
2. T. W. Adorno, Introduction to the Sociology of Music {Hew York: Seabury Press,
1976), p. 31.
3. See Vance Packard, The Hidden Persuaders (New York: McKay Books, 1957).
Packard's influence on the post-World War II generation of Americans, who
comprise the stateside ranks of contemporary film theory, may not only account for
their predisposition to mind-manipulating hypotheses, but also for their frequent
use of advertisements as examples of the ideological operation of representation in
general. Contemporary theorists love to decode advertisements. Indeed, perhaps it
was this obsession with advertisements that enabled Americans to appreciate Bar-
thes as a kindred spirit. O f course, this is just armchair sociology. From the point of
2. MARXISM A N D PSYCHOANALYSIS 239

view of theory, however, one thing needs to be said about the use of advertisements
as privileged examples of how representation operates. Advertisements are repre-
sentations that are expressly made to cultivate capitalist desires. That they succeed in
doing so comes as no surprise. On the other hand, that such representations succeed
in doing so implies nothing about representations not expressly designed to induce
capitalist desires. That is, the fact that advertisements are ideologically inflected—
given that they are made for just such purposes—indicates nothing about whether
or not other sorts of representations, or whether representation as such, are ideolog-
ical instruments.
4. The word "settled" in the above sentence is key. I do not wish to deny that at
various times Marxists have attempted to embrace psychological frameworks other
than the Lacanian one. Pavlovian classical conditioning was at one time thought of
as an appropriately materialist psychology. Nor is Lacanian psychoanalysis the only
sort that people have attempted to link with Marxism. Benjamin, Adorno, and
Marcuse all, in different ways, attempted to adapt Freudian psychoanalysis for
Marxism, while Reich tried to fuse socialism and his own brand of psychoanalysis.
Recendy, Bertell Oilman, in his book Alienation (New York: Cambridge Univer-
sity Press, 1971), criticized Marxism for a lack of psychology. However, in his Social
and Sexual Revolution (Boston: South End Press, 1979), he does not call upon
Lacan to fill this gap but rather Reich. The point here is that there is no settled
Marxist psychology, that is, one that is endorsed in the classical sources of Marxism
and that Marxists agree is correct.
Mention of alternative psychological frameworks to Lacanianism, of course,
prompts the question of why film theorists chose it. Two considerations seem
relevant here. First, as noted in the previous chapter, the notion of the mirror figures
importantly in Lacanian psychoanalysis. Thus, the theoretical language of Lacanian
psychoanalysis contained a potential pun which allowed film theorists to appropri-
ate it by speaking of film as a "mirror." Second, from the earliest stages of the film
renaissance of the sixties, English-speaking film devotees were dependent upon the
French for their most ambitious theories—the auteur theory and the Bazinian
theory of realism. Thus, when the English-speaking film intelligentsia turned to
psychoanalysis, it is not surprising that the favored variety would be French, since it
was from the French that the earlier stages of sixties' film culture had derived its
inspiration. Moreover, many contemporary film theorists were trained in compara-
tive literature departments and French departments, which made it natural for them
to seek their psychoanalytic models in Paris. These observations, of course, are not
made in an attempt to rationally justify contemporary film theory's preference for
Lacanian psychoanalysis overrivalpsychoanalytic positions, but only to explain that
choice.
5. Louis Althusser, "Marxism and Humanism," in his For Marx (New York:
Vintage, 1970), pp. 219-247; "Freud and Lacan," in his Lenin and Philosophy and
Other Essays (New York: Monthly Review Press, 1971), pp. 189-220; and "Ideol-
ogy and Ideological State Apparatuses (Notes towards an Investigation)" in Lenin
and Philosophy, pp. 127-188.
240 2. M A R X I S M A N D P S Y C H O A N A L Y S I S

6. Including Althusser's "The 'Piccolo Teatro': Bertolazzi and Brecht," in For

Marx, pp. 129-153; his "A Letter on Art in Reply to Andre Daspre," in Latin and
Philosophy, pp. 221 —229; and his "Cremonini, Painter of the Abstract," in Lenin and
Philosophy, pp. 229-242.
7. For a defense of Althusser's fiinctionalism see G. McLennon et al., "Al-
thusser's Theory of Ideology," in On Ideology (London: Hutchinson, 1978),
pp. 77-106.
8. Althusser, "Ideology and Ideological State Apparatuses," p. 143.
9. Althusser, "Marxism and Humanism," p. 232.
10. Althusser, "Ideology and Ideological State Apparatuses," p. 162.
11. Althusser, "Ideology and Ideological State Apparatuses," p. 182.
12. It must be noted that some contemporary film theorists regard Althusser's
account here in terms of interpellation as too simplistic. They do not believe that
interpellation is the only element in subject construction. For, it is held, the subject
is not simply positioned by the ideological address, but there is also interaction
between the subject (subjectivity) and discourse. For example, it will be argued that
the subject sutures or fills in the discourse. See, for example, Stephen Heath,
Questions of Cinema (Bloomington: Indiana University Press, 1981), for example,
p. 103. This complication will be taken up in later sections of this book.
13. This is not stricdy accurate. For Althusser, ideology always addresses sub-
jects who have already been affected by previous addresses.
14. The hedge, virtually, here is meant to acknowledge that some Althusserians
might exempt scientific discourse from the realm of ideology. In this, such epigones
would appear to be following the master.
15. Sec Kaja Silverman, The Subject of Semiotics (New York: Oxford University
Press, 1983), p. 50.
16. See, for example, Annette Kuhn, Women's Pictures: Feminism and Cinema
(London: Routledge and Kegan Paul, 1982), p. 47.
17. Althusser, "Freud and Lacan," p. 219.
18. My colleague Christina Crosby has pointed out to me that the Lacanian view
of humans as ever striving for a lost plenitude after having initially fallen from it
sounds suspiciously theological; nor does the explanation of all human behavior in
light of such a master myth seem particularly informative or rich.
19. See Jacques Lacan, "The mirror stage as formative function of the I as
revealed in psychoanalytic experience," in Lacan's Ecrits (New York: Norton,
1977), pp. 1-8.
20. Jacques Lacan, "The mirror stage," p. 2.
21. P. Q. Hirst has questioned the notion advanced by Althusser on the shoul-
ders of Lacan that subjects are constituted in the mirror stage. Hirst's point is logical
and unavoidable. For the infant to recognize itself as a subject in the mirror
presupposes a subject already possessed of the capacities supposedly acquired at that
stage. What does the recognizing, in other words, if not a subject? See P. Q. Hirst,
"Althusser's Theory of Ideology," in Economy and Society (Nov. 1976), 5(4):404ff.
22. Similarly, what can count as a disunity in contemporary film theory is rather
2. MARXISM A N D PSYCHOANALYSIS 241

