Continuum Aesthetics

Series Editor: Derek Matravers, Open University and
Unversity of Cambridge, UK
The Continuum Aesthetics Series looks at the aesthetic questions
and issues raised by all major art forms. Stimulating, engaging and
accessible, the series ofers food for thought not only for students
of aesthetics, but also for anyone with an interest in philosophy and
the as.
Titles available from Continuum:
Aesthetics and Architecture, Edwad Winters
Aesthetics and Literature, David Davies
Aesthetics and Moralit, Elisabeth Schellekens
Aesthetics and Music, Andy Hamilton
Forthcoming in 2008
Aesthetics and Painting, Jason Gaiger
Aesthetics and Nature, Glenn Pasons
Continuum Interational Publishing Group
The Tower Building
1 1 York Road
80 Maiden Lane
Suite 70
New York
NY 1 0038
© Katerine Thomson-Jones 2008
All rights reserved. No part of this publication may be reproduced or
transmitted in any for or by any means, electronic or mechanical,
including photocopying, recording, or any information storage
or reteval system, without prior permission in writing fom the
British Library Cataloguing-in-PubIication Data
A catalogue record for this book is available from the British Libray.
ISBN 1 0: HB: 0-8264-8522-7
PB: 0-8264-8523-5
ISBN 1 3 : HB: 978-0-8264-8522-9
PB: 978-0-8264-8523-6
Library of Congress Catoging-in-PbJication Data
A catalog record for this book is available fom the Library
of Congress.
Tpeset by Newgen Imaging Systems Pvt Ltd, Chennai, India
Printed ad bound in Great Brtain by MPG Books Ltd, Bodmin,
Preface vii
1 Film as an a 1
Scruton: Against flm as an a 4
Responding to Scrton 6
Amheim: The limitations of flm mae it an a 8
A contemporay view: The extra capacities
of flm mae it an a 1 2
Conclusions 14
2 Realism 1 6
Bain: The flm image i s the object 1 8
Walton: The flm image is tansparent 23
Walton' s crtics: The flm image is not tanspaent 28
Curie: The flm image is highly depictive 3 1
Does the flm image really move? 33
Conclusions 38
3 Authorship 40
Who is the cinematic author? 43
Do flms have authors? 47
Author-based flm criticism 52
Conclusions 55
4 The language of flm 57
Film, laguage and montage 60
Film semotics 62
Let' s face it: Film is not a language! , 65
Conclusions 71
5 Naration in the fction flm 72
Must narative flms always have narators? 74
What is it for a flm' s naration to be uneliable? 80
Two-track unreliability 80
Global uneliability 81
How do flms suppor narrative comprehension? 83
Conclusions 85
6 The thinkng viewer 87
Narative comprhension 87
Interpretation 91
Evaluation 95
Conclusions 1 01
7 The feeling flm viewer 1 03
Making sense of our feelings for (flm) fctions 1 05
The paradox of fction 1 06
The paradox of horor 1 1 0
Identifcation and empathy 1 1 3
Psychoanalytic identifcation 1 1 4
Imaginative identifcation 1 1 7
Empathetic identifcation 1 20
Film form and feeling 1 23
Conclusions 1 27
Notes 1 29
Bibliography 1 38
Index 1 43
Among philosophies of the arts, phlosophy of flm is distinguished by
the extent to which it draws upon the broader theoretical tadition for
the art for in question. One explanation of the attention philosophers
give to flm theory has to do with the fact that the technical development
and public distbution of flm at the end of the nineteenth century did not
coincide with wide recognition of its artistic potentia. Film had to ea
a status in the face of considerable technophobic scepticism. This meat
that ealy flm-maing practice tended to involve the self-conscious
application or demonstation of a teory of ar, and ealy flm critica
practice involved the explicit articulation and defense of the theories
behind the flms. In
tur this meant that, from the very beginning, flm
wrting was highly philosophical. And athough the ar status of flm has
long been secure and the preoccupations of flm theorists have broadened
considerably, there is still a philosophical self-consciousness to the tadi­
tion of flm studies.
Broadly speang, there are thee kinds, or developmental phases, of
flm theory that deal with issues taen up by contemporay philosophers.
Classical flm theor fom the 1 920s to the 1 950s primarily aimed to
defend the emergence of a new art for. Contemporay philosophers
who appreciate such classical fgures as Rudolph Aei, Andre Bazin
and Sergei Eisenstein work to complete or extend aspects of this defence
of flm a-for example, by considering whether flm can represent rather
than merely record reality, whether flm-makng has unique resources for
formal play or expression, and whether flm is inherently more realistic
than other a forms.
The kind of flm theory that was dominant in the late 1 960s ad
toughout the 1 970s - now commonly referred to just as ' seventies' flm
theory, involved the application of semiotics, stcturalism and psycho­
analytic theory to flm. Some topics that interest philosophers brdge
the periods of classical and seventies, or psycho-semiotic, theory - for
example, the topic of authorship, which in the perod of classical theory
was of interest for its connection to art status, and in the period of
psycho-semiotic theory was of interest for its connection to theories of
interpretation. The tendency to refer to flm in linguistic ters was most
powerfl during the period of psycho-semotic flm theory. Philosophers
ae interested to see whether the analogy between flm and language can
stand up to rigorous aalysis and whether drawing such an analogy is
helpfl for explaining how we understand or interpret flms.
Cognitive flm theor fom the 1 980s to the present deals with the
nature of our engagement with flm. There is a great deal of collaboration
and conversation between cognitive flm theorists ad philosophers. This
refects a shaed interest in the application of philosophica, psychologi­
cal and neurological theories of mnd to our understanding of flm
engagement - how the viewer makes sense of what she sees on screen
(and heas on the soundtrack) - paiculaly in terms of following a
stor; how she assigns an overall meaning and value to a flm; how she
supplements what she sees and heas with imagination; and how she
responds with feeling to depicted characters and events, as well as to
aspects of the look and feel of the fl itself.
This book is organized around the issues that interest philosophers
who a thinking about the histor as well as the contemporary state of
both flm and flm theory. Since these issues are raised by flms them­
selves but have often aeady been explored by flm theorists, the aim
here is to make philosophical sense of te theor and, i the process,
illuminate something about the nature of flm. This is not to say that
every question or issue addessed in the book comes to philosophy from
flm teory. The question of whether narative fction flms must have
narators, for example, while inspired by work in literary theory, is
almost exclusively pursued by philosophers. Neverheless, the close
working relationship between flm theory and philosophy is evident
toughout the book.
It is import to realize that this is not a book about flm as philoso­
phy-it will not use flms to illustrate or test philosophical views, nor will
it present an agument for the equivalence of philosophical systems ad
flm worlds. This is a book which takes flm per se as the subject of
philosophical inquiry. It is thus for anyone who wishes to better under­
stad an a for that continues to adapt to diferent moving-image
media only to increase in its power ad appeal.
Given its broad focus on flm as an a form, the book will not ofer
lengthy individual flm analyses. As general philosophical claims about
flm ae advaced, however, they will often be brought home to the reader
with paricular examples. These examples are drawn prmaily fom
narative fction flm whch refects the focus of the kind of flm teory
under consideration. This focus is not meant to imply a lack of important
theoretical work on or philosophical signifcance in other kinds of
flm - say, expermental flm or documentary. If anything, this focus sim­
ply indicates a starting point for the philosophy of flm. Insofa as the
forms and genres, as well as the technological resources, of flm continue
to evolve and expand, there is plenty of exciting philosophica work still
to be done. Ultimately, Aesthetics and Film aims to ofer support and
create entusiasm for this imporant future work.
I would lie to than the seres editor, Derek Matravers, as well as
Berys Gaut, Gregory Currie, Andrew Kania, Paisley Livingston, Tess
Takahashi, Tomas Waenberg, George Wilson and Michael Zryd for
thei invaluable feedback on pats of this book. I would aso like to than
Main Thomson-Jones for his constant suppor at every stage of the
book' s prep

Katherine Thomson-Jones
Is flm an at? Before we can answer this question, we need to be clea
on what we mean by 'flm' . The term ' flm' is abiguous; it refers both
to an a for that employs a vaety of physical media - celluloid, video
and digital formats, for exaple - and also to the traditiona medium of
the a form - the projected flm strip that results from the complex tech­
ncal processes of flming and editing. Sometmes 'fl' is also used to
refer to te art form specifcally when it employs the taditiona flm
medium. Classical flm theorists use the ter in ths way simply because
at the time they were wrting, te flm medium was the only medium of
the a form. Despite this ambiguity, however, there is good reason to
hold onto the ter ' flm' . Most imporantly, the ter is still widely used
by ordinary flm-goers, flm crtics and flm theorists, ad it covers
instaces of te a form in every flmmaing tadition, viewed in any
setting. Thus in this book, we will keep the ter ' flm' but use it cae­
fully by makng a three-way distinction between flm the a for, flm
the medium, and 'photographically-based' flm - te art for in its
taditiona medium. We will also follow common usage in keeping te
ter 'cinematic' to describe a flm, a aspect of a flm or a mode of
engagement with a flm that is defned by or relies upon distinctive or
unique features of flm media.
So how should we understand our original question? - Is flm an art?
I we ae referng to an established at form, then ou question is trivial
at best. I fact, however, the frst answer to ths question established the
possibility of using the term 'flm' to refer to a medium-specifc a for.
In the ealy days of flm, frst-generation classica fl theorsts were
interested in the artistic possibilities inerent in traditional flmmang
processes, parculaly in cinematography. Insofar as cinematography
produces a recording on a celluloid stp to be rn though a projector,
classical flm theorsts were thinng about the atistic possibilities of
the flm medium. But not surrsingly, the way they established that flm
can be at is by scrtnizing the results of using te medium - te flms
projected onto a screen for an audience. It was because ealy flm
theorists glimpsed aistry on the screen that they decided that the prod­
uct of cinematography, editing and celluloid projection could be at.
Nowadays, of course, ' flm' still refers to an art form but not to a medium­
specifc one. As we shall see, this raises the question of how to uphold
the status of an art for which was originally justifed in terms of a pa­
ticular medium when that art for has moved beyond its traditional
medium. Before we consider this question, however, let's examine the
orginal justifcation - how flm frst became art.
To answer our starting question, we might begin by pointing out
the existence of cinematic masterpieces like New World (2006) or The
Seven Samurai ( 1 954). But does this show that (photographically-based)
flm per se is an art? It all depends on what makes such flms master­
pieces - whether it is their inherently cinematic qualities or whether it is
qualities derived from other, established art fors - for example, their
dramatic qualities or their painterly qualities of composition. The real
question, then, is whether flm is an at form in its own right and the
answer to this question will depend on whether what makes a flm a flm
can also be what makes it art.
Today most flm-goers assume without question that flm media ca
serve aristic purposes. When flm frst emerged, however, as a mechani­
cal innovation in recording, ther was no such assumption. If anything,
in fact, there was an opposing assumption that flm is merely a recording
device devoid of artistic interest. This meant that ealy flmmakers
flm theorsts frst had to legitimate their practices before they could
secure a receptive audience. Rudolph Aeim, one of the most promi­
nent early flm theorists, was well aware of how much he had to prove for
the sake of an emerging at form. Both the 1933 and the 1 957 versions of
his treatise on the at of silent flm provide a detailed catalogue of all
the creative and expressive possibilities inherent in the flmmakng
process. Essentially, Aeim accepts the assumption that mere mechan­
ical recording cannot be at and then agues that flm a begins where
mechanical recording ends. The result is an authortative articulation of
the anti-realist prnciples of silent flmmaking.
We see these principles applied in diferent ways in each of the major
silent flm movements. In Soviet montage flms like Sergei Eisenstein' s
Battleship Potemkin ( 1 925), editing is used to break up, rearange and
change the meaning of the recording. In Geran expressionist flms
like Te Cabinet of Dr Caligari ( 1 920), highly stylized sets, acting and
nar ation are emphasized with incongrous camerawork and lighting.
Even the flms of Chalie Chaplin wit their use of visual ticks embody
principles for establishing a brea with reality and the creation of some­
thing entirely new.
But what of the assumption that flm cannot be a if it is mere mechan­
ica recording? Behind this assumption is te following line of thought:
Wen a flm is made, however much thought and creativity go into
writing the script, constructing a scene and reheasing the actors, once
the camera is rolling, that's it: the next crcial stage is beyond the cre­
ative contol of the flmmaers. Of course the cinematographer can
creatively contol the angle, direction and distance of the camera, ad the
editor can creatively contol the order and rhythm of images in the fnal
cut. But no one can creatively control the content of those images - if te
camera mechanicaly records a tree, then you end up with an image of
just that te, just as it looked at the moment of recording. It is this lack
of artistic contol at the crucia and distinctively photographic moment of
the flmang process that allegedly prevents photographicaly-based
flm from being an art form.
I is undoubtedly true that flm is sepaated fom taditional as like
painting ad draa by the mechanical nature of its recording process.
The frther question is whether mechanical recording rules out atisty.
Actually, there ae really two questions her: the question of whether
there can be aistry despite mechanical recording
and the question of
whether tere can be artisty in mechanical recording itself. Early flm
theorists like Aheim only considered the frst of these two questions.
The second question is taken up by the frst generation of sound flm
theorists. Among them is the great Andre Bazin who we meet in the next
chapter on realism, and who locates the power of flm in the immediacy
and accessibility of its recorded imager.
By the time we get to the second chapter we shall be able to appreciate
that Bazin's work is made possible by the prior work of silent flm theo­
rsts in legitimzing flmmang and flm study. In paicular, Aeim's
work is historically important for establishing a certain theoretical
approach to flm, one involving close analysis of everything that maes
flm a unique artistic medium. Since we ae interested, not just in con­
fring the art status of photographically-based flm, but confrming its
independent a status, this kind of medium-specifc analysis is extremely
usefl. It is not, however, the only valid theoretical approach to flm
given tat there are many continuities between flm and other art forms.
We will become paticularly aware of tese continuities when we discuss
autorship and nar ation in Chapters 3 and 5, respectively.
Without medium-specifc analysis, however, we might be stuck at the
view that individual flms can be a when they successfully record a
but flm per se cannot be a. On this view, flm is not an independent a
form because the flming process does not contrbute to the atistic value
of the fnal product. This brings our attention to the fact that flm theo­
rists who want to defend the a status of flm are up aganst two distinct
arguments, both of which involve the assumption that flm cannot be
a if it is mere mechanical recording. On the frst view, flm is teated as
the mechanical recording of real life; on the second view, flm is treated
as the mechanical recording of the established art of dama ad is thus
'canned theate' . While Arheim responds to the frst view, contempo­
ra philosophers of flm have tended to focus on the second view. This
is paly because the canned-theatre agument has been revisited by the
contemporay philosopher, Roger Scrton.
I his much-discussed essay, 'Photography and Representaton' , Scruton
agues that flms are just photographs of more or less aristically valuable
damatic representations. 2 Films cannot be atistic representations them­
selves because photographs are not the knd of thing that can represent:
Thei mechanical production blocks any aistic interpretation of what
is being photographed. The debate concering the a status of flm is
thus not merely of historical interest. Scruton's contempora challenge
reminds us that a proper understanding of flm requires an exanation
of the grounds for assuming, as most of us do, that flm is an at. In other
words, Scrton reminds us that as philosophers we ae committed to
uncovering and testing the most basic beliefs that infor our practices as
flmviewers, flmcrtics ad flmakers.
Scrton's refutation of flm as an at form has thee steps: First, he
assumes that the flm medium is an inerently photographic medium.
Then he creates an argument aganst te possibility of photographs being
representationa a. And fnally, he extends this agument to flm.