expansive, for example: uncoordinated motor functions and contradictions, pre-

sumably of both an ideological and a logical nature. But surely these items are too
disparate to be regarded as on a par with each other.
23. Some might argue that the difference cited above suggests that the tide of
this chapter is a misnomer—that it is wrong to call the ruling paradigm "Al-
thusserian." I disagree since despite expansions made on Althusser's "Ideology and
Ideological State Apparatuses," research in contemporary film theory still seems
ultimately based on the Althusserian assertion that ideology is at root subject
construction and that ideology under capitalism is a matter of producing unified,
autonomous subject positions, often misleadingly called Cartesian. For different
reasons, contemporaryfilmtheorist Colin MacCabe would agree with me about the
sustaining influence of Althusscr on recent research. He writes: Throughout the
seventies much work of cultural analysis on the left and particularly within Screen
took place within a space articulated by the work of French Marxist, Louis Al-
thusser." In MacCabe's "Class of'68," in his Tracking the Signifier, Theoretical Essays:
Film, Linguistic, Literature (Minneapolis: University of Minnesota Press, 1985),
p. 13.
24. It is no accident that my example here is of a male child. Freudian and
Lacanian theories both have much neater accounts of the way in which male
children are acculturated than they do of female children. This has long been
recognized by feminists. Those feminists who accept the overall contours of the
Althusserian-Lacanian paradigm, or even merely just psychoanalysis, spend a great
deal of theoretical energy attempting to adjust the approach to accommodate the
female experience into the account. However, this may be tantamount to adding
epicycles to the Ptolemaic theory of the heavens; that is, perhaps feminists would be
better off dropping the paradigm altogether.
25. This belief may, however, be debated. For a brief introduction to some of the
disputes see David Archard, Consciousness and the Unconscious (LaSallc, 111.: Open
Court Publishers, 1986), especially pp. 73 and 83.
26. Jacques Lacan, "The function and field of speech and language in psycho-
analysis," in Ecrits, p. 66. This is sometimes also referred to as "The Rome Dis-
course."
27. Jacques Lacan, "The function and field of speech," p. 67.
28. David Archard has shown to my satisfaction that Lacan's theory of language,
as he develops it from Saussure, is completely ill-founded. See his Consciousness and
the Unconscious, pp. 56-103. Also, for my own objections to what has come to be
called post-Saussurean linguistics in contemporary critical theory, see my "Belsey,
Language and Realism," in Philosophy and Literature (April 1987).
29. Lacanians would object here. They would argue that their theory of meaning
shows there is no fixity of meaning. In this I think that their method of linguistic
theorizing has got things backwards. Linguistic theorizing is supposed to explain
how things like language learning can happen. If you come up with a theory, like
theirs, that cannot account for this because of its denial of any fixity of meaning,
then you should drop your theory. The problem with the use of Saussure in much
242 2. MARXISM AND PSYCHOANALYSIS

contemporary theory is that it proceeds in too metaphysical a manner, deducing

how the world must be from basic premises, come what may. But linguistics is a
science that starts of with some phenomena that wants explaining.
Many contemporary theorists think that fixity of meaning must be denied be-
cause of the supposed arbitrariness of the sign. If meanings exist diacritically, thev
will always be sliding, it is said. However, this is not a very compelling argument.
Consider an analogy. The positions in a baseball game,—pitcher, catcher, short
stop, left fielder, and so on—are relationally defined. But there the relations arc
fixed, fixed by the rules of baseball. You may say the rules of baseball are arbitrary.
But once the rules are in place, there is nothing arbitrary about who is the third
baseman. So also with language.
These comments are probably too brief to convince the proponent of contempo-
rary theory. For more detailed and sustained objections to the use of Saussure in
contemporary theory consult the items in the preceding footnote.
30. Jacques Lacan, 'The function and field of language and speech . . . p . 68.
31. Louis Althusser, "Freud and Lacan," p. 211.
32. Lacan seems to see the Imaginary as developmentally leading to the Sym-
bolic on the basis of such data as the fort/da game Freud observed his grandson
playing (which is noted in Beyond the Pleasure Principle, Standard Edition, p. 18).
That game, played with a ball of string and with a mirror, involves themes both of
difference and of presence and absence. See Lacan, The function and field of
speech," pp. 103-107. One may of course wonder about the advisability of basing
so much theory on this single report.
33. Colin MacCabe, "Theory and Film: Principles of Realism and Pleasure," in
Tracking the Significr, p. 65. We should also note that along with the realm of the
Imaginary and the realm of the Symbolic, Lacan upholds another. This is the realm
of the Real. Often it correlates with themes of death and sexuality and sometimes it
sounds as though it might pertain to what lies outside symbolization. Since it is not
as explicitly crucial to contemporary film theory as the Imaginary and the Symbolic,
we shall, given our purposes, pass over it.
34. Louis Althusser, "Freud and Lacan," p. 216.
35. Kaja Silverman, The Subject of Semiotics, p. 50.
36. Above, I have defined what it is for a belief or an assertion to be ideological.
Some might want to maintain that entire symbol systems—like cinema or lan-
guage—can also be ideological. But such systems would not be susceptible to the
preceding definition for thev are neither true nor false. So what would be the criteria
that identify ideological symbol systems? From my perspective, I would claim that if
an entire symbol system could be characterized as ideological, it would be because it
(1) excludes or represses the representation of certain social facts or relations, bv
virtue of its structure, and (2) is used to uphold some system of social domination
by means of the exclusions it enjoins. I should add that I am not convinced that
either cinema as such or a natural language like English is inherently ideological.
37. In "On the Theory of Ideology: The Politics of Althusser," New German
Critique (Spring 1976), no. 8, pp. 5 4 - 7 9 , J. Ranciere charges that Althusser's view
2. MARXISM A N D PSYCHOANALYSIS 243

of ideology ignores class struggle. The objections in this paragraph follow in that
line of criticism.
38. Louis Althusser, "The 'Piccolo Teatro,'" p. 149.
39. Louis Althusser, "A Letter on Art.'"
40. Louis Althusser, "Cremonini, Painter of the Abstract," pp. 238-239.
41. It may be argued that I have scored some cheap shots in the last two
paragraphs by relying upon a distinction Althusser makes between ideology and
science that few contemporary film theorists accept. However, the way in which this
distinction is usually negotiated by contemporary theorists is to say that there is no
such distinction. And that will only make my objections, except for the last one,
stronger, not weaker.
42. The arguments in the preceding paragraphs correspond to certain points
made by Martin Hollis and Charles Taylor. See Hollis' "Of Masks and Men" and
Taylor's "The Person," both in Michael Carrithers, Steven Collins, and Steven
Lukes, eds., The Category of the Person (Cambridge: Cambridge University Press,
1985), pp. 217-233, and 2 5 7 - 2 8 1 respectively.
43. David Hume, A Treatise of Human Nature, book 1, part 4, sections 1,5, and
especially 6; and Friedrich Nietzsche, Beyond Good and Evil, especially chap. 1,
sections 12 and 16—23.
44. Frederick Engels, Herr Eugen Duhring's Revolution in Science (London),
p. 128.
45. See Immanuel Kant, Critique of Pure Reason, especially "Criticism of the
Third Paralogism of Transcendental Psychology." There Kant points out that the
unity of the self as the transcendental unity of apperception neither entails nor is
equivalent to the identity of a person. Psychoanalysis does not confront the issue of
the transcendental self at the proper level of generality. For an example, though
probably an unsuccessful one, of a direct confrontation with the Kantian variety of
self-identity see Jean-Paul Sartre, The Transcendence of the Ego (New York: Farrar,
Straus and Giroux, 1957).
46. It should be clear that what capitalism needs are subjects, who, for whatever
reason, carry out certain kinds of behaviors. And this type of subject, call it an
economic subject if you want, is compatible with a wide variety of metaphysical
characterizations. Another way to put the point is to say that the abstract subject of
Althusser's apparatuses is not determinate enough to serve as what we just, some-
what loosely, called a capitalist subject.
47. For example, a follower of late Wittgenstein is likely to endorse these as
fundamentals for social science.
48. Karl Marx, Capital, Frederick Engels, ed. (New York: International Pub-
lishers, 1967), 1:737 (emphasis added).
49. A great deal of empirical data concerning working-class beliefs is usefully
summarized in The Dominant Ideology Thesis, by Nicholas Abercrombie, Stephen
Hill, and Bryan S. Turner (London: George Allen & Unwin, 1980), pp. 140-155.
This book defends the notion that what I have called the Economic Force Hypoth-
esis is superior to the Ideology Hypothesis.
244 2. MARXISM AND PSYCHOANALYSIS

50. Ted Benton, The Rise and Fall of Structural Marxism (New York: St. Martin's
Press, 1984), p. 44.
51. Another major problem with the Ideology Hypothesis is that its proponents
constantly speak as though there is a capitalist ideology. But the vexing problem here
is that no one seems to be able to characterize it or its ingredients. This particular
problem is examined at length in the Dominant Ideology Thesis.
52. Perry Anderson has argued that Western Marxists have become embroiled in
questions of method and aesthetics as a result of their impotence in the face of a
stagnant working class; in Considerations on Western Marxism (London: New Left
Books, 1976). Perhaps the elevation of ideology—a form of cultural work—is a
further symptom of that impotence, an attempt to make intellectual work a central
force of capitalism. And if the ideologues are the greatest threat to the revolution,
doesn't that make the antidote—cultural analysis—our greatest hope? One cannot
help feeling that intellectual delusions of grandeur provide emotional support to the
Ideology Hypothesis.
53. A similar point has been noted by contemporary film theorist Philip Rosen in
his " 'Screen' and the Marxist Project in Film Criticism," in Quarterly Review of Film
Studies (August 1977), p. 280.
54. See Thomas L. Friedman, "J.R.'s Message? As Varied as Kibbutz and Ba-
zaar," The New York Times, April 1, 1986.

3. T H E CINEMATIC IMAGE

1. Martin Walsh, The Brechtian Aspect of Radical Cinema (London: British Film
Institute, 1981), p. 12.
2. Bertoit Brecht, "A Short Organum for the Theater," from John Willet, ed.,
Brecht on Theater (New York: Hill and Wang, 1964), section 72, p. 203 (emphasis
added).
3. Brecht, "From the Mother Courage Model," Brecht on Theater, p. 219 (em-
phasis added). One, of course, wonders if Brecht can consistently reject naturalism
tout court given the pragmatic instrumentalism he adopts toward symbol systems in
"Against Georg Lukacs," New Left Review (March-April 1974), no. 84.1 think that
one way to assuage this tension might be for exegetes of Brecht to sav the "illusion-
ism" quotations I stress in this section are off-hand or careless remarks of Brecht that
do not really reflect his considered judgment on these matters. This may be true.
Nevertheless, I still feel it is important to examine the claims of the interpretation of
Brecht that I characterize above because something like that interpretation has been
shared by numerous cine-Brechtians as well as by certain proponents of avant-garde
theater. That is, I am concerned with the way in which Brecht has been used rather
than in a possible, rational reconstruction of his system. On the other hand, for
those who want to claim that I have radically misinterpreted Brecht, it is important
to recall that he really did write what I have quoted.
3. T H E C I N E M A T I C IMAGE 245

One argument in favor of the illusion theory of spectator response—i.e., the

theory that says spectators believe they are somehow witnessing "the real thing" is
to ask a psychological question: "Why would audiences cry and scream unless they
believed the events before them were actually occurring?" I do not examine this type
of argument at length above because it is not one that cine-Brechtians rely on.
However, for the record, let me say that I would begin to answer it by denying the
psychological assumption that we are moved only by events that we believe actual.
We can be moved by the prospect of an event, that is, by thinking that an event
might be possible. For elaboration, see my T h e Nature of Horror," Journal of
Aesthetic and Art Criticism (Fall 1987).
4. Plato attacks mimesis on epistemic grounds in Republic, book 10. It is this
attack that Hegel questions in the opening of his "Introduction" to the Philosophy of
Fine Art. It is also interesting to note, from the perspective of contemporary film
criticism, that Plato not only condemned mimesis on epistemic grounds but also
condemned mimetic drama for inclining spectators toward incorrect behavior (cf.
Republic, book 10, 606d). See also books 2 and 3 of Plato's Republic. And see
Immanuel Kant, "Analytic of the Sublime," in Critique ofJudgment, section 42.
5. Brecht on Theater, p. 215.
6. Ibid.
7. For an account of illusionism, as that is conceived from a Brechtian viewpoint,
by a contemporary film theorist, see the late Martin Walsh's The Brechtian Aspect of
Radical Cinema, pp. 11—14. This interpretation is especially significant in that it
explicitly applies the Brechtian notion of illusionism to cinema. Walsh's account
also corresponds, I believe, to my own.
8. Rosalind Coward and John Ellis, Language and Materialism (London: Rout-
Icdge and Kegan Paul, 1977), pp. 46-47.
9. See Walsh, p. 12, and Stephen Heath, "Le Pere Noel," October (Fall 1983),
no. 26.
10. In his Meditations, Descartes claims that we will our beliefs, a view roundly
challenged by Spinoza in his Ethics.
11. E. H. Gombrich, Art and Illusion (Princeton: Princeton University Press,
1960), p. 29.
12. Dudley Andrew, Concepts in Film Theory (London: Oxford University Press,
1984), p. 42.
13. There is a remaining Brechtian illusion which has not yet been discussed,
namely the audience's purported identification with characters. Since this illusion
has not be developed at length by contemporary film theorists, it seems appropriate
to treat it in a footnote.
Illusionism, according to Martin Walsh, involves "a desire to (psychologically)
penetrate individual experience; its primary appeal is to the emotions rather than
the intellect, desiring the audience's empathetic involvement with the events pre-
sented before them" (Walsh, p. 11). And, as we saw in a previous quotation from
Brccht, this empathetic union with fictional characters distracts us from noting the
ideological contradictions in the representations. Furthermore, this sort of illusion,
246 3. THE CINEMATIC IMAGE

which we can call character identification, can occur with any sort of mimetic-
representation that employs characters, including novels, lyric poems, historical
paintings, theater, and obviously film.
Now there is no question that mimetic representations can arouse the emotions
of spectators, and manipulate those emotions in ideologically significant ways. But
the theoretically relevant question is whether identification is the mechanism
through which our emotions are aroused and directed. Undoubtedly, people—
includingfilmtheorists, critics, and nonspecialist viewers—are always talking about
identifying with characters. But is this way of speaking either accurate or useful?
One problem with evaluating the notion of character identification is that it is not
clear what it is meant to refer to. When a spectator identifies with a character does
she somehow come to believe that she is the character? Do I think that I am James
Bond when I watch From Russia With Love or that I am Rosenthal during Grand
Illusion? If normal viewers identified literally with the characters onscreen, they
would soon become unhinged. They would run from the theater screaming as they
and their fused screen-selves were about to be crushed by a train, pushed off a
building, knifed, or shot at by a German border patrol.
Also, the concept of identification does not appear to have the right logical
structure for analyzing most of our emotional responses to characters. That is,
"identification" suggests that when we bond with characters we somehow become
them, which would entail that we have the same emotions that they do. But this is
very often not the case. Whatever parting emotion King Kong feels for Anne
Darrow, it is not ours; rather, we feel pity for this great beast brought low. We do
not fear for our life during the kangaroo court scene in M, rather we feel the
injustice of the trial. When lovers are united at the end of melodramas, we do not fall
in love; rather, we are glad that people so suited have found each other.
Our emotional responses are certainly related to the emotions of characters. But
they need not and quite often are not identical to the emotions of characters. That is,
a strong sense of identification would suggest that the emotions the characters
evince are the same as those the spectators possess. But this most often is not the
case. Rather, the action, emotions, and plights of the characters generally bring
about qualitatively different emotions in spectators, ones appropriate to spectator-
ship, e.g.: we don't feel humiliated when Harold gives his speech in The Freshman,
we feel amused; when the Tramp wins Georgia, we are gladdened for him, not
smitten with her; and when Mad Max battles Blaster in the Thunderdrome, we feel
suspense, an emotion Max can't afford to indulge if he intends to stay alive.
Stated formally, a strong sense of character identification would imply a sym-
metrical relation of identity between the emotions of spectators and characters. But
generally, the relation is asymmetrical; the characters, in part through their emo-
tions, cause different emotions in spectators. This logical asymmetry indicates that
identification, a symmetrical relation, is not the correct model for describing the
emotional responses of spectators. (This line of objection would also carry against a
weaker sense of identification, one, for example, where identification is defined as an
emotional tie that causes a spectator, consciously or unconsciously, to think, or feel
3. T H E CINEMATIC IMAGE 247