In order to assess Scrton's agument aganst flm art, therefore, we need
to assess his agument against photographic at; tat is, unless we discover
that Scrton's agument against photographic a cannot legitimately be
extended to flm. A photograph that has not been manipulated in any way
records the appeaance of its subject. But, Scruton insists, this does not
mean that the photograph represents its subject. A painting of the ver
same subject, on the other hand, does represent its subject. What' s the
To answer this question, we need to understand how Scruton's account
of representation focuses on the relation, established in its production,
between an image and its subject. This relation deterines the knd of
interest we can tae in the image - whether aesthetic or merely instr­
mental. To tae an aesthetic interest in a representational work of a is to
tae an interest in how the work represents its subject. Scrton claims
that photographs fal to inspire this knd of interest; instead they only
inspire interest in what is represented, namely the subject itself. The pho­
tograph is therefore dispensable as a means to satisfy our curiosity about
the subject. What maes the diference here is the way an image is pro­
duced -whether though mechanical recording or through the intentional,
interpretive activity of a representational arist.
According to Scrton, a painting like the Mona Lisa is representa­
tional because it shows us how the atist saw the subject. The style of the
painting manifests da Vinci's decisions about how to paint his subject
and makes the painting interesting whether or not the subject is aso
interesting. Moreover, given that the panting is the product of artistic
intentions, the subject need not even have existed. Compae this to a
imaginar case of a photograph showing a woman dressed and made-up
to look like the subject of the Mona Lisa. Clearly tis is neither a photo­
graph of a Renaissance gentlewoma nor a photograph of a non-existent
woman in the mind' s eye of the photographer. The photograph cannot be
either because the subject has to exist and be in font of the camera to be
photographed. It is not up to the photographer to create te subject and,
as a result, it is not up to the photographer to decide how the subject is
going to look in the photograph. Since the camera simply records how a
actual subject actually looked at a cerain moment in time, the resulting
image has no aesthetic interest as an arist's interpretation.
When we look at a panting, knowing that it is the product of inten­
tional activity, we assume tat its perceptible details were chosen as part
of the style of the work and thus have meaning. In contast, when we
look at a photograph, we assume that its details were not chosen. In fact,
if it is a te photograph, those details could not have been chosen: they
are just the result of the caera automatically recording all the detals of
te subject itself. Scruton insists that it is precisely this alleged lack
of control over detail on the part of the photographer that prevents her
product fom being representational a. Moreover, it is the same lack of
contol in recording tat prevents flms from being representational a.
Ths is the point at which Scrton extends his argument from photogra­
phy to flm:
A flm is a photograph of a dramatic representation; it is not, because
it cannot be, a photographic representation. It follows that if there is
such a thing as a cinematic masterpiece it will be so because - like
Wild Strawberries and L re
le du jeu ¯ it is in the frst place a
dramatic masteriece.3
Scrton goes fhe� to suggest that it is not just tat the flm-recording
process is neutal in terms of its contibution to the dramatic success of
the fnal work, but that it actually makes a negative contrbution: Again
due to a lack of contol over the detail in flm images, it is going to be
harder for a flm audience to know how to interpret a recorded dramatic
scene than for a theatre audience to know how to interpret an analogous
scene on stage. Let's say that we have a flm scene and a stage scene of a
battle. Since the camera records everything in the scene - every splatter
of mud, every glint of steel - the flm audience can be overwhelmed with
and distacted by a plethora of unorganized detal. In contast, since the
staging of a battle in a play is stylized to allow for the foregrounding of
certain features of the landscape and cerain actions, the theatre audience
is properly drawn to the dramatic locus of the scene.
According to Scruton, since a photograph records rather than represents
its subject, it cannot suppor an aesthetic interest in how its subject is
shown. All it can suppor is an interest in the subject itself. A flm is just
a series of photographs and thus also fails to represent. We cannot take
an aesthetic interest in how something is shown on flm because how that
thing is shown is merely the result of a mechanical recording process and
not the result of creative artistic choices.
Given this line of arument, there ae at least two stategies for
responding to Scrton' s claim against flm:
1. Accept that Scruton' s agument against photography automatically
extends to flm and then show that there are some photogaphs in
which we can take an aesthetic interest and whch thereby qualify as
a in their own right.
2. Leave unquestioned Scrton' s agument against photography and
instead question its extension to flm.
The defence of photographic a involved in the frst strategy is con­
vincigly made by Willia King in the approprately titled, ' Scrton and
Reasons for Looking at Photographs ' .4 The second strategy involves
pointing to ways in which flm is unlie photography, ways that suggest
te requisite creative contol for representation. The most signifcant
way in which flm is unlike photography is of course in being a sequence
of images that are combined in any way that the flm arist wants though
editing.s There are, however, a whole range of devices ad conventions
tat ae distinctive to flm and that ca serve artistic purposes. These are
helpfully cataogued for us by Ameim.
Before we tm back to Aheim, however, let's consider King' s
response to Scruton. Remember that Scrton takes it as evidence of the
inability of photographs to represent that the only reason we can have
for looking at them is to satisfy our curiosity about thei subjects. This
way of thinkng about photographs, King responds, can only be a result
of a lack of awaeness or appreciation of the range of photographic tech­
niques and the consequent range of possibilities for atistic intervention
in the photographic process. If Scrton had actually considered real
exaples of photogaphs and the way people talk about them, he would
have reaized that there ae may knds of reason for l<okng at a photo­
gaph. As well as curiosity about the subject, there are reasons having to
do with the evocative power of the image, its fora properties, and its
history of production. Most importantly, however, rasons for looking
at non-abstact ' ar photography' invaably include an interest in the
manner of representation.
King gives us three compelling examples of photographs which
involve artistic interpretation: William Kein' s 'Entrance to Beach, Ostia,
Italy, 1 956' , Ansel Adams' s ' Moon ad Half Dome' , and Ralph Gibson' s
' Te Prest' . In each case, the photographer has efected an aestetic
tasformation of his subject such that the photograph has qualities that
the subject does not have. Moreover, this is done solely by photographic
means, including the use of diferent lenses and development metods.
Thus Klein distances and renders enigmatic an otherwise slightly theat­
ening group of young men by imposing a grainy ' photographic' texture
on the image. Adams lends a quality of uneaity to a moon-lit landscape
by making objects appear larger and closer than normal. And Gibson
foralizes a human subject by framing only the very bottom of the
priest's face ad the top of his vestments in smooth, sharp contrast.6
With these examples, King implies that due to the complexity of the
photographic method, the photographer has just as many options for
presenting her subject as the painter. If photography ca be art and flm
is essentially photographic, then surely flm can be at too. This is not
enough to show that flm is its own a form, however. We need a sense
of the diferences between the aistic resources of photography and flm,
something that Amheim indirectly supplies. But Amheim only takes us
so far in accounting for the aistic clams of flm. It is up to contempo­
ray scholas to complete the account by suggesting that flm ofers
more creative possibilities than almost any other art form, inclUding
panting and drama.
Amheim' s agument for photographically-based fm as an a form has
two stages. First he catalogues all the medium-specifc ways in which the
flm image is diferent fom perceived reality. Then he illustrates all the
ways in which these peculiarties of the flm medium can be exploited for
artistic efect. Since Amheim is workng under the assumption that a
ought to be expressive, he urges flmakers to use all aspects of the flm
medium - fom camera angles to editing - as expressively as possible.
As a result, what we end up with is a handbook of flm techniques for
expression and the creation of meaning.
Aeim is certainly not alone in thining that flm is a expressive
medium. What is distinctive about his view, besides its technical detail,
is the key assumption about the source of expression in flm. According
to Aeim, expression is possible when a flm fals to record something
accurately. It is precisely in the limtations of flm as a mechanica record­
ing device that possibilities for artistic interpretation emerge. It might
seem strange to want to af rm the a status of flm in such a negative
way, and indeed, as we shall see, there ae limts to Aeim's approach.
In focusing on the ways in which we fail to experence a flm as real life,
Aeim misses al the ways in which we experience a flm as more
tan real life. We' ll come to this idea later, however. First let' s see how
Aeim responds to the claim that flm cannot be ar because it is mere
mechancal recording.
According to Aeim, we see that flm fails to record accurately in
all the discrepancies between perceived reality and the flm image.
These include: ( 1 ) a discrepancy in perceived depth; (2) a discrepacy in
the perceived size and shape of objects; (3) a discrepancy in the limits on
our range of view; and, (4) a discrepancy in the experience of the fow of
space and time.
The frst discrepancy in perceived depth is the result of recording and
ten projecting three-dimensional, real-life events (or performances)
onto the fat surfaces of flm stock and screen. With the subsequent loss
of depth, the flm image is experienced as something in between two
and three dimensions. Take te example of an aerial shot of two passing
tains. In the three-dimensional world represented on screen, the viewer
sees one tain moving away from her and one towads her. But on the fat
screen she can also see the tains as moving towards the upper and lower
edges of the fame.7 Moreover, this second impression modifes the frst
so that even in three dimensions there is a signifcant loss of perceived
depth. This is one way, according to Aheim, in which 'flm is most
satisfactorily denuded of its realism' . 8
With the loss of depth comes the loss of what Aheim calls 'constan­
cies of size ad shape' in our perceptual experience. In real life we
perceive an object moving away or towads us as remaining constant in
size even though the image of the object on ou retina changes in size.
When we watch a fl, on the other hand, an object moving towads the
caera appeas to grow lager and one moving away from the camera
appeas to grow smaller. Similaly with shape: in real life we perceive a
rectangula table as rectangula even though te image on our retina is
wider at the font tan at the back. In a flm, however, a rectangular table,
paiculaly one close to te camera, may appea wider at the front than
at the back.9 Of couse Aeim is not saying that you are unable to see
the rectangula table as rectangular on screen, just that this way of seeing
is not automatic.
While te limts of the fat screen contribute to a loss of perceived
depth in flm, they also contbute to te loss of the fll range of vision
that we have in everyday life. By moving our eyes ad turing our heads,
we can see a continuous panorama of our surroundings. But when we
watch a flm we cannot t our heads to see beyond the frame. In this
way, Aheim tinks we are remnded once again of the limitations of the
flmecording medium. l o
Perhaps the most profound discrepancy, however, between perceived
reaity ad te flm image is a fnction of editing. In real life, we cannot
jump instantly to 5 mnutes later or to 5 mles away. We have to pass
trough all 5 minutes and cross all 5 miles. Not so in flm. A scene at one
time and place may be immediately followed by a scene at a totally
diferent time and place. Amheim suggests that it is the pictorial qualit
of flm that prevents the juxtaposition of scenes and shots from disturb­
ing or confusing the viewer: ' One looks at [the juxtaposed scenes and
shots] as calmly as one would at a collection of pictre postcards' .
Once Aheim has listed these and other discrepancies between per­
ceived reality and the flm image, he moves onto the second stage of his
argument. At this stage, he uses a wide variety of examples to show how
te flm arist can exploit te failings of flm as a mechanical recording
device. Behind this project is a paricula theory of a, one that can be
questioned independently of Aheim 's flm analysis. Ultimately, whether
Aheim can succeed in defending the a status of flm is going to depend
less on his brilliant catalogue of expressive flm techniques and more on
the acceptability of his criteria for art status.
According to Aheim, a ought to be expressive in order to serve a
defnite purpose: By highlighting and consequently dawing our atten­
tion to those qualities of things that we would miss in a mechanical
recording, expressive art helps us to understand te te nature of things
and what tey have in common. Now we can begin to see why Amheim
spends so much time listing the failings of flm as a recording device. If
flm were entirely successful in recording exactly how things look,
then, as Scruton suggested, flm would just give us what everyday expe­
rence gives us; namely, an undiferentiated, pragmatic, quantitative view
of the world. It is only when flm fails to record accurately that expres­
sive patters can emerge. The world is interpreted for us on flm and
given meaning. 1
Thus Aheim urges the flm artist not to accept 'shape­
less reproduction' , but to ' stress the peculiarities of his medium' in a way
that 'the objects represented should not th
reby be destoyed but rater
strengthened, concentrated, interreted' .13
For each of the discrepancies between the flm image and perceived
reality, Aheim provides vaious examples of their expressive potential.
Tae the loss of depth in the flm image and the resulting loss of con­
stancy in the perceived size and shape of objects on screen: Amheim
suggests low-angle shots gain their expressive power as a result of these
discrepancies. A close-up, low-angle shot of a police ofcer communi­
cates forceflness because te police ofcer appears to tower above us
with a huge body and a small head. 14 This distorion that makes the shot
expressive is due entirely to the fattening out of the image and our literal
interpretation of its distored proportions. Moreover, the same conditions
of distortion can be used to suggest the relative importance of chaacters
on screen. If one character is signifcantly closer to te camera than
another, the frst chaacter appeas to dwa the second, both physically
and psychcally.
The viewer's loss of an unlimited view due to the framing of te shot
has many expressive functions, one of which is the creation of suspense.
For example, it is a convention in horror flms to show a monster's next
victim on screen and leave, at least temporarly, the monster of screen.
Te fame prevents the viewer fom doing what he would normally do;
namely, turing his head to get a look at what is mang the victim
scream. Instead the viewer must imagine the horors faced by the victim
while axiously awaiting the revelation of the monster. 1
Te fnal discrepancy mentioned above concerng the continuity of
space and time is perhaps the one richest in expressive potential. As we
have seen, Amheim points to the way editing, as an essential component
of the flm medium, subverts mechaical recording by changing the way
we perceive the passage of time and the unity of space. Through editing,
messages can be conveyed, associations and oppositions created. Rapid
cutting can give a sense of frenetic activity or confusion; slow cutting
ca give a sense of lingering nostagia. Once one considers both the
expressive potential of individual shots and the expressive potential of
their combination, the aistic possibilities ae practically limitless. 17
Aeim' s work provides a fascinating account of the technical meas
to expression in silent flm. It has its problems, however. Although we
have not explored this here, Amheim's aguments for why a should be
expressive, what purpose aistic expression seres, and what it is for
works of a to be expressive ae plagued by ambiguity and inconsis­
tency.IS More relevant for us, however, is the restrctedness of his strategy
for defending te a status of flm. As we have seen, Ameim locates the
expressive potental of the flm medium in discrepancies between the
flm image and perceived reality. This means that innovations for achi­
eving greater realism in flm merely reduce the chances of mang flm
a. One such innovation is, of course, sound, which Ameim fatly con­
demns. In fact, for several decades after the tansition fom silent to
sound flm, Amheim continued to insist on the degeneracy of sound
flm. To think the opposite, that te intoduction of sound is a improve­
ment, 'is just as senseless as if the invention of three-dimensional oil
painting were hailed as an advance on the hithero known prnciples of
pantng' . 19
Clealy something has gone wrong if Amheim's approach requires us
to denounce every sound flm as aistically compromised. The historica
explanation is of course that Aeim was working aganst the assump­
tion that flm cannot be art because it is mere mechanical recording. Thus
what he needed' to emphasize were all the ways in which flm fails to
accurately reproduce reality. But to think only in terms of the limitations
of flm as a recording device is to mss the extra capacities of flm as a
representational at for. Instead of thinking of the introduction of sound
as takng away fom the expressive potential of the flm medium, one can
tink of it as adding to the resources of te flm-maker; giving her even
more choices for how to tell a story, convey meaning and evoke emo­
tions. Thining about the sheer diversity of aistic resources available to
the flm-maker has led some contemporar philosophers to propose
diferent kinds of agument for the independent art status of flm. One
such agument is made by Alexader Sesonske in a series of articles tat
focus on our unique experience of flm.20 There is much to remind us of
Arheim in Sesonske' s work but whereas Aheim distinguished flm art
technically, Sesonske distinguishes flm art formaly.
By the time that Sesonske is wrting about flm in the 1 970s and 80s, the
atistic possibilities of the medium have been widely acknowledged such
that there is little danger of flm being dismissed as mere mechanica
recording. However, since new flm media ae also available by this time,
questions arise about the advisability of relying on medium-specifc
arguments for flm at. This could explan why Sesone' s account is
formally rather than technically based -whether a flm is shown on video
or celluloid, many of its formal qualites will be the same. Moreover, flm
may no longer be dismissed as mere mechanical recording but that does
not mean it has secured independent ar status. In fact, as Sesonske
descrbes, the tendency among flmmakers, flm critics and flm viewers
is always to think of flm in terms of some other established art for - for
example, as visual poety or recorded drama. The task for Sesonske is
thus to fnd a way to understad and appreciate flm on its own terms - in
other words, to create an aesthetics of flm. But where to begin?
The frst step, according to Sesonske, in creating the aesthetics of any
a form is an articulation of the range of formal possibilities inherent in
the medium (or, perhaps, media).21 Thus we can say that the defning
formal dimensions of flm are space, time and motion. Sound is imporant
1 2
but not essential, since a flm can be complete without sound. Film shares
space with painting, sculpture and achitecture, time with music, and
motion with dance. You might think that flm shares all its formal catego­
ries wit drama. But space, time and motion ae only integral to dramatic
perormance and not to the draatic work itself, which is defned in
terms of chaacter and action.2
Despite this distincton, however, the
foral overlap between flm and practcally every other art form paly
explains the tendency to think of it as a derivative a. But Sesonske
insists that this is the wrong way of thnking. For every a form that has
a formal category in common with another at form there ae a unique
range of forma possibilities within that category. Thus what flm can
do with space, time, and motion - how it represents them for us to expe­
rence23 - is completely new. According to Sesonske, ' [w]hen we view a
flm our experience of space, time, and motion difers from any other
context of our lives' . 24
Sesonske continues by arguing that the space and time we exper­
ence in flm have a unique duality. Like the space of paintings, flm
space has a two-dimensiona surace (of moving colours and fors) and
a three-dimensional represented depth. But unlike painting-space the
thee-dimensional space of flm is an 'action-space' in which motion can
occur. Ts action-space is discontinuous both with real-life 'natural'
space and with itself. We can have the sense of moving through a flm's
action-space while remaining in our seats in the Cinema. And both our­
selves and the characters (though not usually in the fction) can jump
instantaneously from one location to another.
As well, however, we can jump fom one point of time to another,
thus indicating the paallel discontinuity of flmtime. While there is a
paicular length of time it taes to watch a flm, tere is also a particular
length of time in which the events depicted on screen occur. The 'dra­
matic time' of represented events is more highly controlled tha in any
other a form, including literature: A prehistoric scene in a flm can be
imediately followed by a contemporary scene, or a segment of time in
a continuous event ca simply be cut out. This contol is such that we
may even experence a change in the form of time. For example, a feeze
fame may be experienced, not merely as interpted motion, but as
tough, in the world of the flm, time itself has stopped.25
Despite these peculiarties, space and time in flm do not feel that
diferent to us. We may feel like we have seen an entire event even when
only its highlights are shown on screen and, as Aeim points out, we
accept jumps in location as calmly as though we were tuing the pages
1 3
of a picture book. Indeed, it is ofen the sign of a good flm tat we fal
to notice its unique teatment of space and time. Motion in flm, however,
is a diferent matter. Sesonske thinks that te way that motion in flm
is famed, edited and highlighted by camera angulation is hard to miss.
The frame created by the screen gives motion a direction and magnitde
that it lacks in real life; editing can give motion a new and aesthetically
signifcant rhythm; and, camerawork can lend expressive force to even
the tiniest movement. Moreover, given the discontinuity of flm space,
our relation to perceived motion in flm has temendous range: One
moment we can see a movement on the distant horizon and in the next
moment we can be engulfed and swept along by it through the world of
the flm.
6 If we think of a flm like Cruching Tiger Hidden Dragon
(2000) with its swooping camerawork, it is easy to appreciate Sesonske' s
point here.
Thus Sesonske concludes with the following remark:
In each of these formal categores, space, time, and motion, the
modes that can be realzed in cinema ae unique to cinema. And
though it is sometimes suggested that cinema must be inferior as an
a because of its dependence on mechanical devices, we might note
here that the creative possibilities in flm ae at least as great as in
any a. As in literature, the whole of the world of the work is to be
created, not only the characters and their actions but the ver forms
of space and time in which they act.
The challenge in this chapter has been to show that despite what Seson­
ske describes as a 'dependence on mechanica devices,' flm counts as an
art form in its own rght. It is not enough to show that flm can incoro­
rate aspects of taditional a forms, aspects lie dramatization and
painterly composition. Rather, it must be shown that flm has its own
methods for creating a world on screen for the viewer to enter in imagi­
Here is what we have covered in this chapter:
1 . Scruton' s sophisticated version of the 'canned-theatre' agument
aganst flm as an a for.
2. To kinds of response to this agument; one which stats with Kng' s
defence of te a status of photography as the basis of flm, and one
which daws upon Aeim' s detailed acount of creative uses of the
flm medium.
3. Sesonske' s frther agument for flm as an a for in terms of the
formal possibilities inherent in its media.
Sesonke' s agument shows us tat we do not need to limit our atten­
tion to the taditional medium of flm in order to uphold the a status of
flm. It is also worh noting that skeptical arguments like Scruton' s which
assume that flm is at best a photograph of a dramatic representation can­
not even get of the ground with flms involving computer-generated
imagery (CGI). Wen flmmaers employ CGI, they a not recording
anyting but instead doing something an to painting. Indeed, CGI
is most commonly used to depict scenes and entities that could not
be recorded simply because they ae fantastical or at least wholly fc­
tional - for example, AsIan in Naia or Gollum in Middle Ea. Even
though CGI is usually used to create the illusion of reality - as though
AsIan is part of the 'real' world caught on flm, it still gives te flm­
maer complete freedom over what to represent and how to represent it.
Presumably, then, we can take an aesthetic interest in CGI. The skeptic
might want to claim that flms involving CGI ae not really flms. But
sice ordinay viewers and critics consider them to be flms, this move
reduces the skeptic' s agument to an anachronistic and mainly verbal
dispute about a ealy phase of flmmaking.
1 5
In the previous chapter we were concered with establishing tat flm
is an a in its own right. The form of argument we considered empha­
sizes all te atistically signifcant ways in which flm is not merely the
mechanical recording of reality or theate. In this for, the agument
for flm a leaves a key sceptical assumption untouched; namely, tat
mechanical recording canot be a. Perhaps one could ague that flm is
a signifcant and distinct a for precisely because it involves mechani­
cal recording. Then one would have to say what it is about flmrecording
that has artistic signifcance. This may not be so difcult to do if one
looks to the history of flmmaking.
I the late 1 920s, a revolution occured in flmmaking with the intro­
duction of synchronous sound. Once the sound flm was a technical (and
commercial) possibility, tere was no looking back; indeed, flmmaers
and flm-goers aike tended to assume that the intoduction of sound
represented a great improvement to the a form. Why would they assume
this? If they were thinking like Rudolph Amheim, after all, they would
assume the opposite. As we know, Amheim saw the introduction of
sound as a frher obstacle to creative expression which requires a diver­
gence between flm and reality. Given that sound flm advocates did not
share Arheim' s worries about sound, they must either have given up on
flm being a or have switched to a new flm aesthetic. Thinking about
some of the flms made during this period - flms like Rules of the Game
( 1 939) and Citizen Kane ( 1 941 ) , convinces us that the latter and not the
former possibility must be the case.
As usual, flm theory was left to catch up with a flmmaking practice
that embodied a new set of assumptions about the nature and value of
flm. Fortunately, one flm theorist was paying close attention to the
changing flm world and understood the theoretical implications of
the sound-flm revolution. Andre Bazin, who began writing about flm in
the 1 940s and continued until his deat in 1 959, is considered by many
to be the most imporant fgure in the history of flm theory.29 This is for
two reasons: ( 1 ) Bazin was active as a flm critic, historian and theorist
during the most important period in the history of the flm - the period in
which sound flm came into its own and established a new aesthetic of
flm; and (2) Bazin wrote about the flms and flm-maers of this period
in a highly original, provocative and insightfl way. The fndamental
theoretica claim in his writings is that flm is distinguished from other
a fors by its distinctive capacity for realism, a capacity tat derives
fom the nature of the flmaking process. With te introduction of
sound, flm-makers cae to rcognize this capacity and exploit it in ever
more refned ways.
Bazin' s intuition about the distinct and supreme realism of flm is
widely shad. Even though some flms and some flm styles are consid­
ered more reaistic than others, if you compare flm to other art forms, it
seems to stand out for its realism in a surprising way. To see what is
surprising here, compare flm and teate: When you watch a play you
ae seeing rea people - namely, actors - up on stage. When you watch
a fction flm, on the other hand, you ae merely seeing images of actors
on a fat screen. Surely this implies that theatre is more realistic than
flm. And yet our intuitions point in a diferent direction as refected in
te tendency to describe some compaatively unrealistic flms as 'theatri­
cal' . Why is this? If we t to explain flm realism in terms of te way
cinematic porayals are often particularly intimate, raw and afecting,
we ae merely going around in circles. For it would seem that the afec­
tive power of flms is due in pa to thei realism and this is what we
ae trying to explain.
As we shall see, philosophers of flm have tended to follow Bazin
in ting to understand flm reaism in terms of the flmmakng process.
The idea is that flms ae made up of photographs ad photographs, as we
leat in Chapter I , have a mechanical basis and thus a causal connection
to thei subjects. This connection is taken by diferent philosophers to
imply diferent things. Kendall Walton taes it to imply that photographs
ae tsparent - we see their objects though them.30 Gregory Currie, on
the other hand, who is one of several crtics of Walton, takes the casua
connection to imply that photographic and cinematographic images ae
in a class of representations tat aso includes such natural signs as
tee rings and footprints but excludes paintings and drawings. 3l To see
what is appealing about each of these views, we need to consider them as
pa of a long debate tat began with Bazin in the 1 940s, sufered a
extended interption durng the heyday of psycho-semiotic flm theory,
ad then recently reignited among aalytic philosophers with interests in
perception and the mind mor generaly.
1 7
In a way it is surrsing that the contemporary argument about flm
reaism still focuses on the photogaphic basis of flm given that the
traditional medium of flm is fast being replaced by digital formats. It is
an open question, therefore, as to how one would defend the realism
of digital flm. Insofar as digital images ca be as much mechanical
rcording as celluloid photographs, some traditional aguments for flm
realism may still be applicable. But this will only be in a limited way
given that digital images can also be endlessly maipulated.
It is worh noting tat digital flms in which extensive manipulation
is obvious - if only because the flms portay fctional locations and
fctional entities - can still be considered hghly reaistic. Perhaps this is
a hold-over from the days of photographically-based flm: Perhaps, at
some level, we still tend to think of flm, the a for, in terms of flm,
the photographic medium. On the other had, it might suggest that the
realism of flm images had nothing to do with their causal history in the
frst place.
Another limitation of arguments for flm reaism tat start with the
photograph is tat tey have little to say about the realistic contbution
of one seemingly essential32 ad distinctive feature of flm; namely,
motion. Now that flm is a multi-media a form, the impression of motion
on the screen can be created in a varety of ways. When flm involves a
celluloid flm stp, however, this impression is created by rapidly run­
ning a series of static photographic fames though a projector. In light of
this method for creating apparent motion, a debate has sprung up among
philosophers as to whether the motion we see on screen is a straightfor­
ward illusion or in some sense real. Perhaps it seems obvious to you
tat flms do not really move. But the obviousness of this conclusion may
be an imporat clue to understanding our entrenched and lagely unques­
tioned ways of thinng about flm realism.
In a series of essays collected in What is Cinema ?33 Bain defends real­
ism in two stages. First, he examines the nature of the flm image - its
mechanical production and causal connection to the object - in order to
show that realism is the essence of flm a. Then he agues that a
particular style of flmang is the realist style. This is the style that
Jean Renoir perfected in the 1 930s with flms like Rules of the Game. It
is also the style of several post-war America directors, including Orson
1 8
Welles and Billy Wilder, and Italian neorealists like Vittorio De Sica and
Roberto Rossellini. We will review the defning characterstics of this
style in a moment, but before we do, we need to consider the intelligibil­
it of Bazin' s approach. If he is going to argue that flm by its ver nature
is realistic, thereby implying tat any flm in any style would be realistic,
how can Bazin also t to isolate a particula realist flm style? The way
Bazin tes to do this is by claiming that the style of Renoir, Welles, et al.
is the style that most fully realizes the nata purpose of the traditional
flm medium. Given the mechanical basis of flmmang, this purpose is
to 're-present' reality in an 'objective' and thus insightful way. There are,
however, problems with what Bazin is tying to do here. These will
become clear as we examne his view in more detail.
The so-caled realist style that emerged in the flms of Renoir is in
many ways the same style used in manstream flm today, although tech­
nical innovations have rendered the style gradually more transpaent
over time. Among its chaacteristics ae the following: ( 1 ) The use of
medium-long shots - shots in which the whole of an actor's body is
visible on screen, often with space above his head; (2) the use of deep
focus techiques as opposed to the sof focus techniques of 1 930s Holly­
wood flms; (3) the composition of shots with dramatic details located on
diferent pictorial planes rather tan concentrated in the foreground; (4)
the use of the long take to capture a whole scene instead of breakng it
down into several shots; (5) the use of camera movement instead of edit­
ing to follow the action; and, (6) a non-theatrcal, non-painterly use of
te fae as a rsult of the camera following the action.34
All of these chaacteristics are evident in the style of Rules of the
Game though not all of tem in every shot. And tough there ae impor­
tant diferences in narative stucture and subject-matter between Renoir,
the Italian neorealists, and the American post-war directors, tey shae a
comtment to the use ofsequence shots in continuous space and thus an
opposition to the highly interentionist techniques of montage. But why
does Bain think tat this style and no oter is realistic? The reason is
that he tnks this style comes closest to a simple recording of what is
going on in font of the camera. If the actors' performances are captured
in a single long take with a moving camera, rater than in a series of
edited shots, the natural unity of space and continuity of time is pre­
sered. The moving camera also gives the action a feeling of independence
- as though the camera is simply following what happens instead of
fng a staged scene. Finally, and perhaps most imporantly, the use
of deep focus photography enables the viewer to scan the recorded scene
1 9
for interesting details just as she scas her surroundings in real life. The
overall perceptual experence of a Renoir flm is therefore closer to
ordinay perceptual experience precisely because the recording aspect of
the medium is emphasizedø
This maks a tansitional point in Bazin's argument - the point at
which he identifes his favoured flm style with the essence of flm art.
Recall that for Aheim, the essentially cinematic qualities of paicula
flms ae their exprssive qualities, and these qualities ae the result of
medium-specifc divergences fom mechanical recordng+ For Bazin, in
direct contast, the essentially cinematic qualities of particula flms a
their reproductive or presentationa qualities, and these qualities result
directly from mechanical recordinge Indeed, it is te mechanical basis of
flm that raises the a form in its original medium above the traditional
reprsentational as: Wereas a representatonal painting, for example,
only stands in for realityg a flm or a photograph retrieves reality. Bazin
explains this rather strange idea in the following way:
The objective nature of photography confers on it a quality of cred­
ibility absent fom all other picture¯mang« In spite of any objec­
tions our crtical spirt may ofer, we are forced to accept as rea the
existence of the object rproduced, actually re-presented, set before
us, that is to say, in time and space. Photography enjoys a cerain
advantage in vire of this tansference of rality from the thing to
its reproductionr 35
Bazin then goes even furher with te following claim: 'The photo­
graphic image is the object itself, the object feed from the conditions
of time and space that gover it' . 36 Wat could Bazin mean here given
that we usually have no difculty distinguishing between a photograph
and the thing photographed? Perhaps we should take it as mere hyper­
bole or metaphor ad assume that Bazin is simply ting to emphasize
the diference between te photographc as and the taditional repre­
sentational as. Takng the claim literally, ater all, would seem to
commt us to a doubtfl metaphysics ¯ one that allows for the identity of
tings which ae clearly not one and the same; namely, a photograph
of something and the thing itself. Perhaps, though, Bazin' s claim is to be
takn literaly once it is more flly spelt out. From some of his other
remaks, it is plausible to take Bazin to mean that the way in which the
image and its object are identical is akin to the way in which two casts
from the same mold are identical. Alteratively, it is also plausible to
tae Bazin to mean that the image and object ae identical insofa as they
produce the same image on the retina. These interpretations a certainly
possible but they need to be explained and their implications explored.