or act as she imagines the character with whom she has the tic does. For, as we have
seen, the spectator in a large number of cases does not feel as one imagines the
character does.)
I would not wish to deny that there are cases where the audience and a character
may share parallel feelings. In horrorfilms,for example, we tend to be reviled by the
monster in the same way that the character is. However, cases of matching emotion
like these are only one of the ways in which a spectator can relate to characters. So,
even if there is some place for some as yet to be described empathetic relation
between audience and character, identification cannot be the whole story of the
audience/spectator connection, nor is it clear that it is even most of the story.
Often in American films, audience concern for certain characters, like cowboy
heroes and misunderstood gangsters, is encouraged by emphasizing that character's
courteous, caring, respectful, or thoughtful treatment of supporting characters,
especially ones who are poor, old, weak, lame, oppressed, children, etc.—that is,
characters who are the protagonists' inferiors in some sense, but whom the protago-
nist treats with consideration (conversely, villains are often marked by their mis-
treatment of such characters). Thus, by presenting characters who display values or
interpersonal ideals that the audience acknowledges, shares, or feels allegiance
toward, films nurture a pro-attitude in the audience on behalf of the spectator.
But feeling allegiance or having a pro-attitude is not the same as identification. I
can feel allegiance to South African blacks and Irish Catholics in Belfast without
imagining myself to be one of them. Nor, by the way, returning to the issue of
illusion, does a response to shared values seem to me to be the sort of thing that is
usefully described as a matter of deception. In sum, though I do not deny that,
broadly speaking, something like empathy may be relevant to some of our emo-
tional responses to characters, I suspect that allegiance to shared values may be a
more critical lever for shaping our emotional responses in film. For some further
elaboration see my "Toward a Theory of Film Suspense," in Persistence of Vision
(1984), no. 1.
14. Quoted in Aaron Scharf, Art and Photography (Baltimore: Penguin Books,
1975), p. 149.
15. Benedetto Croce, Aesthetic (Boston: Nonpareil Books, 1978).
16. Clive Bell, Art (New York: Capricorn Books, 1958), p. 23.
17. Bill Nichols, Ideology and the Image (Bloomington: Indiana University Press,
1981), and John Ellis Visible Fictions (London: Routledge and Kegan Paul, 1982).
18. John Ellis, Visible Fictions, p. 4 1 ^ 2 .
19. Emile Benveniste, Problems in General Linguistics (Coral Gables, Fla.: Uni-
versity of Miami Press, 1971), pp. 217-222.
20. Also, the historic mode of representation omits use of the second person
pronoun as well as use of the first person. This is taken to imply that the realist
representation does not acknowledge the presence and the productive activity (e.g.,
interpretive activity) of the reader. This purportedly reinforces the impression of
"self-containedness" of the representation which reinforces the impression that it
"comes from nowhere."
248 3. T H E C I N E M A T I C IMAGE

21. Metz, "Story/Discourse (A Note on Two Kinds of Voyeurism)," in The

Imaginary Signifier (Bloomington: Indiana University Press, 1982), pp. 91-98.
22. For further criticisms of the notions of illusion, naturalness, and trans-
parency as discussed in the preceding two sections of this chapter see my "Conspir-
acy Theories of Representation," in Philosophy of the Social Sciences ( 1987), vol. 17.
23. Jean-Louis Comolli, "Technique and Ideology: Camera, Perspective, Depth
of Field," in Film Reader (Northwestern University) (January 1977), no. 2,
pp. 131-132.
24. Stephen Heath, Questions of Cinema (Bloomington: Indiana University
Press, 1981), p. 28.
25. Heath Questions of Cinema, p. 53.
26. Bill Nichols, Ideology and the Image, p. 159.
27. David Bordwell, Narration in the Fiction Film (Madison: University of
Wisconsin Press, 1985), p. 106. Bordwell's discussion of perspective is very clear-
headed, perhaps the most useful summary of the perspective issue in recent writing
on film theory.
28. Jan B. Deregowski, "Illusion and Culture," in R. L. Gregory and E. H.
Gombrich, eds., Illusion, Nature and Art (New York: Scribner's, 1973), p. 183.
29. Ichitaro Hondo, "History of Japanese Painting," in Painting, 14-19 Cen-
turies: Pageant ofJapanese A rt (Tokyo: Tokyo National Museum, 1957), 2:54—55.
30. Bordwell, p. 107.
31. Stephen Heath, "Le Père Noël," October (Fall 1983), p. 88.
32. Stephen Heath, "Le Père Noël," pp. 85-87. For a sustained rebuttal of the
charges Heath lays at my door step, see Noel Carroll, "A Reply to Heath," October
(Winter 1983), pp. 81-102.
33. In denying that perspective is appropriately described as a convention, I am
not saying that perspective pictures lack conventions altogether. I am only holding
that conveying a sense of spatial depth by means of perspective is not an arbitrary
convention. For a convention is something adopted in the context where there are
alternative ways of achieving the same effect and it is a matter of indifference as to
which of these alternatives is adopted, such as driving on the left or right hand side
of the road. But if perspective is more accurate spatially, then it is not the case that it
is one among numerous, indifferent alternatives for depicting the appearance of
spatial layouts.
34. Bordwell, p. 107.
35. Similar objections to Heath can be found in Bordwell, Narration in the
Fiction Film, p. 25.
36. J. E. Hochberg and V. Brooks, "Pictorial Recognition as an Unlearned
Ability," American Journal of Psychology, (1962), no. 75, pp. 624-628.
37. J. B. Deregowski, E. S. Muldrow, and W. F. Muldrow, "Pictorial Recogni-
tion in a Remote Ethopian Population, Perception (1972), no. 1, pp. 417—425.
38. John M. Kennedy, A Psychology of Picture Perception (San Francisco: Jossey-
Bass, 1974), p. 79.
39. Margaret A. Hagen, Varieties of Realism: Geometries of Representational Art
4. NARRATION 249

(Cambridge: Cambridge University Press, 1986), pp. 86-87.1 quote Hagen not to
subscribe to her overall position but for her evaluation of the literature of cross-
cultural perception.
40. K. J. Hayes and C. Hayes, "Picture Perception in a Home-Raised Chim-
panzee," Journal of Comparative and Physiological Psychology (1953), no. 46,
pp. 470-474.
41. John Searle, Minds, Brains and Science (Cambridge: Harvard University
Press, 1984), pp. 51-52.
42. I don't think that the preceding paragraph demonstrates its case conclusively,
but only that it renders its hypothesis plausible, something that I take many
contemporary film theorists would deny.
43. Where a cinematic image conveys a message statable in propositional form, it
will be ideological roughly in accordance with the definition given of ideological
beliefs in the previous chapter. However, it is also the case that many would be
loathe to apply the predicate "false" to an image. Where an image does not have a
propositionally formulatable message, nevertheless, some might still wish to desig-
nate it to be ideological. Perhaps, in such cases, we shall want to say that a cinematic
image is ideological just in case the image and/or its contextualization in a con-
catenation of images obscures or otherwise fails to represent certain social relations
or facts accurately, and that inaccuracy plays a role in some system of social
domination.