Just as the static photograph re-presents a moment of reality, so the
'moving' flm image, on Bain's account, re-presents a segment of reality
with a paicula duration:
Viewed in this perspective, the cinema is objectivity in time. The
flm is no longer content to preserve the object, enshouded as it
wer in an instant, as the bodies of insects a preserved intact, out
of the distant past, in amber . . . . Now for the frst time, the image of
things is likewise the image of thei duration, change mummifed as
it were.37
At ths point, one might begin to wonder if, by focusing on the record­
ing aspect of the medium, Bazin has not so much succeeded in defending
flm as the supremely realistic a for, but rather, has succeeded in
confrng the suspicions of those who want to deny tat flm is any knd
of a for. Afer all, mechanical recording, even if it is facilitated cre­
atively, is not in itself either creative or expressive. It might seem,
terefore, that either Bazin has to give up his account of flm realism or
he has to give up on flm's claim to ar status. But, not surprisingly, he
does neither. Rather, he works with diferent assumptions about the
natre of a in order to claim that flm is art precisely because it is inher­
ently realistic. Iterestingly, however, it seems that realsm is an aesthetic
value for Bazin for some of the sae reasons that expressiveness is an
aestetic value for Amheim: A highly realistic visual artwork just like a
exprssive visua awork gives us a new way of seeing the world:
The aestetic qualities of photography ae to be sought in its power
to lay bare the realities. It is not for me to sepaate of, in the com­
plex fabric of the objective world, here a refection on a damp
sidewalk, there the gestre of a child. Only te impassive lens, stp­
ping its object of al those ways of seeing it, those piled-up
preconceptions, that spirtal dust and grime with which the eyes
have covered it, is able to present it in al its virginal purty to my
attention and consequently to my love.38
As we have aleady mentioned, Bazin wants to derve a paricula flm
style from his account of the nature of the flm image. But it is not clear
2 1
how he can do this without contadiction. Given his clam that the photo­
graphic process necessarly results in images that re-present reality,
Bazin has surely committed himself to describing as realistic any flm in
any style that is made in the usual, photographic way. Indeed, Bazin
himself seems to realize this commitment when he claims that no matter
how distorted or fzzy a photographic image, ' it shares, by viue of the
very process of its becoming, the being of the model of which it is the
reproduction' . 39 If Bazin will allow that fuzzy shots are realistic, why not
highly expressionistic or otherwise stylized ones? However, just as with
his rather extravagat claim tat the image is the object, there may be a
way to interpret Bazin's attempted dervation more moderately: Perhaps
Bazin's point is just that, since our experence of flms in the style of
Renoir, Welles, et al. reproduces our ordinay perceptual experience in
some respects more closely, than ealier styles like Soviet montage and
Gera expressionism, the style of Renoi, Welles, et al. beter expresses
the inherent realism of the medium. This is not an implausible sugges­
tion even though Bazin does not suppor it with a fully developed
As we shall see, the idea that flm and photography are realistic
because their manner of production suppors a certain kind of perceptual
experience is taen up by the philosopher, Kendall Walton. Walton pro­
vides the agument that Bazin needed, avoiding both metaphysical
absurdity and contadiction, though, as his several critics will tell us,
there is still something a litle stange about his claims. Just as we
may balk at Bazin' s claim that the object is 1e image, we may bak at
Walton's claim that we see the object tough the image.
Before we come to Walton, however, ther is one other crticism of
Bazin's account worh mentioning. The criticism concers the applica­
tion of his account to fction flms. As Noel Caroll points out,4
that flm images re-present their objects, Bazin is · committed to saying
that thei subjects ae always real. Thus, in the case of a fction flm,
Bazin can only tal about the flm re-presenting actors and sets rather
tha characters in fctiona settings. But this is surely a stange way of
talking given that what is most relevant to our viewing of a fction flm is
that it represents certain chaacters. In addition, it is not as though the
r of fction f
ms is a minor or maginal one; on the contray, the
bulk of flms most of us watch, and indeed the bulk of flms tat ae
made, ae fctional naratives. And Bazin himself spends most of his time
with fction flms. Yet he does not have an account of the way that the real
object allegedly re-presented by te flm image can also stand in for or
nominally portay somethig fctional.
Once more, however, Walton comes to the rescue. He suggests that in
ou experience of fction flms, it is actally the case that we see the
actors through the photographs projected on screen, but it is also fction­
aly te case that we see the characters. In other words, we actually see
the actors, albeit indiectly, but we imagine seeing the characters. If a
fction flm is flmed on location rater than in a studio, it is bot actally
ad fctionaly te case that we see the scene of te story. This explains
why some flmaers aim to give their works of fction the quality of
home movies or documenties - say, by using hand-held cameras ad
only natural lighting. This stylistic choice emphasizes te recording
aspect of te medium ad, consequently, may make it easier to imagine
that you ae seeing the characters, rater than just the actors, trough
te flm. Thus, on Waton's view, fction flms ca functon on diferent
levels, both realistically and imaginatively. Wheter or not this solution
is a good one, however, it is not a solution that is available to Bain. On
his view, the image just is te object, ad insofa as the object is not
a fctional porayal of anythng, neiter is te image.
A related worry will emerge later in the chapter with Gregory Cure's
response to Waton. As we shall see, Curre is concered that Walton's
agument for tanspaency is just like Bai' s in that it takes photographs
out of the category of representatons altogeter. Even tough Walton
addresses this concer in a response to Curie, it is interesting that it
keeps re-emerging whenever the causal relation between a photograph
ad its object is emphasized. You may remember fom Chapter J that
Roger Scruton had exactly te same concer and was quite adamant that
te photographs, given teir mechanical basis, canot represent. The
challenge, therefore, is to fnd a way to ackowledge the imporance of
te photographc process for an account of flm realism and, at the sae
tie, account for te realism of fction flms as fctions.
Like us, Walton is interested in what is behind Bazin' s confusing clam
tat the photographic image is the object. Thus he is interested in the idea
tat photographs ae reastic by thei very nature, in viue of having
been made in a cerain way. This idea seems rght to Walton and serves
as a staring point for his account of the tanspaency of photographs.
Walton agrees with Bazin that photogaphs ae not realistic in virtue of
thei appearance or in virtue of their accuracy. Even a fzzy or distored
photograph is realistic on their views, and in a way that a painting cannot
be no matter how naturalistic or illusionistic.
Thus it is suggested that there is a fundamental gap between photogra­
phy and painting in terms of realism. A point tat is often made much of
in establishing this gap is that paintings can be of things that do not exist
whereas photographs cannot be. Walton ad Cure, however, both thin
that this point is not that signifcant. Cure suspects that it is mere preju­
dice to say tat a painting that uses a model to represent Saint Anne, for
example, is a panting of Saint Anne but a compaable photograph that
also uses a model is not thereby a photograph of Saint Anne but only a
photograph of a model representing Saint Anne. 41 Walton, on the other
hand, points out that even when a painting is of something actual, it
cannot be as realistic as a photograph. His exaple is a painted porait
compaed to a photographic porait of Abraha Lincoln: Why, Walton
asks, is the latter more realistic than the forer?4
The answer, Walton thinks, is that we quite literally see Lincoln
through te photograph but not through the painting. In other words,
when we look at the photograph of Lincoln, while we recognize it as a
photograph, we neverteless see through it as we see through such tings
as eye glasses and mcroscopes. We do not merely have the impression
of seeing Lincoln, nor do we actually just see a duplicate of Lincoln.
Rather tan being a way of gaining access to something tat we cannot
see, the photograph of Lincoln is a way of indirectly seeing what we
cannot see directly. Thus a photograph is like eyeglasses and microscopes
both in its tanspaency and also in being an aid to vision. But, you may
ask, how can a photograph allow us to see Lincoln when he is dead? In
other words, how can a photograph allow us to see into the past? Recall
that Bain confronted tis issue when he claimed that the photographic
image is the object 'freed fom the conditions of time and space that
gover it' .43 In other words, on Bazin' s view, a photograph preserves or
' mummifes' an object at a particula moment in its existence so that it is
always available to us as long as the photograph exists. Walton does not
want to say that an object from the past is literally brought forward into
the present by a photograph. But he is quite comfortable with the idea
that we see into the past when we see an object trough a photograph.
This is not because he relies on some special sense of ' seeing' - he takes
himself to be using the ordinary sense throughout his discussion - but
because he establishes imporant analogies between other fors of indi­
rect seeing and our experience of photographs.
The way tat Walton does ths is by using a slippery slope argument
tat gives his transparency claim ' initial plausibility' . First Wlton asserts
that no one will deny that we see through eyeglasses, mirors, and tele­
scopes. How, he then asks, could someone deny that a security guard also
sees via closed circuit television or a spors fan sees via a live television
broadcast? But ater going this fa, surely we ca also speak of seeing
athletic events via delayed broadcasts. The diference is that now we're
seeing events in the past. But Walton is not sure this diference matters.
Afer all, we tal of seeing the explosion of a star that occured millions
of years ago through a high-powered telescope. The question is whether
any of the diferences on the slide down the slippery slope are enough to
mak a theoretical distinction between seeing and not seeing something.4
Walton thins that we cannot justifably stop the slide before we get to
photographs, and photographs ae at the bottom of the slope. However,
some of Walton' s crtics dig their heels in on the slippery slope by speci­
fing conditions for seeing that photographs fal to satisfy. We will tm
to the challenge of the slippery slope in the next section. First we need to
consider Walton' s necessary and sufcient conditions for transparency.
According to Walton, the reason that we can see through photographs
is that they ae caused by thei objects in a mechanical way.45 Although
the photographer can play an important role in determning what we see
through her photographs, we nevereless see through them a a result
of a mechanical process that preseres the necessary causal relation for
perceptual contact. As a result of this causal relation, the appeaance of a
photograph is 'counterfactually dependent' on the appeaance of the
object photographed. This means that the photograph would have looked
diferent if the object had looked diferent. Counterfactual dependence
can also be presered without such a causal relation, however. Lincoln' s
painted porait might also have looked diferent if Lincoln looked difer­
ent. But tis is the result of a change in the painter's beliefs mediating te
relation between the painting and its object. If Lincoln had looked difer­
ent to the panter, then the painter would have painted Lincoln lookng
diferent. Unless, of course, te painter is not really seeing what Lincoln
looks like because the painter is hallucinating. Then the appearance of
te painting will not reliably depend on the appearance of its subject.
Wit the photographic porait of Lincoln, however, even if the photog­
rapher were hallucinating, the appeaace of the photograph will reliably
depend on Lincoln' s appeaace. To sum up, photographs are tanspaent
because they ae caused by their objects and thus exhibit counteractual
dependence which is not mediated by beliefs. Cure helpflly calls this
kind of counterfactual dependence, 'natural dependence' , in contrast to
the 'intentional dependence' exhibited by paintings of actual things.4
The next question to consider is whether natural dependence is a
necessay and sufcient condition for tanspaency. Walton defends the
necessity of natural dependence with the imagined case of blind Helen
who is ftted with a prosthetic device through which a neurosurgeon
feeds her visual experences. Walton thinks that our intuitions tell us
that blind Helen is not seeing through her prosthesis because her visual
experiences a mediated by the neurosurgeon's beliefs. Even if the neu­
rosurgeon is completely tustwory and makes every efor to feed Helen
accurate inforaton, Walton thinks this does not count as a case of
seeing. It is worh noting, however, that the intuitive force of this case
may not be as strong as Walton thinks. Cure, for one, thinks that Helen
is seeing through her prosthesis precisely because the neurosurgeon is
trustworthy. Imagine instead, Currie suggests, that it is an omnipotent
and omnibenevolent god that feeds all of us our visual experences. Even
though our visual experences ae mediateq in this case, could we not
still say that we see, thanks to the complete reliability of our source of
visual experences?
Leaving this question aside, Walton comes up himself with a case
that shows that natra dependence is not sufcient, on its own, for trans­
paency. He asks us to imagine a machine which is sensitive to the light
refected by an object such that it accurately records the object, but as a
verbal description rather tan as a picture. In this case, the machine' s
prnt-out is caused mechanically by the object and would have been
diferent if the object were diferent. The prnt-out thus exibits natural
dependency. But, Walton concedes, te machine's prnt-out is surely not
This case leads Waton to propose another condition with which natu­
ral dependency is jointly sufcient for tanspaency: the preseration of
real similarity relations. To see what this condition consists in, consider
again the description-generating machine. Walton decides that the reason
that this machine's prnt-out is not tansparent is that it suppors a difer­
ent knd of discrimination than the knd suppored by ordinary perception
of the object described. The best way to see this is to compae the knds
of erors we ae prone to make about te descrption and about the object
described. Let's say the description is of a house. In ordinary perceptual
experence, it may be easy to mistae a house for a ba. But in reading
te descrption of the house, this is not the kind of mistake we a liely
to mae. With the description, it is much more liely that we will mstake
the house for a hearse because the words ' house' and ' hease' look simi­
la. With a photograph of the house, on the other hand, we ae likely to
mae the ba mstake and not the hease mistae. In oter words, we ae
likely to make te same knd of mstae 's we make in ordinary percep­
tual experience. The knd of discrimination suppored by the photograph
establishes the same kind of similaties and diferences as in ordinary
perception. And, according to Walton, a 'process of discrimination counts
a perceptual only if its structure is thus analogous to the stucture of
te world. When we perceive, we ae in this way, intimate wit what is
pereived' . We cannot see trough the mechanical description because
discrimnating between words for things is crucially unlike discrminat­
ing between the things themselves.48
Thus, on Walton's view, the taspaency of photographs must be
distinguished fom the opacity of both paintings and descriptions. A
painting may be counteractually dependent on its object but only via
someone' s beliefs about the object. Thus a painting does not satisfy the
condition of natural counterfactual dependence. A mechanical descrip­
ton, on the other hand, exhibits natura dependence and thus has what
Walton considers necessay - depending on the intuitive force of the case
of blind Helen - for transpaency. But since we cannot, in fat, see
tough the description, natural dependence must not be sufcient on its
own for tanspaency. Te frher condition for tanspaency that Walton
intoduces - the preservaton of real similaity relations - is perhaps not
tat easy to grasp, in pat because it is not that easy to compae modes of
discrimination. As Walton' s critics point <ut,49
it may be that some but
not all of the erors we ae likely to mae when looking at a certain kind
of representation ae te same as te erors we make in ordinar percep­
tua experience. Then the question becomes, how much discriminator
overlap is enough to establish the tansparency of a certain kind of repre­
sentation over aother kind.
In response to tis question, Currie provides a case involving two
clocks, A and B, designed to show that natual dependence and the pres­
ervation of ral similaity relations a not jointly sufcient for
tsparency. The orientation of clock Ps hands govers the orentation
of clock B' s hands by means of a radio signal. Thus there is a natral
dependence between my visual experience of clock B and te appea­
ance of clock A such tat any erors I make in perceiving B, I would also
make in perceiving A. But if clock A were out of sight and I were just
looking at B, there is no plausibility to the claim that I a seeing clock
A when I look at clock B. In this case, even though there is complete
discrminatory overlap, such tat clock B undoubtedly preserves real
similarity relations with clock A while being naturally dependent on
clock A, there is no tansparency. Whether or not we ae convinced by
this case, the use of counterexaples more generally represents one pop­
ular kind of critical response to Walton. Clearly it is time to consider
more closely the overall structure of the critical debate surounding
Walton' s account.