4. NARRATION

1. John Ellis, Visible Fictions (London: Routledge and Kegan Paul, 1981), p 8.
2. John EUis, Visible Fictions, p. 7 (parenthesis added).
3. Ellis, p. 9.
4. For an account of the coherence of Anger's Inauguration of the Pleasure Dome,
see Noel Carroll, "Identity and Difference: From Ritual Symbolism to Condensa-
tion in Inauguration of the Pleasure Dome," in Millennium Film Journal (Spring
1980), no. 6, pp. 31-42.
5. I am following Kaja Silverman's translation of Benveniste's terms here. See her
Subject of Semiotics (New York: Oxford University Press, 1983), p. 196.
6. Emile Benveniste, Problems in General Linguistics (Coral Gables, Fla.: Univer-
sity of Miami Press, 1971), p. 218.
7. Quoted by David Bordwell, in Narration in the Fiction Film (Madison: Uni-
versity of Wisconsin Press, 1985), p. 341.
8. Coward and Ellis write "The narration does not appear to be the voice of an
author; its source appears to be a true reality which speaks." In Rosalind Coward
and John Ellis, Language and Materialism (London: Routledge and Kegan Paul,
1977), p. 49.
9. Catherine Bclsev, Critical Practice (London: Methuen, 1980), p. 72, and
250 4. NARRATION

Jenny Simonin-Grumbach, "Pour une typologie des discours," in Julia Kristeva,

Jean-Claude Milner, and Nicolas Ruwet, eds., Langue, discours, société (Paris: Seuil,
1975), p. 103.
10. David Bordwell, Narration in the Fiction Film, p. 21. In the section of this
book entided "Film Narration as Enunciation," pp. 21 - 2 6 , Bordwell is particularly
good at showing the confusion, often unnoticed, of contemporary theorists in their
attempts to loosely extrapolate Benveniste's linguistic categories to texts and to
films. Bordwell shows that what has transpired is pretty much a fanciful free-for-all
with different theorists making of the original categories whatever thev will.
11. See Raymond Bellour, "Hitchcock the Enunciator," in Camera Obscura (Fall
1977), no. 2."
12. Metz, "Story/Discourse (A Note on Two Kinds of Voyeurism," in The
Imaginary Signifier (Bloomington: Indiana University Press, 1982), pp. 91-97.
13. David Bordwell, Narration in the Fiction Film, pp. 2 1 - 2 6 .
14. There is a particularly byzantine and probably incoherent formulation of the
significance of enunciation in Stephen Heath's Questions of Cinema (Bloomington:
Indiana University Press, 1981 ). Since I have dealt with this specific variant at some
length in my "Address to the Heathen," October, (Winter 1982) no. 23, and my "A
Reply to Heath," October (Winter 1983), no. 27, I shall not dwell here on the
precise details of Heath's characterization of enunciation. Interested readers should
refer to Heath's book and our ensuing debate in October for more intricate dialectics
in the matter of enunciation.
15. Metz, "Story/Discourse," p. 96.
16. One reason to drop the personification of the film is the sheer silliness that
the metaphor induces. Metz writes: "The film is not an exhibitionist. I watch it, but
it doesn't watch me watching it. Nevertheless, it knows that I am watching it. But it
doesn't want to know" (Imaginary Signifier, p. 96). He then goes on to attribute to
the film a psychic act of disavowal. But this is word play, not theory. Films don't
disavow anything in the psychic sense Metz implies, and I don't think we learn
anything through the elaboration of this utterly fanciful conceit.
17. Mark Nash attempts to argue the case for the second person in cinema in his
"Vampyr and the Fantastic," Screen (Autumn 1976), 17(3):29-67. Bordwell shows
the error of his ways in Narration in the Fiction Film, pp. 23-24.
18. Also the notion that viewers do not regard films as constructions seems to be
inconsistent with Metz's account of film fetishism. Purportedly, we are fetishists
when we think that a film is well made; but doesn't this mean that when we arc
behaving as film fetishists, we believe that the film is constructed; indeed, we believe
that it is well constructed. This observation is not offered in order to accept Metz's
view of film fetishism, but only to indicate that it sits oddly with the theorv
discussed above.
19. Interestingly, the relativistic claims concerning interpretation that are widely
popular throughout American culture converge on the type of interpretive freedom
espoused by many literary critics and their followers in film. Perhaps those critics.
4. N A R R A T I O N 251

rather than fomenting a revolution, are reflecting the tendency of the culture at large
to transform endemic principles of tolerance and pluralism into an epistemological
creed.
20. See my "Address to the Heathen" and UA Reply to Heath" for further
discussion of this variant of the enunciation effect.
21. It is ironic that while contemporary film theorists often derive their impetus
from anti-formalism, they, nevertheless, appear to offer very generic accounts of
representation which are not very specific historically and which seem to speak of
invariant structures of realist representation, such as the effacement of enunciation.
Such analyses could be called a species of formalist functionalism.
22. See my "Toward a Theory of Film Suspense" Persistence of Vision (1984), no.
1, for an amplification of the role of virtues in popular entertainment.
23. Stephen Heath, Questions of Cinema, p. 54.
24. Heath, Questions of Cinema, p. 136. Heath calls this a simple definition. He
seems to think that it is not complex because it does not address the purported fact
that the narrative cannot suture every element in the film. None of the ensuing
arguments against Heath's theory of movie narrative, however, is based on exploit-
ing the lack of this complication in the equilibrium definition of film (movie)
narrative.
25. Heath, Questions of Cinema, p. 133.
26. Heath, Questions of Cinema, p. 157.
27. Heath, Questions of Cinema, p. 36.
28. For example, Stephen Neale in his book Genre (London: British Film In-
stitute, 1980), tries to apply systematically Heath's idea that genres are "instances of
equilibrium." In the attempt to identify equilibriums and disequilibriums for every
film genre, these concepts are bent completely out of shape. Equilibrium can mean
order, law, harmony, and so on, while disequilibrium takes into account dramatic
conflict, criminality, and discord. Neale does not confront the issue of genre films
like The Red Shoes and Von Ryan's Express which do not end in equilibriums as he has
characterized them. Are unhappy or tragic endings equilibriums even though they
would seem to be discordances in the system of homologies Neale outlines? Neale
might say that these endings are equilibriums because they reinforce some ideologi-
cal tenet. But if we begin talking about disequilibrium-equilibriums, the system
seems very ad hoc and equilibrium is emptied of explanatory power.
Another quite different problem with Neale's book is that he proceeds under the
presumption that every film genre must correlate to some sort of psychic phenome-
non or syndrome. He spends his procrustean energies matching each genre with
exhibitionism, paranoia, and so on. But all his word play is to little avail since he has
never provided any reason for us to believe that every genre has a psychoanalytic
correlate; nor does such a presupposition appear probable.
29. Once the film scholar takes possession of a theoretical term like violence, she
can begin reading the theory into the story of the film—finding it remarkable, for
example, that the initiating change in such and such a movie literally is a violent act.
252 4. NARRATION

Thus, a description becomes parlayed into something that sounds like an analysis.
But in fact such exercises are merely specious specimens of allegorizing—i.e., laving
the scenario of the theory on the narrative by means of a pun.
30. Heath himself tends to identify the balanced repetition of a narrative with
narrative coherence as such; see Questions of Cinema, p. 157.
31. For example, see my "The Moral Ecology of Melodrama: The Family Plot in
Magnificent ObsessionNew York Literary Forum (a CUNY publication), the Melo-
drama Issue, 1980.
32. The contemporary film theorist is prone to describe everything—film de-
vices, whole films, and spectator-subjects—with the same words (position, center,
and so on). This enables the theorist to slide rhetorically from the description of a
device to an effect on a subject; a centered narrative centers a subject. One can only
guess at the internal structure of these cause and effect relations. Is it contagion? Or
is the appearance of a causal explanation advanced merely by means of equivoca-
tion?
33. David Bordwell comes to roughly the same conclusion in his Narration in the
Fiction Film, p. 25.
34. V. I. Pudovkin, Film Techniques and Film Acting (New York: Grove Press,
1960).
35. Lev Kuleshov, Kuleshov on Film (Berkeley: University of California Press,
1974).
36. It should be noted that it was not Pudovkin's purpose to proffer a picture of
only one basic plot structure. Rather he intended to enumerate an exhaustive
account of the principles that justify the insertion of a scene in a narrative film. He
believes a scene can be added to a film if it is (1) an answer to a previous question;
(2) a parallelism; (3) symbolism; (4) an instance of simultaneity; (5) a leitmotif.
Only 1 and 4 , 1 presume, are relevant to the discussion of the basic linear narrative
that I take up in the current chapter. On the other hand, 2, 3, and 5 would be
features of something more elaborate than a basic linear narrative. Tangentially, it is
interesting to speculate that perhaps Pudovkin (mistakenly) believed that film is a
language because his formulation of these principles for legitimately adding scenes
to a film bear a passing resemblance to (recursively stated) grammatical rules. This is
not to say that Pudovkin discovered a grammar of narrative or of film, but that he
stated his recommendations in a way that suggests something like a recursive
definition for well-formed sequences—rules for licitlv adding scenes to scenes.
37. John Holloway calls this "proponing" in his Narrative and Structure
(Cambridge: Cambridge University Press, 1979).
38. In some cases, the questions that are answered by a scene or an event may not
feel very acute. For example, the causal circumstances in an earlier scene may appear
so implacablv set out (a typhoon is heading toward an island, for instance), that we
don't have much of a question to ask. Or a character may state emphatically what he
intends to do in the next scene: ' T m going to the saloon to shoot Billy Ringo." The
eventuations of such causes and intentions in later scenes seem better described not
merely as answering scenes but as fulfilling scenes—i.e., scenes that fulfill what is
4. NARRATION 253

predicted, not simply asked, by earlier scenes. Consequently, one should perhaps
enter a special sub-category—that of the fulfilling scene—to the above list. Yet one
should recognize that it is a sub-category of the answering scene since in film fiction
it is always possible for causal and intentional trajectories to make unexpected,
hairpin turns. There is always a question—will the typhoon hit?—(though some-
times a slightly felt one) that underwrites a fulfilling scene. Evidence for this might
be that if we have to leave the movie before we see the typhoon hit the island we are
likely to ask an informed viewer something like "I didn't see it but the typhoon hit
the island, didn't it?"
39. In many cases, an incomplete answering scene may not be recognized as such
at first glance. We may only retrospectively realize that such a scene gave us a partial
answer to a preceding question. Many of the "clues" in classical detective movies
function this way. We might want to call such scenes ambiguous, incomplete
answering scenes; they are ambiguous because their initial significance is different
from their retrospective significance.
40. Pudovkin's parallelisms, leitmotifs, and symbols are digressions from this
point of view. Also, a saloon chanteuse singing a barroom ballad in a western of a
ccrtain period would count as a digression. It would be interesting to investigate the
types of digressions that appear mandatory in given genres in stipulated historical
periods. One might be able to develop a list of recurring types of digressions in
popular film. Needless to say, digressions in a linear narrative should not be
mechanistically treated as automatic deconsmictions of classical cinema; they are
very often part and parcel of the form a filmmaker is working in.
Because of digressions, because of the various types of questioning scenes, be-
cause of the insertion of establishing scenes after the film is on its way, and because
of the possibility of complex temporal relations between scenes (e.g., parallel
narration), we should not anticipate that answering scenes will always follow scenes
that initiate a question. Were we to diagram many movies in order to oudine their
question/answer plan, we would often discover that we would have to leapfrog, so
to speak, several scenes in order to connect question scenes with their answers. The
question scene still cognitively generates the need for an answering scene, but the
answer may not appear immediately after the question. A question scene may be
followed by a digression, followed by an establishing scene, followed by another
questioning scene before we get an answering scene that correlates with our first
question, for example. The more leapfrogging the film involves, the more complex
(and less basic) we tend to think the narrative structure is. By the same token, if a
question is raised that is presented as important—what will happen to character
x?—and it is not answered, even via a complex process of leapfrogging, then we
tend to find the movie incomplete. And if we cannot justify that incompleteness in
terms of some meaningful point or quality that the film is projecting, then we tend
to see such incompleteness in a negative light. A case in point is the failure in The
Hills Have Eyes, Part II, to tell the audience what finally becomes of Jupiter's
daughter.
41. For example, in Theory ofFilm (New York: Oxford University Press, 1960),
254 4. NARRATION

Kracauer often attempts to illustrate the nature of the episodic form by contrast to
what he calls "the intrigue" and to studio-fabricated plots. I am less concerned with
whether K.racauer's argument is sound than with the fart that he represents the
tendency not only to associate the episodic narrative structure with realism and but
to attempt to get a handle on that structure by contrasting it with mass movie
plotting which is fundamentally e roteric.
One might argue that the episodic form has the credentials for being considered
the basic film plot because of its close resemblance to the chronicle. And undoubt-
edly the episodic form is supposed to appear to be a chronicle. But it is in fact a
highly mediated imitation of a chronicle rather than a chronicle pure and simple,
since the events it strings together are selected not because they happened one after
the other but in order to make a point—e.g., "love is fleeting" in La Ronde—or to
evoke a quality—e.g., a sense of a social totality in The Tree of Wooden Clogs. The
kinds of spectator responses required to properly engage such episodic structures
and to divine their purpose is far more demanding and complex than that required
by an erotetic linear narrative.
42. After the last macro-question is answered and Johnny becomes a Confeder-
ate lieutenant, there is one more scene in the film, the saluting gag. From the
viewpoint of the core structure of the linear erotetic film plot, such scenes are
optional, though, of course, they may gready enrich the film as a whole. For an
analysis of the thematic significance of the saluting scene and its relationship to the
rest of the film see my doctoral dissertation, "An In-Depth Analysis of Buster
Keaton's The General (New York University, 1976).
Also, it should be mentioned that there is a standard deviation from the account
of closure in movies just offered. Sometimes, especially in horror films, a movie mav
end by introducing one last unanswered macro-question e.g., is Freddy back at the
conclusions of the first two installments of the series Nightmare on Elm Street. This
secures the feeling of uneasiness and uncertainty to which such films are dedicated.
43. In The General, the battle sequences toward the end—through which
Johnny wins his uniform—may appear tacked on. One reason for this is that the
macro-question of whether Johnny will be allowed to enlist has not been as sus-
tained as it might have been throughout the film. The battle seems extraneous to the
most animated questions the film has raised and the film might have successfully
terminated when Johnny makes its safely to town. In retrospect we recall that his
girlfriend has made enlistment a condition for their relationship, but as the film
unravels the issue gets lost and thus gives the battle scenes, wherein the uniform is
won, an aura of superfluousness.
Isolating the macro-questions in a film—something most easily achieved after
one has knowledge of the complete film—provides a powerful perspective from
which to analyze the entire narrative structure of a film. We observe what questions
are answered last in a film and then back up and enumerate all the scenes that set
forth and sustain the question that elicits the final answer. For example, looking at
the last scenes of Bride of Frankenstein, we see that they answer three narrative
questions: will Baron Frankenstein be persuaded to perform the experiment? (he
5. CINEMATIC N A R R A T I O N 255

will); will the monsterfinallyhave a friend? (he won't); will the Baron and his wife
escape? (they do). Of these questions the last one is a micro-question generated by
the circumstances of the experiment scene. The other two questions, however, are
alternatively the basic issues of the majority of scenes in thefilm.The monster keeps
searching for a friend in scene after abortive scene—thereby reasserting the ques-
tion—while Dr. Pretorius tempts Frankenstein in alternating scenes. Finally, the
questions converge when the object of the experiment becomes the creation of a
female, potential friend for the monster.
What I am calling macro-questions might be referred to by manyfilmtheorists as
"enigmas." This terminology derives from Roland Barthes's S/Z (New York: Hill
and Wang, 1974). At best, this characterization is misleading. Most narrative
questions are not obscure or unfathomable mysteries. Identifying them as such
seems a rhetorical gambit that enables the psychoanalytically inclined theorist to
conflate narrative questions with things that one might more appropriately think of
as enigmas, e.g., the nature of the human subject. Perhaps by calling such narrative
elements macro-questions, we can make a contribution to short-circuiting some of
the most egregious arguments by equivocation that are rampant in film scholarship
today.
44. The historical point made above requires some clarification. Feature film-
making, stricdy speaking, did not supplant serial filmmaking; both arose roughly
contemporaneously, and both continued to be made throughout the twenties and
thirties, though featurefilmmakingclearly became the ascendent form. However, it
also seems evident that when the feature form was evolving, with such examples as
Lang's Spiders, one notion of the way in which to organize feature-lengthfilmswas
to string serial scenes together back to back, resulting in a great deal of aimless
peripeteia; in this light, the introduction of the framing story in Lang's Destiny
might be seen as a means of bringing order between the disparate action scenes by
way of a presiding macro-question.
Of course, the serial mode of narration continues today in such genres as the soap
opera; for an analysis, see my "As the Dial Turns: Thoughts on Soap," The Boston
Review (February 1988).