The controversial nature of Walton' s claim that we literally see through
photographs has motivated close crtical attention to his argument.
The resulting debate can be initially chaacterized in ters of two gen­
eral responses to Walton's slippery slope agument. The frst response
involves taking up the challenge to fnd theoretically signifcant difer­
ences between photographs and oter things that we ca agree we see
though, like mirors, eyeglasses, and mcroscopes. In this way, critics
try to stop the slide down Walton's slippery slope. The second response
also involves stopping the slide but stopping it before it stas. A crtic
can refse to tae the frst step onto the slippery slope by denying tat we
see through mirors, eyeglasses, mcroscopes, telescopes and television
footage. Then a comparison between these things and photographs will
not yield te conclusion that photographs are tanspaent.
There ae at least two forms that such a refsal has taen in te litera­
ture. Gregor Currie efectively refuses to step on to the slipper slope
when he gives an account of pictoral representation that would seem to
apply to mirrors. 5
If mr ors ae representatons, on Cure' s view, then
we do not see though them. But mirors are close to the top of Walton's
slippery slope - something that we can supposedly all agree we see
through. More generally, Jonathan Frday refuses to step onto the slip­
pery slope when he defends a theor of perception that rles out any
for of indirect seeingY On the direct realist theory of perception, we
cannot be said to see through any device, be it a miror, a microscope,
or a photograph. The frther challenge for tose crtics, like Currie
and Frday, who refuse Walton's chalenge of stopping the slide down
te slippery slope once it has stared is to account for why it is tat we
nevereless have the sense that we see through such things as mirors
and microscopes. Friday attempts to give an eror theory52 to explain
this sense and criticizes Cure for failing to do likewiseø
Tere are a number of suggestions for theoretically signifcant difer­
ences between photographs and the varous devices that precede them on
Walton' s slippery slopeø Although we can only briefy mention them
here, they are all worth closer consideration. The general aim with these
suggestions is to show that, just because we think that we see through
mrors, telescopes and the like, we ae not thereby comitted to thin­
ing that we also see though photographs. Edwin Matin suggests two
diferences that might stop the slide: the diference between real and
viua images,53 ad the comparative length of the causal chain between
te object of perception and the medium of perception. Nigel Waburon
gives four conditions for seeing that photographs fail to satisfy.54 These
ae virtual simUltaneity' between what is seen ad our seeing it, 'tempo­
ra congrity' between the duration of our visual experience and the
duration of the event that we experience, the sensitivity of our visual
experience to changes in what we experience, and knowledge of what
caused us to have a cerain visual experience such that we can position
ouselves in relation to what we see. There are, however, ways of chal­
lenging the relevance of all of these diferences.
Walton himself gives counterexamples that efectively undermine
Main's conditions for seeing. The case of the compound microscope,
the lower lens of which produces a real image seen though the upper
lens, shows that seeing need not only involve virtual images. The case
of a series of 1 0,000 mirors aranged to relay light from the object to
the viewer shows that te length of the causal chain need not afect
seeing ¯ that is, as long as one thinks that we can at least see through
one mror. In response to Warburon, Jonatha Frday suggests that
Waburon s conditions for seeing ae not obviously essential to seeing.
Virtual simultaneity may not be essential depending on how we inter­
pret a case Friday imagines that involves adding a delay function to a
prosthesis which facilitates sight. The conditions of temporal congrity
ad sensitivity to change only apply to the seeing of events and not to the
seeing of objects. And, fnally, in relation to Warburton' s fourth condi­
tion, Friday suggests that it may be sufcient for seeing to know that
there is a causal link between our visual experience and its object wit­
out having the kind of kow ledge of the causal process that would allow
us to position ourselves in relation to the object.
These arguments deserve closer atention than we can give them here.
But the way is still open for others to take up the challenge of stopping
the slide down Walton's slippery slope. Perhaps there ae other signif­
cant ways in which photographs ae unlike those devices that we agree
we see though. At present, however, it appears that Walton' s slope is
even more slippery than his critics have assumed. Given just how slip­
pery, perhaps it is best not to step onto the slope in the frst place. This
brings us to Cure' s classifcation of photographs as natural representa­
tions. Cure agrees with Walton' s emphasis on the diference between
what Cure calls the natural counterfactua dependence of photographs
and the intentional counterfactua dependence of paintings. Whereas the
appeaance of a panting of something actual depends on te appeaace
of that thing only insofa as the painter' s beliefs ae simlaly dependent,
the appearance of a photograph depends on the appearance of its object
regadless of what the photographer believes.
Walton can t to use this distinction to ground the tanspaency of
photographs because our ordina perceptual experiences exhibit the
sae natra dependency as photographs. However, as we have seen,
Cure thinks that the argument for traspaency fals because Walton is
unable to successfully specif necessay and sufcient conditions for
transparency. This falure simply confrs Currie's sense that the difer­
ence between natural and intentional dependency does not mak a difer­
ence between seeing things and merely seeing pictres of things. Rather,
it marks out two categores of representations. In the category of 'natu­
ra' representations we fnd photographs, but also natura signs like tree
rngs and footprnts. In the category of 'intentional' representtions we
fnd drawings and paintings. Photographs and paintings may be diferent
kinds of representation, and tis is important to realize, but as representa­
tions, both photographs and paintings stand in for things; they do not
mediate our seeing of things.
Interestingly, Walton claims that, in fact, he does not mean to deny
that photographs ae representations. 55 As representations, photographs
function to generate imagined perceptual contact with fctional charac­
ters and events. Walton insists that his point is that photographs are
representations tat we also see though. Tis of course raises the ques­
ton of whether representations, particulaly pictures, ae the kind of
thing that one can see through (to thei Obj ects). Cure seems to assume
they are not when he argues that photographs ae natural representations,
not tanspaencies. But what we really need here is a theory of pictoral
representation or depiction. In the next section, we shall see how Currie
derves a paicular explanation of flm realism from just such a theory.
Frday suggests that what maes Walton's slippery slope so slipper is
a paicular kind of causal theory of perception; namely a rpresentaton­
alist causal theory of perception. For te representationalist, ordinay
seeing is mediated by a mental representation whereas for the direct
realist, ordinay seeing is essentially unmediated. Furhermore, for the
representationalist, the diference between what counts as seeing and
what is mere hallucination depends on the relation between te mental
representations we see through and the tngs we see though tem. And
this relation is just the kind that Walton insists upon for tanspaency;
namely, a relation of natural counterfactual dependence. Thus we can see
why representationalism might be te lubricant for Walton' s slippery
slope: As Friday explains, 'if there is no objection to saying that we see
tough mental picturs that meet cerain conditions, why should we
object to te claim that we can see through physical intermediaies that
meet the same conditions?
Te issue for Friday is that repre�entationalist teories of perception
have been widely discredited and replaced by direct reaist theories.
But direct realist theories do not allow for seeing in virue of any media­
tors, mental or physical. Thus the diect realist does not step onto the
Walton's slippery slope. There is, however, according to Friday, a cost
associated with resisting the slippery slope. The cost is having to explain
ou intuition that we see through such things as mr ors, microscopes,
ad indeed, photogaphs, even though, according to the reaist, our intu­
iton is mistaken. Friday accepts this cost ad argues that the intuition is
te result of our confsing a new way of seeing wit a new attitude
towads what is seen. (This is Friday' s eror theory, mentoned ealier.)
In light of the way that photographs ae produced, we tend to take a prac­
tcal attitude towards tem as means of discovering thei objects. In other
words, we teat photogaphs as though tey give us direct perceptual con­
tact to tei objects when we teat them as evidence, proof of presence or
as windows onto past events. 57 We may teat photographs as though tey
a tanspant but tat does not mean they really ae tansparent.
So far we have considered two explanations of the distinctive realism of
flms and photographs. Bain' s explanation is tat a photograph literally
puts us in the presence of its object. Walton' s explanation is that we liter­
aly see te object through the photograph. Both of these explaations
mak a fundamental divide between photographs and other kinds of
3 1
pictures on the basis of their distinct modes of production. Cure's
explanation, on the other hand, is an explanation of the realism of pic­
tures in general, or, more specifcally, an explanation of what maes
depiction, the distinctive way in which pictures represent, distinctly real­
istic. The degree to which a representational work of a is realistic
depends on how much it depicts rather than represents in some other
way -say, linguistically. What distinguishes flm realism, on this account,
is the 'depictive potentia' of flm media. Other knds of visual a can
depict states of afairs but only flm can also depict time and space. In
other words, flms can be more thoroughly realistic than other kinds
of representational a because they can represent more of the world real­
istically. But what maes depiction so realistic?
Currie's account of depiction is based on a paricular theory of mind
and an attention to what is going on with the viewer. On this account,
when a picture depicts an object, we employ the same recognitiona
capacities in relation to the depiction as we would in relation to the
object itself.5
8 The likeness between a horse and a picture of a horse, for
example, is one of appearance whereby the horse ad the picture shae
properes signifcant for our recognition of horses - te horse ad the
picture tgger the same recognitonal capacity.59 Of course there a
many properties tat a picture of a horse has that a horse doesn't have
(fatness, for example), and vice versa. But all that matters is that the
horse and the horse picture shae the propery (or set of properties) that
tiggers the horse recognitional capacit, whatever that property may be.
Wen the picture tiggers your horse-recognitional capacity this does
not mean that you think there is a horse in front of you. 6 Your horse-rec­
ognitional capacity is a diferent kind of menta operation tha rational
judgment - Cure describes the capacity as a ' quick-and-dirty mecha­
nism which, somewhere deep in my visual-processing system, identifes
a certain input as a horse' . In other words, it is a more automatic and less
fexible process that works, ' not on the basis of a detailed, comprehen­
sive examination of the visual input in the light of background belief and
all the rest, but on the basis of just a few clues extracted from the visual
input itself' .6
It is, however, precisely because my horse-recognitional
capacity is the knd of menta operation it is that depiction works as it
does. Because my horse-recognitional capacity is quick-and-diry, it can
be fooled by things like donkeys at dusk ad stufed horses, but particu­
laly by horse pictures.
Tus depiction is an essentially reaistic form of representation because
it works by exploiting our visua capacities to recognize the objects
depicted. And flm, according to Curre, is the a form that uses depic­
ton most pervasively and naturally. This is seen best with the aesthetically
signifcant decision to depict, rather tan represent in some other way,
spatial and temporal relations between objects as well as the objects
themselves. A flm depicts time and space when it uses the temporal and
spatial properies of its elements to represent te temporal and spatial
properies of te things represented. A succession of events, for example,
ca be represented by te length of time it taes to obsere them on screen
ad the order in which they a experienced on screen. And te height of
a six-foot man on screen can be represented by his being a deterinate
height in relation to te other characters and the scene as a whole.
Even tough depiction is not the only way to represent space and
time, it is te default in flm. A flm-maker could decide to represent the
passage of time wit a fade, wit a dissolve or with the words 'twenty
yeas later' suddenly appeang on the scren. 62 But in the absence of
such devices, we assume that the time it takes to represent an event is the
time the event itself takes. If it taes 5 minutes to represent a conversa­
tion, we assume the conversation lasts 5 mnutes. That is, unless the
conversation is obviously broken up, perhaps by being intercut with
other scenes.
Curie can then explain why Bazin took the style of flmmakers like
Renoir, which involves long-take, deep-focus photography, to be inher­
ently realistic. This is because Renoir's style enhances our ability to
detect the spatial and temporal properties of a flm scene by using the
natural capacity we have to detect those properties of things in the real
world.63 Deep focus, in paricular, allows us to shift our attention at will
fom one object on screen to another, just as we ae able to do in the rea
world. With montage style, on the other hand, quick cutting between
distinct spatial and temporal perspectives requires us to judge spatial and
temporal properies by means of inference fom the overall dramatic
structure of te flm. Curie is thus able to show, with a coherent and
plausible account of depiction, that Bazin was really onto something
when he interpreted the aesthetic signifcance of intoducing sound to
flm in trs of realism.
Perhaps so far in ths discussion of flm realism, you have felt there is
something missig. In moving seamessly fom the analysis of photographs
to the analysis of flm images, we have not drawn attention to the fact
that flm has extra capacities for realism that photography lacks. The
most obvious of these capacities is the capacity for depicting motion.
Whereas a photograph might capture a kung-fu fghter frozen in mid-air,
a flm can capture te kung-f fghter's entire balletic tajectory. Perhaps
Bazin' s descrption of the way flm ' mumifes' chage fails to do jus­
tice to the dynamc quality of the a for - just as our world is fll of
movement, so are te worlds of flm. We can follow the kung-fu fghter' s
moves because the image moves - or does it?
If the kung-f fghter'S leap is captured in a single shot from a fxed
perspective, we can assume that the actor playing the kung-fu fghter
really moved before the caera. But if we are watching the kung-f flm
in a cinema, the movement we see on the screen is taditionally generated
by the projection of 24 static frames per second.6 Knowing this, we might
say that the taditional projection mechanism for photographically­
based flm merely creates an illusion of moton for creatures with ou
particular perceptual apparatus. In other words, the answer to te ques­
tion in this section' s title - ' does the flm image really move?' - is
staghtforwadly, 'no' .
But perhaps this answer is not so straightorard. Curie
5 and Noel
are fully aware of the static mechanical basis of the impression
of motion in flm, and yet they suggest that te flm image really does
move just because it moves for us. While tis does not mean tat flm
motion is real in some ultimate, metaphysical sense, it does mea that
flm motion is not an illusion. In the same way, colours perceived under
the right conditions ae not illusor even if tey are merely appaent.
If colours were always illusor, we could not make sense of the fact that
we distinguish special cases of the illusion of colour.
7 We can be right
and wrong about what colour someting really is because we take the
rea colour of something to be the colour it appears to normal viewers
under normal viewing conditions. This is why colours ae commonly
referred to as real, response-dependent properties. Currie and Caroll
thnk that flm motion is real in the same way because normal viewers
under normal flm-viewing conditions realy see the flm image move.
Currie does not think he has to give a positive agument for his view
because te burden of proof is on tose who distst our experence of
flm motion when, in general, we should sta out by tustig our percep­
tua experiences.
Cure therefore proceeds by attempting to undermne
te opposing view. Specifcally, he challenges two common arguments for
'illusionism,' the commonsense view that motion in flm is an illusion.
On the frst agument, the motion of flm images is an illusion because
when we look at the flm strp apat from the projector, there is no motion;
tere is just a series of still negative images. Both Currie and Caroll
have a ready response to this agument: The illusion tat is up for debate
is the illusion of motion on the screen, not on the flm strp. It is the flm
engaged by the projector that appeas to move. Consider the analogous
case of musical recordings: If we hold an audio CD up to our ear, we
don' t hea any music. But that does not mean that te music we hear
when we play the CD on our stereo is an illusion.
The second agument for illusionism that Currie considers points to
the fact that it is specifcally creatures like us that see motion on the
screen. On a scientifcally objective descrption, the flm projection pro­
cess is just a matter of light paicles hitting the screen. We could also
give an objective descrption of the way that the projection apparatus
engages the human perceptual appaatus. But movement only enters in
when we descrbe the subjective experience that consistently results fom
this engagement. Thus only a subjective and not an objective description
of the flm projection process involves motion. But Currie also has a
ready response to this agument: All that the agument indicates is that
flm motion is a rea, response-dependent propery like colour.
Cure then considers a compelling objection to his own proposal;
namely, that it destroys the distinction between the real and the merely
apparent, eliminating many or perhaps all illusions. Consider the famous
Miller-Lyer illusion, which involves two hori;ontal lines of equal length,
the top one with an outwad-pointing arowhead at each end and te
bottom one with an inwad-pointing arowhead at each end:
( )
The illusion we have when we look at these lines is that the bottom
line is longer than te top line. But, on Currie's account, couldn' t we say
that the greater length of the inwad-pointing-arows line is a real,
response-dependent property? Afer all, none of us can help but see the
geater length, given the knd of perceptual appaatus we have. If we
accept Currie's proposal, couldn' t we say that anything we perceive is
real in a response-dependent way - the mirage shimmering in te desert,
te way that flm chaacters seem to approach us when we are watching
a flm with 3D glasses, or the way that water in waves seems to move
towads us on the beach?