5. CINEMATIC NARRATION

1. lean-Pierre Oudart, "Cinema and Suture," Screen (Winter 1977-78), vol. 18,
no. 4.
2. These characterizations of the shot/reverse shot figure derive from David
Bordwell, Narration in the Fiction Film (Madison: University of Wisconsin Press,
1985), p. 110.
3. Kaja Silverman, The Subject ofSemiotics (New York: Oxford University Press,
1983), p. 203.
4. Jean-Pierre Oudart, "Cinema and Suture," p. 43.
256 5. C I N E M A T I C N A R R A T I O N

5. Daniel Dayan, "The Tutor-Code of Classical Cinema," in Bill Nichols, ed..

Movies and Methods (Berkeley: University of California Press, 1976), p. 448. This
article has been criticized by William Rothman in "Against the System of Suture,"
which is also anthologized in the Nichols' volume.
6. Oudart, "Cinema and Suture," p. 37.
7. David Bordwell, Narration in the Fiction Film, p. I l l
8. Annette Kuhn, Women's Pictures: Feminism and Cinema (London: Routledge
and Kegan Paul, 1982), pp. 56-57.
9. Daniel Dayan, "The Tutor-Code of Classical Cinema," p. 449.
10. Barry Salt, "Film Style and Technology in the Forties," in Film Quarterly
(Fall 1977).
11. Stephen Heath, Questions of Cinema, (Bloomington: Indiana University
Press, 1981), p. 97. For examples of followers of Heath's expanded view of suture,
see Kaja Silverman, The Subject ofSemiotics, pp. 213-214, and Philip Rosen, "The
Politics of the Sign and Film Theory," October (Summer 1981), no. 17, p. 19.
12. Jacques-Alain Miller, "Suture," Screen (Winter 1977-78), vol. 18. no. 4.
13. Stephen Heath, Questions of Cinema, pp. 52-53.
14. Colin MacCabe, Tracking the Stgntfier; Theoretical Essays (Minneapolis: Uni-
versity of Minnesota Press, 1985), pp. 87-88.
15. Note that the problem I point to here does not rely on a view on my part that
evidence can be divorced from theory. I admit that theory may shape observation.
However, the objection above is that we are confronted only with theory sans
evidence of any sort.
16. David Bordwell, Narration in the Fiction Film, p. 111.
17. This is a point that Richard Wollheim makes against all Lacanian theory in
his T h e Cabinet of Dr. Lacan," New York Review ofBooks (1979), vol. 25, nos. 20
and 22.
18. Philip Rosen, "Adorno and Film Music: Theoretical Notes on Composing for
the Films," in Tale French Studies (1980), no. 60, pp. 157-183.
19. Hanns Eisler, Composing for the Films (New York: Oxford University Press,
1947). Though Adomo and Eisler co-operated on the theories in the book onlv
Eisler signed it.
20. Rosen, p. 174.
21. Rosen, p. 175.
22. Adorno and Eisler, Composing for the Films, pp. 75-76.
23. Adorno and Eisler, Composing for the Films, p. 78.
24. Perhaps the use of ghosts in this account reminds contemporary film theo-
rists of Dayan's talk of ghosts in relation to suture. See the first section of this
chapter.
25. Aaron Copland, 'Tip to Moviegoers: Take off" Those Ear-Muffs," in the
New York Times, Nov. 6,1949, section 6, p. 28. This article is discussed at length in
Roy M. Prcridergast's Film Music: A NegUcted Art (New York: Norton, 1977),ch.
6.
CONCLUSION 257

26. Perhaps the use of the horns here is also suggestive of the hunt, another form
of violent activity associated with high-spiritedness.
27. Friedrich Nietzsche, The Birth of Tragedy and the Case of Wagner (New
York: Random House, 1967), p. 103.
28. The discussion here is limited to nonvocal music because songs with words
generally provide the referential material needed to secure some measure of emotive
explicitness.
29. Peter Kiw, The Corded Shell (Princeton: Princeton University Press, 1980),
p. 98.
30. Johann Adam Hiller, "Abhandlung von der Nachahmung der Natur der
Musik," in Friedrich Wilhelm Marpurg, ed., Historisch-Kritische Beytrage, (Berlin,
1754), 1:524.
31. James Beattie, The Philosophical and Critical Works (Hildesheim and New
York: Georg Olms, 1975), p. 463.
32. Kivy, The Corded Shell.
33. Eduard Hanslick, The Beautiful in Music (New York: Liberal Arts Press,
1957); Nietzsche, The Birth of Tragedy, p. 102.
34. For one source of this view of the emotions see Anthony Kenny, Action,
Emotion and Will (London: Routledge and Regan Paul, 1963).
35. For a thorough account of musical representation see Peter Kivy, Sound and
Semblance (Princeton: Princeton University Press, 1984).
36. Some (Goodmanians) might prefer to say here that the music labels the scene
with a feeling, or that it attaches a feeling label to the scene.
37. In previous discussions of the factors that give rise to the power of movies—
especially the discussion of pictorial representation—I have argued that little or no
training is required on the part of the audience in order for it to comprehend the
symbol in question. It may be the case with movie music that some familiarity with
the musical idiom being employed is requisite. This probably does not require
formal training, and, given the kinds of music movies employ, it is likely that it is
picked up rather rapidly, almost by osmosis. However, even if movie music is more
a matter of cultural conditioning than pictorial recognition, it is still appropriate to
refer to its reception as direct and immediate in the sense that once one is familiar
with the musical idiom one recognizes the qualities in the music without perform-
ing operations of inference, decoding, or reading.

CONCLUSION: PROBLEMS AND PROSPECTS

OF CONTEMPORARY FILM T H E O R Y

1. The modern locus classicus of the position against metaphor as an essential

part of science is Pierre Duhem, The Aim and Structure of Physical Theory (New
York: Atheneum, 1962), especially part 1, ch. 4.
258 CONCLUSION

2. The locus dassicus for the defense of analogy in science is Norman Campbell,
What Is Science? (New York: Dover, 1953), ch. 5. See also Rom Harri, The
Principles of Scientific Thinking (Chicago: University of Chicago Press, 1970).
3. See Mary B. Hesse, Models and Analogies in Science (Notre Dame, Indiana:
University of Notre Dame Press, 1970), for a thorough discussion of the logic of
scientific metaphor.
4. Stephen Heath, Questions of Cinema (Bloomington: Indiana University Press,
1981), p. 13.
5. Some readers may be curious about the relation of piecemeal theorizing to an
earlier essay of mine entitled "Film History and Film Theory," which appeared in
The Film Reader (1979), no. 4. At this time, I regard that essay as really addressing
the small-scale question of the way in which to identify a film as art in terms of the
evolution of style. For further philosophical developments along this line of
thought see my "Art, Practice and Narrative," in The Monist (April 1988).
 INDEX

Abraham, K., 12 Bell, C , 108

Acker, K., 148 Belson, J., 104
Adorno, T.W., 53, 54, 55, 58, 214, 215 Benjamin, W., 53
African Queen, The, 225 Benveniste, E., 120, 121, 123, 124, 150,
Age of Gold, The, 11 151, 152, 153
All That Jazz, 176 Bergman, I., 11
All Quiet on the Western Front, 106 BicycU Thief, 147
Althusser, Louis, 5, 6, 5 3 - 6 2 , 64, 67, 72, Birds, The, 128
74, 77, 79 Birth of a Nation, The, 75, 104
Amatori, 175 Blow-out, 41
Andalusia» Dog, The, 11 Blow-up, 41
Andrew, Dudley, 1 0 0 - 1 BordweU, D., 129, 151, 152, 153, 185,
Anger, K., 148 191-92
Annie, 167 Brakhage, S., 26, 140
Apparatus: cinematic, 10, 1 3 - 3 2 ; ideologi- Brecht, B „ 91, 92, 93, 103, 104, 106
cal, 57, 59 Bringing Up Baby, 164
Aristotle, 42 Bunuel, L., 11
Arnheim, R., 107 Butch Cassidy and the Sundance Kid, 126,
Aura, 123 166
Auteurism, 3
Avenging Conscience, The 11 Cabiria, 203
Cahiers du Cinema, 2
Bandwagon, The, 148 Captain Blood, 25
Back to the Future, 106 Casablanca, 25, 31
Balasz, B., 107 Cavell, S., 1, 107, 112
Barthes, R „ 5, 6, 10, 32, 91, 112 Chabrol, C., 3
Baudry, J. L., 5, 10, 11, 1 3 - 3 2 , 44, 48, Chomsky, N., 2
49, 51, 52, 180, 205, 209, 210 Circus, The, 166
Bazin, A., 3, 4, 40, 107, 111, 114, 119, Classical film theory, 1
125 Clowns, 105
Beattie, J., 219 Comolli, J. L., 127, 137, 138
260 INDEX

Conner, B., 127, 168 Heath, S., 128, 134, 135, 136, 137, 1 6 0 -
Conversation, The, 41 70, 186, 187, 188, 197, 229
Copland, A., 216 Hiller, J. A., 219
Courbet, G., 148 Hitchcock, A., 3, 12, 34, 183
Creation of the Humanoids, 123 Hofmansthal, H., 12, 36
Cremonini, L., 77 How Green Was My Valley, 123
Crocc, B., 108 Hume, D., 81, 83

Dallas, 85 Imaginary, the, 36, 39, 63, 6 4 - 6 6 , 70, 72,

Danto, A., 42 74, 85; 121, 126, 129, 153, 1 6 9 - 7 0 ,
Dayan, D., 184, 185, 186, 188 185
Dead of Night, 166 Inauguration of the Pleasure Dome, 149
Death Wish, 147 Indexing, 2 0 1 - 2
Descartcs, R., 9 9 India Song, 176
Disavowal, 36, 43, 44, 99, 160 Intolerance, 175
DOA, 123 Into the Night, 178
Double Indemnity, 123 Introjection, 6 7
Dream of a Rarebit Fiend, 11 Invaders from Mars, 166
Duras, M „ 148, 170 Invasion of the Body Snatchers, 106
Invasion U.S.A., 88
Irigaray, L., 189
Eislcr, H „ 214, 215
Ellis, J., 115, 1 1 7 - 2 6
Emptre, 27, 189 Kammeradschaft, 106
Kant, I., 92
End of St. Petersburg, The, 8 9
Keaton, B., 178
Engels, F., 81
King Kong, 166
Enunciation, 1 5 0 - 6 0
Kivy, P., 219
Erotctic narration, 1 7 1 - 8 1 , 2 0 6 - 8
Kiss Me Deadly, 164
Kiss Me Goodbye, 106
Fantomas, 178 Klein, M-, 11
Fellini, F., 105 Kracauer, S „ 107, 112
Fetishism, 37, 38, 4 2 - 4 3 , 4 4 Kuhle Wampe, 148
Force of Evil, 222 Kuhn, A., 185
Ford, J., 3 Kuleshov, L., 170, 171
Fort Apache: The Bronx, 104
Frampton, H., 168
Lacan, J., 5, 35, 36, 56, 61, 6 2 - 7 3
Freud, S „ 62, 67, 68, 192
Langer, S., 13
Lang, F., 34, 40, 202
General, The, 178 Last Laugh, The, 203
General Line, The, 147 Last Tear at Marienbad, 49, 170, 176
Gidal, P., 147 La Terra Trema, 106
Godard, J. L., 3, 91, 148 Lawrence ofArabia, 42
Golem, The, 148 LeGrice, M., 231
Gombrich, E., 1 0 0 - 1 Lévi-Strauss, C., 69
Greed, 27, 148 Lewin, B., 11, 16, 17, 28, 29
Green Berets, The, 157 Lifefime, 25
Griffith, D. W „ 4 Lindgren, E., 107
Grand syntagmatique, 5, 9, 10 Lonedale Operator, 176
GungaDin, 217, 218, 221 Lord of the Flies, 106
Lukacs, G., 51
Hagan, M., 141
Halloween, 42, 158 M, 40, 202
Hanslick, E., 220 McCabe, C., 71, 189
INDEX 261

Macro-questions, 178 Rhythmus 21, 140

Mad Max, 159 Road Warrior, 159
Magnificent Ambersons, The, 147 Robbe-Grillet, A., 148
Man With A Movie Camera, The, 148 Roderick Random, 123
Marcuse, H., 53, 54, 55, 58 Rohmcr, E., 3
Marx, K., 53, 84, 85 Roma, 105
Metz, C., 1, 4, 5, 10, 11, 12, 13, 17, 19, Rope, 183
26, 27, 32-52, 121, 126, 146, 152, Rosen, P., 2, 2 1 4 - 1 5
153, 154, 227 Rules of the Game, The, 106, 149
Micro-qucstions, 178
Miller, J.-A., 187
Sachs, H., 12
Mirror stage, 35, 36, 6 3 - 6 6 , 72
Satyriam, 176
Modifying music, 216-25
Saussurc, F., 1, 4, 32, 33, 69
Mother, 88, 106
Scaling, 201, 202
Mothra, 96
Schopenhauer, A., 216
Munsterbcrg, H., 48
Screen, 2
Murder My Sweet, 123
Searlc, J., 144
Secrets of a Soul, 12
NaJud Jungle, The, 42 Siegfried, 24
Nashville, 164 Silverman, K., 71, 78, 79
Neo-Realism, 3, 111 Sontag, S., 112
Nichols, B., 114-17, 128-29, 192 Spellbound, 12
Nietzsche, F., 81, 219, 220 Spies, 178
Street Scene, 222
Oedipus complex, 67, 68 Sturgcs, P., 4
Oudart, J. P., 184, 185, 186, 188, 189 Superman, 149
Sunrise, 27, 110, 203
Pabst, G. W„ 12 Sunset Boulevard, 123
Packard, V., 55 Suture, 6, 185-99
Paitan, 149 Swept Away, 159
Panofsky, E., 107, 112 Symbolic, the, 6 7 - 7 2 , 74, 191
Passion ofJoan of Arc, The, 27
Peck, G., 38, 217, 218
Text of Light, 140
Perkins, V., 1, 112
Tom Jones, 123
Perspective, 127-38, 162-63
Top Gun, 88
Photography, 106-27
Touch of Evil, 161, 166
Plato, 19, 20, 21, 22, 24, 30, 31, 92
Tree of Wooden Clogs, The, 175, 177
Polanvi, M., 43
Trial, The, 176
Porter, E., 11
Triumph cfthe Will, 156
Poseidon Adventure, The, 42
Potemkin, 27 Truffaut, F., 3
Psycho, 222 2001, 43
Typee, 106
Pudovkin, V., 170, 171
Purple Rase of Cairo, The, 106
Uki-e pictures, 133
Rombo: First Blood Part II, 104
Rear Window, 41 Vagabond, The, 203
Red Dawn, 101, 104 Valse Triste, 108
Rebel Without a Cause, 217, 218 Variable framing, 2 0 0 - 8
Regeneration, 203 Voyeurism, 36, 38, 4 1 - 4 2
Renoir, J., 3, 111
Richter, H., 140 Wargoma, 178
Riefensthal, L., 156 Warhol, A., 189
262 INDEX

Wayne, J., 157 Yearling, The, 217, 218, 221

Welles, O., 3, 34, 111, 176 Toung Werther, 123
Wertmullcr, L., 159 Tou Only Live Once, 166
Wild Strawberries, 11
Wittgenstein, L., 100, 192 Zola, E., 108
Wyler, W„ 3, 111 Zorns Lemma, 168