Cure says 'no' . Our intuitions tell us that there is a important difer­
ence between genuine illusions like the Miller-Lyer illusion and te
phenomenon of flm motion. The reason te Miller-Lyer illusion is a gen­
uine illusion is because te misleading appeaance of the two lines can
be undermined by a simple and readily available 'independent check' -
say, by measurement with an ordinary ruler or even with a piece of sting.
By contast, Curie tinks that te ordinary flm viewer cannot check
whether he is really seeing motion on the screen. This is because our
experience is of images, ad not physical objects, moving on screen, and
there is no way to check te movement of images independently of our
experience of them.
Andrew Kania objects to this response, however. Surely, he counters,
there is an independent test for the illusion of flm motion: Simply by
slowing down the projector, the viewer can see, with the naed eye, the
static projected frames go by, one by one.70 Moreover, to object tat tis
merely alters te conditions required for the impression of motion misses
the point. Kaia argues that the point is not to claim that since you don' t
see motion on screen when the projector is slowed down, i t must be an
illusion tat you see motion on screen when the projector runs at noral
speed. This would not convince anyone, just as it would not convince
anyone that the Miller-Lyer illusion is not an illusion if you covered the
arowheads so that we no longer see a diference in the length of the two
lines. Instead, the point of slowing down the projector is to bring some­
one to an understanding of what is going on at a lower, mechanical level
so that she changes her mind about what is going on when the projector
operates at normal speed. 7l
To this, a Curiean might respond by invoking te case of solidity from
the history of science. When scientists discovered tat nothing is solid in
the way we tought - that is, nothing is thoroughly homogeneous, they
didn' t conclude that tables, chairs and bas of gold ae not solid. Analo­
gously, it is not clear that we should deny motion to flms just because we
know tat their underlying bases do not exactly move.7
But Kania insists that ts is a bad aaogy. I te case of solidity, paa­
digmatic cases were examined and it was found that nothing is solid in
the way we tought. However, since there really is a distinction to be
maked between things we cal solid and things we cal liquid and gas­
eous, we kept tang about solidity in the same way but with the under­
stding tat solidity is someting diferent than we had tought. By
contast, flm motion is not a paadigmatc case of motion. I we discovered
that paadigmatic cases actually involve an object being stationay in
24 slightly diferent positons at 24 diferent moments in a second, then
we might say that motion consists in something other than we thought
and flm images move paradigmatically. But when Cure asks us to
acknowledge a new knd of
response-dependent motion, he is not sug­
gesting that we have been confused about the natre of ordinay motion
al along.
This is the point at which Kania brngs the debate to a close. He
concludes that the initial objection to Currie' s notion of real, response­
dependent flm motion remains: By invoking the category of rsponse­
dependent properies, Cure secures te reality of flm motion only by
destroying, or at least severely depleting, the category of illusions as a
It is worh noting that the illusionist can allow for technica conditions
under which the flm image really would be moving. Kania asks us to
imagine that the impression of motion is generated by a projector con­
tinuously shining light through a single flm fame that changes over
time - perhaps, as Kania suggests, the frame is a liquid crstal display
contolled by a computer. Since the images projected on screen would
paallel the actual movements within the fame, they would really be
continuously and contiguously moving.73 Kania sees no problem with
saying that the same flm projected in the standard way and projected in
ths special way realy moves in the latter case but not in the foner case.
At te very least, the compaison between the two cases makes it more
difcult to say that the standadly projected flm really does move: ' Just
as you can set up what looks like a Miller-Lyer illusion, but in which the
line that looks longer actually is longer, there can be something that you
might assume to be an illusion of movement but which is in fact an
example of movement proper' .74
Car 911, on the other hand, would likely fnd such a position incoher­
ent. In response to a diferent case, he insists that to say that the same
flm viewed in the cinema and viewed at home on the television really
moves in the latter case and not in the foner is 'patently contradic­
tory' .75 The mechanism that generates the impression of movement in the
cinema - namely, the projector - can (at least in prnciple) be decon­
stucted in such a way that we can see with the naed eye how movement
is generated. But the mechanism that generates the impression of move­
ment on TV cannot be deconstcted in an analogous way. If the
slowed-down projector is the test for an illusion of movement, and the
same knd of test cannot be conducted on a TV-screening of a cerain
flm, Caroll suggests that the illusionists' agument is defeated simply
by the intoduction of new technology for screening flms.
This is rather a hasty conclusion, however. There could be other kinds
of test applied to other kinds of flm-screening technology that revea
perceived motion to be an illusion. Just because the slowed-down projec­
tor test fails to apply to flms shown on TV, the illusionist is not forced to
conclude that the motion on TV is real. To think otherise is to set up a
false dichotomy. And even if there ae no such tests for the illusion of
movement produced by other flm-screening technologies, tis need not
afect the illusionists' argument for the case of flms projected in the
traditional way. While some illusionists might like to show that flm
motion is always a illusion, there is noting contadictor about their
saying that, depending on how the impression of motion is generated, it
either is or is not illUSOry.
As a reminder, here is what we have covered i n this chapter:
1 . Bain's confsed but fascinating attempt to show that tere is a single
reaist flm style that realizes te true purpose of photographically­
based flm.
2. Walton's agument for the transpaency of photographs tat preserves
the Bazinian insight that te realism of photographs derves fom the
photographic process.
3. The crtical response to Walton that raises questions about the condi­
tions for seeing though a photograph, the very possibility - given the
nature of perception - of seeing through a photograph or any other
kind of mediator, and potentially signifcant diferences between see­
ing through photographs and seeing though other things such as
mr ors and mcroscopes.
4. Cure's agument for the realism of depiction and the superior depic­
tive potentia of flm media.
5. Te debate between Cure, Caroll and Kania about wheter the ta­
ditionally projected flm image really moves just because we cannot
help but see it move.
Before taing up the question of flm motion, we had come fll cicle
in our quest for a proper understanding of the pictorial realism of flm.
Currie's crtique of Walton's broadly Bainian account of photographic
tansparency leads him to give an account of flm realism that aso pre­
serves some of Bazin' s centa insights. Cure himself declaes that
one's view about the pictorial realism of flm does not ental a paricular
view about the realism of motion in flm.76 But in te Bazinian spirit;
Currie seems compelled to push reaism as far as it will go in accounting
for the nature of flm. Perhaps you think he has gone too far when he
posits a new knd of response-dependent motion. But if nothing else,
Currie's view about flm motion forces us to a new awarness of the
grounds of the commonsense illusion view. Such, an awaeness is invalu­
able given te increasing diversifcation of technological means for
generating the impression of motion in flm.
In Chapters 1 and 2, we leat that photographicaly-based flm has dis­
tinctive capacities for astically signifcant efects like representation,
expression, formal play and realism. We might, therefore, assume that
someone who creatively exploits these capacities is a flm arist. More
generally, if there a works of flm a, then surely there must be flm
artists. But who ae they? Given the technically complex and frequently
collaborative nature of flmmaking, it is often difcult to kow where to
assign creative ownership and responsibility. This problem is tradition­
ally construed as a problem about the authorship of a flm, which refects
the long-standing infuence of the literary paadigm in flm teory. In this
chapter we will consider whether some flms have authors and what role,
if any, the flm author plays in interpretation ad criticism. We will
encounter various complications along the way, most notably with the
distinction between an actal author - an actual person with a crcial
role in flmang, often thought to be the director, and an implied
author - an imagined fgure constrcted in interprtation. But before we
enter into the debate about flm authorship, we need to gain a sense of its
The tradition of thinking about flm in terms of authorship or creative
individuality frst gained promnence in France with the flm joural that
Bazin co-founded in 1 95 1 , the Cahiers du Cinema. The critics of the
Cahiers advocated a new approach to flmmaing which would result in
'auteur cinema,' ' auteur' just being the French word for 'author' . The
critic spearheading the campaign for the new approach was Fran�ois
Trufautg who later became an imporant New Wave director.
In the January 1 954 edition of Cahiers, Trfaut mounted a attack
on what he caled the 'tadition of qUaity' in French flmmang which
involved the adaptation of French litera classics. Films in this tradition
were, according to Trufaut, derivative, stufy ad stylistically forulac.
Most imporantly, they failed to explore the possibilities of the flm
medium. In contrast, auteur flm was inherently cinematic, beaing the
mak of an orginal and creative ' cineaste' . Among French directors,
Trufaut points to Jean Renoir ad Robert Bresson as makng innovative
ad highly individual flms. These directors were more easily able to do
this because they wrote teir own screenplays. But wit vision and a
proper appreciation of flm form, even a diector workng with someone
else's screenplay in a restctive studio system could contrbute to 'auteur
cinema' . Thus Trufaut and other Cahiers critics celebrated the work of
many Hollywood directors, including Alfred Hitchcock, Orson Welles,
Rober Aldrch and Nicholas Ray. Moreover, when the American critic,
Andew Sarris, tried to tu Trufaut's ideas into a theor of criticism, he
went so far as to claim that only studio directors working with other
people's screenplays ae candidates for the highest auteur status.
The sudden popularty of American flm among post-wa French
critics was partly due to historical circumstance. During the Liberation
perod in France, all the Hollywood flms from the late 1 930s and ealy
1940s that had been banned during the Occupation, fooded into Pars.
The heady experence of watching so many Amercan flms closely
together gave French cinephiles a unique appreciation of continuities
of style and the emerging personalities of directors across bodies of
work. Tus in his 1 957 Cahiers article, 'La Politique des auteur,' Bazin
described auteurist crtical practice as ' choosing in te artistic creation
te personal factor as a criteron of reference, and then postulating its
permanence and even its progress fom one work to the next' .77
The aims of the Cahiers crtics who advocated auteur cinema were
polemcal rather than theoretical. Trufaut's aims, in particular, were to
make room for a new knd of flmmakng in the conserative ad hierar­
chcal French system, and force an appreciation of distinctly cinematic
achievements by individual artists. Unavoidably, however, Trufaut com­
mitted himself to varous theoretical assumptions which were later
arculated by Sars. The problem, as we shall discover, was that once
te theory behind auteurist practice was made explicit, its contradictions
ad limtations were hard to ignore.
Since Sars thinks that a good flm is typicaly the product of a good
diector, he works to specify directorial standards tat also fnction as
critical standads. Criticism of these standads coincided with a lager
theoretical shft in flm studies under the infuence of semiotics and
strcturalism. At frst, attempts were made to reconcile the role of the
auteur wit a new emphasis on the impersonal codes and structures that
supposedly run across groups of flms. Auteur-stcturalism was one
such attempt: Emerging in the late 1 960s, it postulated the auteur as
te stcturing unconscious latent in the work of a paicula diector.
4 1
On this view, the actal director cannot be the auteur because the auteur
is a fnction of a particular kind of flm anaysis. The immediate result of
the structuralist infuence was thus a shf fom thinking of the auteur as
an actual person - for example, the directors Renoir and Bresson, to
thinking of te auteur as a crtical construct. In literay theory, the move
fom stcturalism to post-stcturalism meant that that the author-fgure
came to be regarded wit great suspicion. If you have read any post­
stucturalist or postmoder literary theory - particularly the work of
Michele Foucault or Roland Barhes, you will aleady know about the
ultimate rejection of the notion of authorship as a repressive principle of
interpretation. This idea has had signifcant infuence over flm theor.
And yet, flm critics have continued to refer to the author or auteur, as
though we cannot help but think of certain flms as te product of a
creative pla or as vehicles for personal expression.
Recently, philosophers have become interested in the tenacity of te
notion of te flm author. By forulating a clear conception of the flm
author and her role in interretation and criticism, philosophers can hope
to ret some theoretical legitimacy to the endurng critical interest in
authorship. But as we shall discover, this is no easy task, pariculaly
because the authorship debate takes up three diferent kinds of claim
about the flm author. Most of te theorists we will consider touch upon
all thee. However, it is fa to say that taditional auteursts like Sarris
were mostly concered with the third claim, auteur-stcturalists were
mostly concered with the second claim, and contempora philosophers
ae mostly concered wit te frst claim:
1 . The ontological claim: Many flms ae authored. A flm can be seen as
the creative product of a single individual.
2. The interpretive claim: The best way to understand a flm is as the
creative product of a single individual - that is, as manifesting an
individual' s artistic vision.
3. The evaluative claim: We can judge a flm to be good or bad on the
basis of its author' s reputation or creative stuggle.
In order to defend eiter the interpretive or the evaluative claim, one
frst has to assume the ontological claim. But there is a fndamental chal­
lenge to the ontological claim which derives fom the way most flms a
made. If you ae someone who likes to watch the credits roll at the end
of a flm screening, you will know just how many technicians, crat­
speople and aists ae involved in any given flm project. Many of the
roles listed in the credits - from lead actor to best boy, are highly
specialized such that it is unlikely that any one person involved in te
project - even the diector or producer, controls every detail of the flm­
maing process. And yet despite kowing how complex and collaborative
flmmaking can be, we still regulaly credit flms to a single person, usu­
ally the director. Is tis just a convenient shorthand or does it refect some
tth about flm authorship? The question that a defender of the flm
autor needs to confont is whether it maes sense to claim authorship
for flms that ae serially manufactured - in other words, the vast major­
ity of flms, made by a series of specialists - fom the screenwriter al the
way to the editor, building on each other' s work at each stage of produc­
tion. Tere ae several replies to the question of whether serially
manufactured flms can have authors:
1 . The diector ca be the sole author of a flm because his' -role in super­
vising and coordinating the activities of others deterines the aesthetic
signifcance of the flm.
2. Someone (or something) besides te director can be sole author for
the same reason that their role deteimines the aesthetic signifcance of
the flm - for example, the screnwrter, the sta, the producer or even
the stdio system.
3. Insofa as there are mUltiple atistic collaborators on a flm project,
there can be multiple authors of a flm.
4. A flm can have a single, implied author who is constcted to help
with understanding te flm.
5. Films do not have authors. Their aesthetic Signifcace is not deter­
mned by a generative agent, actual or implied. Instead, their aesthetic
signifcance might be determined by unconscious forces or by the
In this chapter, we will discover that there is some truth in every one
of these replies.
To ask this question maes it sound as though a crcial assumption about
the nature of authorshp in flm has aeady been made; namely, that there
is only ever one author for an authored flm. There could be several rea­
sons why most theorists interested in authorship make this assumption.
For example, it could be due to the infuence of the literay paadigm
ad its emphasis on single authorship, or the infuence of the Romantic
conception of aists as solitar, tortured souls aiming at personal expres­
sion. Watever the reason, however, there is an imporant challenge to
the single authorship assumption, which we will consider in the for of
a agument by a contemporar philosopher, Bers Gaut. Before we do
this, however, let us consider the most common answer to our staing
question, which is tat, if a flm has a author at all, it is the diector.
It is easy to understand why this answer is so popula. Among ordinary
flmgoers, it is common practice to identify flms by their directors - as,
for example, when we talk about seeing the latest Scorcese or admirng
the work of Ang Lee. It is aso common practice for critics to refer to
a cerain knd of director as an auteur, one whose flms are all recogniz­
ably hers. However, as well as the assumption of single authorship,
claiming that a flm's author is the director rests upon a second, crcial
assumption; namely, that the cinematic author is a rea person. As we
know from our brief surey of the history of authorship in flm stdies, a
shift away from thinng of the author as a real person and towards think­
ing of him as an interprtive constct occured under the infuence of
semiotics and stcturalism.
One reason to be wary about makng te author a real person derives
from the single authorship assumption. If one is hesitant to single out
one of the many real-life collaborators on a flm project as the flm' s
author but neverheless holds onto the assumption that a flm can only
have one author, then one may be attracted to the idea of constructing a
single author for a suitably unifed collaborative flm. Other reasons for
adopting a constctivist view of autorship a tied up with particular
theoretical comtments. For example, we shall see that Peter Wollen' s
commitment to strcturalist analysis leads him to identify the director' s
unconscious preoccupations, as tey emerge throughout a body of work,
with a implied author. As well, under the infuence of postmodem liter­
ary theory, some flm theorsts conceive of authorship as one subject
position among many in relation to 'reading' the flm 'text' .
But maybe we have got aead of ourselves. Before considerng who
occupies te authoral role in flm, we must make sure that we understand
what this role consists in. In oter words, we must have a solid concep­
tion of the author, which fts both with the ordinar use of te ter
ad its use in discourse about the as. In what follows, we will compae
two philosophical conceptions of the author: Paisley Livingston' s realist
conception ad Alexander Nehamas's constuctivist - or, as he likes to
call it, tanscendental - conception.
The most general defnition of an author in the Oxford English Dic­
tionar is 'the person who originates or gives existence to anything' . As
well, however, an author can be an inventor, constructor or founder, the
cause of events, a prompter or mover, an authority or informant, a com­
poser or wrter, a diector, rler or commader, and even the ' one who
begets' - in other words, a father or ancestor. In discourse about the arts,
the author is frst ad foremost the creator of a literay work. Architects
ad composers ae sometimes refered to as authors, as well as the
creative authority behind a lage-scale aistic project like the ceiling of
the Sistine Chapel. But the litera author is paradigmatic; indeed, even
outside discourse about the as, this kind of author may be the frst that
comes to mind for many people.
If we want to apply the literay notion of authorship to flm, we need
to decide how the literay notion is related to the complex ordinary­
language notion. In the past, literary theorists with cerain agendas have
been tempted merely to stipulate a use of the term 'author' . But, as
the contemporary philosopher, Pasley Livingston, points out with pa­
ticula reference to Michel Foucault, a stipulated usage is not going to
help us make sense of how we already think about the author. This leads
Livingston to a pragmatic analysis of the term as a means of attribution
for the sake of establishing socio-cultura context or the proper taget
of critique and response. The result is a 'provisional' defnition of
'author' as a 'term of a' which coincides with but also helps disam­
biguate ordinary usage of the term. According to Livingston, an autor is
' a agent who intentionally makes an utterance, where the making of an
utterance is an action, a intended function of which is expression or
communication' .7
For an action to be an utterance it must be intended to express or com­
municate some attitude, be it a desire, belief, or intention. The fact that
we ae in control of and responsible for our utterances is enough to make
us authors of them, even if most of our utterances ae tte, formulaic, or
unimaginative. To be the author of a paicula utterance of ' Good Mor­
ing' , one need not have invented the phrase or the social function it
flfls. One says it on pupose, though not necessaly with deliberation,
and in so doing, exercises 'one's lingstic and social kow-how' .79 Author­
ship on Livingston's account is therefore just a fact of our being rational,
socia, and discursive creatures; it is pa of a 'prgmatic faework' that
'remains a deeply entenched, valuable, and arguably indispensable
schema of interaction. It is, moreover, a schema that we fequently apply
in discussions of the as' . 80
Even though Livingston does not explore this possibility for the case
of studio flm, his defnition allows for a single utterance having multiple
authors - the exaple he gives is of John and May jointly sending a
video letter to thei paeBts with holiday greetings and news.
For John
and Mar to be joint-authors requires tat they both intend to express
their own attitudes in the letter -for example, they ae not being coerced
into saying things they don' t mean, and though each of them may not
know exactly what the other will say during taping, they each kow tat
the other is committed to the same, shaed project. 82 This example also
seres to highlight the fact that an utterance need not be linguistic. There
ae many diferent kinds of expressive and comunicative actons, as
well as products of these actions - a flm, for example - that ae authord
because they a ' identifed with reference to the relevant features of
their context of production' . 83
Wereas Livingston gives us an entirly general account of the actual
author, which he himself later applies to flm, Nehamas gives us a spe­
cifcally literar account of the implied autor, which it is up to us to
apply to flm. Despite their many diferences, tese accounts share an
emphasis on action and intention. According to Nehamas, when we con­
struct an author for a litera work, we are thinking of te work as te
product of an action such that we mae sense of it by conceiving of its
possible motvation. The one who ' owns' the hypothesized action behind
the work is the author. The sense in which one owns an action is ver
diferent fom the sense in which one owns a pair of shoes or a ca: One
cannot be sepaated from one' s actions, since an agent and her agency
ae mutually defned. This helps explain why Nehamas constes the
author's ownership of the action resultng in a literay work in tens of
the author being pa of that work.
According to Nehamas, when we construe a text as a paricular work,
we create a context for that text that can be extended as far as interest
allows. If you a tying to fgure out why someone peroned a certan
action - say murdering his lover, you can stop at the imediate cause -
say, the murderer's discover of his lover's infdelity, or you can keep
going and fnd other more distat causes. Mter al, murder is not the only
way that one can respond to the discovery of infdelity. Wy did the jeal­
ous lover commit murder instead of just shouting and thowing things
or ending the relationship? In other words, why this paicula action?
I order to answer this question, you are going to have to think harder
about the jealous lover's other past experiences, general psychological
tendencies and expectations. This involves expanding the explanatory
context for his action. According to Nehamas, this is also what we do in
interpretation, understood in ters of expansion rather than in terms of
depth and hidden meaning. The one diference is that in interretation,
we cannot interogate the agent whose action we ae trying to under­
stand. Instead we imagine what the agent would have to be like to have
perormed the paicula action that results in the literary work.
Fortunately, our imagining can be guided by knowledge of the actual,
historical fgure who wrote the text that we are interreting. Nehamas
agues that the author can be considered 'a plausible historical vaant of
the wrter, a chaacter the wrter could have been, someone who means
what the writer could have meant, but never, in any sense, did mean' .
A wrter as an actal person cannot contol and i s not awae of all the
features of her writing. ' But the author, produced jointly by writer and
text, by work and
ritic, is not a person; it is a character who is ever­
tng the text shdts it to be and who in tu determines what the text
shows' .8 We a not, therefore, going to conceive of a
edieval text
wrtten by a medieval monk as having a modem author. We a going to
thin of the work in relation to other medieval works and construct an
author with a suitably medieval motivation. But this process is open­
ended in that we can keep comparng te work to other works and
complicating the explanatory context for the author' s activity.
We now have two very diferent conceptions of the author available to
us - Livingston's pragmatic conception of the author as any actual per­
son who perors an utterance, and Nehamas' s tanscendental conception
of the author as a constct generated by and tough literary interpreta­
tion. The next question to consider is wheter we can apply either or both
conceptions to flm. In the following secton, we will examine Livings­
ton's use of a modifed version of his conception to show that some
serially manufactured flms have single authors. Then we will consider
the way that auteur-stcturalist, Peter Wollen, constucts an author in
his interpretation of groups of flms by paicula diectors.
The obvious place to star in adapting Livingston's broad defnition of
authorship for flm is by replacing 'utterance' with ' cinematic utterance' .
Livingston suggests that ' an utterance is a cinematic one just in case the
agent or agents who produce it employ photographic (and other) means
in order to create an appaently moving image projected on a screen
(or other surface)' . 8
Now the question is whether there are any serially manufactured flms
which count as the product of a cinematic utterance such that they count
as authored. Livingston uses a series of real and imagined cases to show
that, on his account, there ae some serially maufactured flms which
clearly are not authored; there are others which ae authored to a cerain
degree; and there ae still others that are successflly completely
authored. These cases ultimately reveal two criteria for full authorship:
sufcient control over the fnal cut, since the fnal cut is the primary
means by which a cinematic author can intend to make some attitudes
manifest; and, having in mind some paicular attitudes for the fnal cut
to manifest. Thus the kind of control had by a flm author involves having
and being able to implement some knd of creative plan. Insofar as either
the plan or its implementation can be partly compromised, a flm can be
only patly authored. Although this may fequently occur under studio
conditions, fll authorship remains a very real possibility - as Livingston
demonstrates with the case of Ingmar Bergman' s Winter Light ( 1 962).
In the makng of Winter Light, Bergman was crucially involved at
every stage. He wrote his own script, did some of the casting, coached
the actors, supervised the editing and sound-mixing, and worked closely
with the cinematographer. He also exercised a high degree of contol
over the choice of location, props and make-up. Even those elements of
the flm that he did not help create himself - like the music by Bach - he
appropriates for his own particular ends. At no point in production was
Bergman coerced into changing his creative plan. To illustate the bal­
ance achieved by Bergman between control and collaboration, Livingston
compares him to the foreman of a constction project. Just as a foreman
discusses the building plan with his crew and then trusts them to cary
out specifc tasks under his direction, so Bergman aims to shae his artis­
tic vision with his collaborators in order to constrain and synchronize
their specialized roles. 8
In claiming sole authorship for Bergman, Livingston is not overlook­
ing the fact that many other talented people worked on Winter Light.
His point is just that all the other collaborators were working to make
Bergman 's flm, to realize Bergman's aristic vision. In other words,
being an artistic contrbutor to a project does not make one its co-author
unless one 'exercises decisive control over the creative process and taes
credit for the work' .
7 This distinction between a non-controlling artistic
contributor and an author is not one that everyone recognizes, however.
Gaut, in particula, considers anyone who maes 'a signifcant aistic
diference to the work' to have a share in its ownership. In painting, for
example, it is now standard to ascribe a work, not just to a master painter
like Rembrandt, but also to the master's workshop. But even though there
is collaboration in many of the arts, Gaut goes on to note that te aristic
efects of collaboration in flm tend to be fa more signifcant.
Compare the case of flm with that of architecture: When we say that
a building is by Frank Gehry, we mean that Gehry is its sole author even
though he has no hand in its actual constction. The builders ae not
Gehry' s co-a
thors because thei work involves simply realizing Gehry's
deSign. Whereas Gehry can specify the exact dimensions of his building
and the exact materals to be used in its constction, a flm director has
to leave it up to her collaborators to decide exactly how to cary out many
of th
irfsigned tasks. There ae simply too many possible dimensions
of vaation in how these tasks ae perormed; practically speang, a
single person could not issue directorial instructions precise enough to
accomodate every possible variation. The task of acting illustrates this
point nicely: No amount of directoral instction can contol exactly
how an actor says his lines and the exact nuances of emotion and mean­
ing that he brings to them. The director cannot simply manipulate
actors like puppets in order to realize her personal aristic vision through
them, and yet an actor's performance may contribute signifcantly to
the aesthetic signifcance of . a flm - take for example, Daniel Day
Lewis' s magnifcent performance in There Will be Blood (2008). This
suggests to Gaut that at least some flm actors are co-authors alongside
the director. 89
The same is not te for theate actors, however. Gaut points to a
fndamental ontological diference between flm, as an audio-visual
recording art, and theatre, as a literary ar, that explains why theatre
actors are not usually co-authors of te plays which they perfor. Since
a play is individuated by its scrpt, the playwright fully determines the
characters befor they are porayed by diferent actors. No doubt the
frst porayal of Othello by the famous Renassance actor, Richad
Burbage, was very diferent fom the acclaimed early-twentieth century
portrayal by Paul Robeson. But the fact remains that both Burbage and
Robeson ae poraying the sae character from Shaespeare' s play. By
contast, a chaacter in a flm cannot be flly determned by, say, a
screenwriter or a director before he is porayed by an actor on camera.
This is because a recorded perfonance is part of what makes a flm what
it is; namely a series of moving images, (ofen) with sound. Since a fl
cannot be perfoned but instead incorporates recorded perfonances,
and since flm actors give the perfonances that, when recorded, become
pat of the flm as an object of aesthetic appreciation, Gaut argues that
actors a one kind of flm collaborator among many that have a clam to
Whereas Gaut is willing to count as a flm's author anyone who plays
an aesthetically signifcant role in its production, Livingston only counts
as a flm' s author someone for whom the flm is a direct and personal
vehicle of expression. This means that, overall, Gaut will consider many
more serally manufactued flms tan Livingston to be authored. But he
will consider far fewer flms than Livingston to have actual, single
In the case of large-scale studio flm production, if not in the case of
small-scale independent, amateur, or expermental flm production, one
might be tempted to give up tying to identify an actual, single author for
a flm and instead settle for constructing a author when it helps with
interpretation. But can we do this? Can we intelligibly and usefully con­
stuct a single author for a highly collaborative flm in the way that
Nehamas suggests we constct a single author for a litera work? In
order to answer ths question, we will frst assess one fon that the con­
stuctivist strategy has taken in flm theory. Then we will consider some
general objectons to the constctivist stategy.
In the late 1 960s, a group of British flm theorsts under the infuence
of the French structural antopologist, Claude Levi-Strauss, developed
a new approach to flm interretation called auteur-stcturalism. On this
approach, the auteur is bot te product and locus of structural analysis.
On Levi-Strauss' s model, strctural analysis involves tacing the com­
plex cultural expression of varous binary opposites that stcture our
thinking. Applying this model to flm, auteur-stucturalists ted to iden­
tfy the basic binary motifs that chaacterize a body of work ad trace the
development of thei expression. The way that the auteur enters into this
approach is by representing the source of the stcture of thinng
expressed in te stcture of a cerain group of flms. According to one
leading auteur-strcturalist, Peter Wollen, the auteur is identifed with
the 'flm unconscious' or its expression in the binary thematic strcture
of the flm. Thus the basic oppositions between gaden and wilderess,
and nature and culture that characterze flms by John Ford ae attributed
to 'Ford' the auteur, the imagined mind, but not to the actual person,
Ford, since his relationship to the flm is mediated by the work of other
collaborators. 9
According to Wollen, the auteur-stucturalist approach reinforces
many of our common judgments about the work of diferent directors.
The richness and complexity of Ford' s flms at te level of stuctural
analysis explains their widely recognized critical imporance. In t,
tis suggests that the vaue of a flm like The Searchers ( 1 956) lies in its
potential for re¯stucturing culture and our thinkinge Thus Wollen con­
cludes, ' [a] valuable work, a powerfl work at least, is one which
chalenges codes, overthrows established ways of reading or looking, not
simply to establish new ones, but to compel a unending dialogue, not at
random but productively' . 91
We will ret to the question of what makes a flm good or bad for a
auteurst in the next section (and we will examine te nature of so-called
cinem�ic ' codes' in the next chapter). First we need to consider whether
the general strategy behind auteur-structuralism - the strategy of con­
stcting an auteur as par of flm interretation - is a good one. The
reason we will evaluate this general strategy and not the specifcs of
auteur-strcturalism is because, in the history of flm theor, auteur­
stctralism represents little more than a tempora brdge between two
opposed methodologies. As the flm autor was swept away by the tide
of postmoderism, auteur-stcturalism was abandoned before it was
flly developed.
Let us, therefore, retur to a more general question: Can we really
constuct a flm author? Nehamas suggests that to do so in literary inter­
pretation is an inevitable result of consting a text as a work ¯ the
product of an imagined action. But what of flm interpretation? Is it also
te case that we can make sense of a flm in terms of an imagined flm­
maer with a particular motivation? Once again, this brings us back to
te complex ad highly collaborative nature of flmmang. Even if you
do not know much about flmmang, the flm itself, with its diverse
combination of visual and acoustic elements, betrays something of the
enormity of the task of its creation. But if one has this sense of a flm's
collaborative orgins, perhaps it would be difcult and unhelpfl to at­
bute the fnal work to a single, imagined author. We would have to
imaginatively ascribe to this author responsibility for every aspect of a
flm fom its musical score to its lighting and cinematography, and even
all the perorances of the actors. But then, as Gaut suggests, we would
be workng with such an alien entity, one with superhuman abilities
such as te ability to control actors like puppets while simultaneously
5 1
planning every camera move, that we would not be able to understand the
author's motivation. According to Nehamas, the whole point of construct­
ing an author is to rationalize a work as te product of a particular action,
ascribable to a paicula kind of agent with a particula motivation.
The difculty in fully imagining a single, implied author for a serially
manufactured flm could explain the vagueness of Wollen' s description
of the auteur as ' an unconscious catalyst' after whom the defning struc­
ture of a group of flms is named. Indeed, it is difcult to see why Wollen
bothers wit the auteur at all if he is really interested in strctures and
codes. In Gaut's case, however, he remains sensitive to the way that flm
interpretation is guided by signs of creative intelligence in the work
itself. But he insists that these signs do not need to be signs of a single
author; they can be signs of multiple authors who we may imagine as
coordinating their creative activities mor or less successfully in a more
or less unifed flm.
Even if it is difcult, in the case of studio flm, to imagine a single
author, the method of interretation involving the constuction of authors
is still permitted by Gaut. Livingston, on the other hand, sees little to
recomend such a method: Either a flm has an actual author, in which
case there is no need to constrct one, or it does not have a actua author,
and then it would be misleading to imagine that it does.
Thus Livingston, unlike te constructivists, does not view the attribution
of authorship as an integral and essentia pa of flm interpretation. We
can only decide on a case-by-case basis whether a flm' s expressed atti­
tudes can be seen as te intended results of an actual author' s activities.
Of course this requires us to kow quite a bit about how a flm was made
and many ordinary viewers do not seek out this knd of inforation
before watching a flm. But Livingston insists that this does not mae
such information ay less valuable for interpretation. Any interretation
that we ofer of a flm has to match up with rlevant featres of that flm,
and some of those features will be tied to the flm's 'causal histor' .9
However, Livingston would be the frst to admit that only for certain
flms will these causal features be features of authorship. He does not
specify te diferences in how we interret flms with and without
autors. And although his example of Bergman' s Winter Light suggests
a connection between the value of a flm and its having a visionary
author, Livingston does not develop a theory of author-based crticism.
By contast, for traditional auteurists, this was prcisely the goal behind
the frst step of identifing a flm's actual author.
The idea developed by the critics of the Cahiers, paiculaly Francois
Trfaut, that tere is a certain knd of flm-maker who makes truly cin­
ematic and personally expressive flms, was impored to the United States
in the ealy 1 960s by the flm critic, Andrew Saris. Keeping 'auteur' as
the label for ths kind of director, Saris forulated explicit principles for
autor-based flm criticism. These were to be applied to Amercan flm,
which ever since the First World Wa had been almost entirely based in
Hollywood. Tied of the critical overshadowing of Hollywood ' enter­
tanment' by European ' a' flm, Sarris saw auteur teory as a way of
elevating the Amerca director and celebrating te rch tradition of
can flm. As a more ' scholarly' approach compaed to the ' subjec­
tive perls of. impressionistic and ideological' jouralistic criticism,
Saris also saw auteur theory as a way to institutionalize the study of flm
in North Amerca.93 While acknowledging the established European and
Asian auteurs, Sarrs recasts the Amercan director as a cowboy-fgure
engaged in a heroic creative stuggle against the oppressive machinery of
the Hollywood studio system. Notable 'cowboys' include Raoul Walsh,
Howad Hawks, and John Ford. The good diector, according to Sars,
makes good flms by winning his stuggle against the system. The critic's
task is to identify the signs of a successful stuggle across a diector's
body of work.
In paticular, the crtic looks for thee things: Signs of technical com­
petence on the pa of the director, signs of the director's personality,
ad signs of the creation of what Sarris calls, ' interior meaning' . Since a
Hollywood director is usually workng with someone else' s script, his
personality emerges, not in the subject-matter of his flms, but in his
treatment of the material. The inevitable tension that persists between a
flm's materal and directoral personality is the source of interior mean­
ing. Although interor meaning is supposed to be the hghest mark of a
flm's value, Sars admts that it is hard to say what it is exactly. This is
because interor meaning 'is imbedded in te stuf of the cinema and
cannot be rendered in noncinematic terms' . The most we can do is point
to examples of moments in particula flms when we ae brought to rec­
ognize interior meaning. The example Sars gives is, surprisingly, not
fom an American flm: It is a particular scene in Rules of the Game
( 1 939) when the chaacter Octave (played by the director, Jean Renoi)
pauses in his 'bearishly shambling jourey to the heroine's boudoi' .
Sar is remaks, ' [i]f I could describe the musical gace note of that
momentay suspension, and I can' t, I might be able to provide a more
precise defnition of the auteur theory' .
The more closely you study a body of work, compaing flms by the
same director, the more instaces of the expression of personality and
interior meaning you will fnd. In other words, you will notice greater
stylistic continuity across a body of work - say, through the repeated use·
of a pacula narative device, and you will gain a stonger sense of the
director' s particular outlook and sensibility. The 'joys of the auteur
theory' derive fom the surprising discoveres the critc can make in her
compaative analysis. Ultately, then, Saris seems to recommend his
critical approach as a way to get more out of watching studio flms. But
whether the approach has interal coherence ad intuitive plausibility is
a fer question.
It is another American critic, and one of Sarris' s contemporaries, who
pursues this question most vehemently. Pauline Kael maes no efort to
hide her derision for auteur theory, and sometimes sacrifces logical
argument for the sake of witty retort. Nevertheless, she raises some
important objections to Sar is' s evaluative criteria. 95 Technical compe­
tence, Kael argues, cannot be a prerequisite for good direction becaus

there are directors whose greatness derives in part from their inventive­
ness in the absence of technical competence - her exaple is the Italian
diector, Michelangelo Antonioni. As for the expression of personality
in the treatment of material, this is not just unnecessary for a good flm,
but might actually make for a bad flm. This is because we tend to notice
te personality of the director most in his worst flms, those that simply
re-use particulaly efective devices that characterize his style. By link­
ing unifority of style and the expression of personality, auteur theory
ends up celebrating, not atistic originaity, but the mastery of a set of
'tricks' for manipulating the audience. The result is that 'ofen the works
auteur critics call masterpieces ae ones that seem to reveal the contempt
of the director for the audience' .96
Finally, Kael considers interior meaing, the ultimate stamp of the
auteur, and decides that it is illUSOry. For one thing, it is rare to fnd the
knd of tension that is supposed to be productive of interior meaing
because any ' competent commercial director generally does the best he
can with what he's got to work with' . But even if there is such tension in
a flm, Kael asks, 'what kind of meaning could you draw out of it except
that the director's having a bad time wit lousy material or material he
doesn' t like?' . 97 Lookng for interior meaning is just the auteur critics'
excuse for repeatedly watchng mediocre flms, flms which only display
te requisite tension at the expense of a unity of for ad content.
Much of the force of Kael's atack on Sars derives from her convic­
tion that auteur theory violates the ethics of criticism. It does this, she
tins, by encouraging praise for the occasional bad flm by a favoured
director and censure of the occasional good flm by an unpredictable
director, as well as by regading any director who works with her own
material as suspect. Thus despite the vitrol, Kael' s analysis reminds us
of the danger that author-based criticism become merely a cult of person­
aty. From what we have seen in this chapter, this danger might be
avoided in at least two ways: By refering to a single, implied author
instead of the actual director, or by acknowledging the authorial role of
multiple collaborators. This is not to deny, however, tat a suitably qual­
ifed reference to the director of a flm could also guide interpretation and
Autor-based criticism is of course not te only form of flm evaluation.
I Chapter 5, we will consider other fors that involve only perpheral
reference to the flm-maker. The point of considering author-based criti­
cism here is that it involves cerain assumptions about the nature and
status of the flm autor. Despite disagreement over te acceptability of
these assumptions, paicipants in the flm authorship debate can agree
tat in the complex and highly collaborative production of serally manu­
factred flms, there is ofen a domnant collaborator ad tis is ofen the
director. It would be interesting to ty to ariculate what it is about the
role of a director and flm culte in genera tat ofen maes the director
assume dominace in the mnds of crtics and ordina viewers. But
te more pressing phlosophcal question concers whether a dominant
collaborator has creative ownership over a flm, and if so, what knd. As
we have leat, there are several ways of answering ths question: The
dominant collaborator on a paricula flm could be
1 . the sole author;
2. one author among many; or
3. not an author at all.
If we choose the third answer, this could be for one of two reasons:
i. with a multitude of collaborators, there can be no actal authors;
ii. the author is not an actual person but a interrtive construct.
Finally, even if we accept that the dominant collaborator on a particu­
lar flm is its actual author, we can still interpret the flm in terms of the
role of an implied author.
These ae the options for assigning authorship made available by our
discussion in this chapter. There a, however, furher considerations
concering new technology that might complicate these options. One
such consideration derives fom the introduction of DVDs.
As well as the flm itself, many DVDs include commentar by direc­
tors and other collaborators, as well as out-takes and even alterate
endings. A DVD may also give two or more versions of the same flm.
For example, in 2006, a two-DVD 'collector's edition' of Apocalypse
Now ( 1 979) was released, which includes both the orginal theatrcal
version of the flm as well as the 2001 re-released version, Apocalypse
Now Redu, with over 49 minutes of extra footage. The DVD set also
includes such items as a clip of Malon Brando reading T. S. Eliot's 'The
Hollow Men' , new production 'featurettes' , and cast member interiews.
It is interesting to consider whether the addition of all this 'supplemen­
tar material' makes it easier or more difcult to assign authorship.
On the one hand, the supplementary materal may help us to determine
whether the director' s or some other collaborator's aistic aims have
been realized in the fnal work. On the other hand, it may also draw our
attention to the high degree of contingency and circumstance governg
how this flm, like any other massive production, turs out. In comparing
the two versions of Apocalypse Now we could decide that one has an
(actual or implied) author and the other does not, or that they have difer­
ent (actual or implied) authors (which then raises the question of whether
the two versions really count as two versions of the same flm). One thing
is clear, however. We would not be in a position to make any knd of
decision about a complex case like Apocalypse Now on DVD without the
work we have already done in this chapter - without frst having care­
fully distinguished kinds of autorship, crtera for the assignment of
authorship ad criteria for author-based flm crticism.
The infuence of a literar paadigm is felt, not only in the debate about
flm authorship, but also in another debate that bridges classical and
psycho-semotic flm theor. This is the debate about whether there is
a language of flm and, consequently, whether we understand and appre­
ciate a flm in fundamentally the same way as we understand and appre­
ciatt-a literary work. Given that languages have to be leat by grasping
the c
nventions tat gover the meaning of words and the operations of
gramma, th

possibility that flm functions lie a language has serous
implications for its claim to reaism. As we leamt in Chapter 2, the real­
ism of flm is often explained in terms of our seemingly natural ability to
recognize on the flm screen what was flmed. But if we ae 'reading'
flm like a language, then this recognitional capacity is not natura, ater
al, but conventional, and the accessibility of flm images cannot be taken
as a sign of their realism. If we wish, therefore, to settle on a coherent
picture of the nature of flm, paiculaly in its traditional medium, we
must take up te debate about flm and language.
I te heyday of psycho-semotic flm teor, it was quite common for
scholas to talk about the ' grama' of the flm or te shot,
and even
today, it is comon for scholas to t about 'readig' a fl or flm
genre.99 In most of these instances, however, language ters ae only being
used metaphorically, and witout ay ontological commitments, to refer
to diferent modes of flm analysis and interpretation. This refects a
broader tendency in the ats to use language terms to suggest the serous­
ness and rigor of a particular mode of anaysis or the elevated status of a
certain art for or genre - as when art historians identify a mature style
of panting or achitecture in ters of its distinctive 'grama' . But in
relation to a in general, language terms are not only used metaphori­
cally. Nelson Goodman, for example, does not call his seminal work in
aesthetics, Lnguages of Ar, 1 0 just because it sounds serous and intrgu­
ing. Rater, he uses this title because his work contains a theor about the
language-like conventionality of aistic representation. Similaly, when
flm theorist Christian Metz uses language terms, it is not merely for
rhetorical fourish. I his case, it is because he is applying the general semi­
otics of the Swiss linguist Ferdinad de Saussure t flm. Since Saussure
prioritized language a the paadigmatic semiotic system, it is not surprs­
ing that Metz refers to ' cinematographic language,' and aims ' to study
the orderngs and fnctionings of the main signifying units used in the
cinematic message'
. 1 01
When language ters ae used literally in relation to flm, the philoso­
pher natrally pricks her ears. Now we have a genuinely philosophical
question to consider concering the nature of flm. The question of
whether there is a language of flm is not the question of whether flms
involve language - say with dialogue or text, but whether there is a
laguage of flm images. The images in a flm are the individual shots,
taditionally animated as a succession of frames passing before the pro­
jector beam, and then combined into sequences through editing. Thus
the question of whether there is a language of flm images is te question
of whether individua shots and teir combinations constitute a language.
In other words, it is either the question of whether individual shots fnc­
tion like words or sentences, or it is the question of whether the combina­
tion of individual shots occurs in the same way as the combination of
words and sentences. Furhermore, the question of whether shots fnc­
tion as words or sentences concers whether they acquire their literal or
intnsic meaning in the same way as words and sentences. This meas
that the starting question, whether there is a language of flm, tends to
reduce to the . question of whether we understand flm as a language.
When the focus is on narative flm, this question becomes even more
focused: Do we understand the stor in a flm in the same way that we
understand a stor in a novel, say, which is written out in a familia
Even though there may be no defnition of language, we can agree on
a range of uncontroversial examples like English, French and Arabic.
These and other natural languages ae the subject of linguistics, and the
aim for many flm theorists in defending the language-like nature of flm
is to qualify flm for the rigours of linguistic analysis. Metz sums up this
aim when he suggests that the precise, analytical methods of linguistics
'provide the semiotics of the cinema with a constant and precious aid in
establishing units that, though they are still approximate, are liable over
time (and, one hopes, through the work of many scholars) to become
progressively refned'
. 1 02
Given that the purpose of the flm-language
analogy is to qualify flm for linguistic analysis, the best way to test the
analogy is by compang flm to a language like English that already
qualifes for such analysis.
In an initial comparson, we immediately see that the ' laguage' of
flm images is more limted than English in its modes of tansmission
and representation. Wereas English is available through all our senses
- though sight in its written for, through heang in its spoken for,
tough touch in Braille, and at least in principle though taste ad smell
in coded fors, flm ' language' is only available to us though sight.
Furherore, in the wrtten for of English, the wide vaiety of fonts and
styles of handwriting suggest a degree of fexibility in exactly how we
represent a natural language. But with a flm image, there is no such fex­
ibility: Any modifcation to how the image looks changes its meaning. 1 03

'diferences raise interesting questions about whether flm ' lan­
guage' can be translated in te way that all natural languages can be
tanslated into one another (albeit with some loss in shades of meaning).
The Russian silent-flm-maer and theorist, Vsevolod Pudovkin, argued
that the task of the flm diector is to tanslate a flly written-out sce­
naio, word-by-word, into images tat ae then edited together just as te
words of
e scenaio ae combined into sentences and phases. 1
mght stre one as a highly idiosyncratic view of flm-mang, however.
If it does, the underlying rason may have to do with fher diferences
between language and flm that we shall discuss in tis chapter.
If you ae already suspicious of the flm-language analogy, you may
want to consider a weaker claim: Not that flm is a laguage, but that
flm belongs to a lager category which also includes languages. Ths is
the category of semiotic systems or systems of signs. Other tngs that
have been called semiotic systems include the natural languages, trafc
signs, ships' signalling systems, the gestures of Trappist monks, sema­
phore, conventions of dress or costume, and even myths. Semiotics is the
stdy of systems of signs and specifcally the codes that deterine the
meaning of signs. Film semiotics, in paicula, reduces the task of the
flm theorist and critic to one of ' de-coding' flms, thereby determining,
not just what every shot and sequence means, but how ever shot and
sequence meas what it means - by what cultural, artistic or cinematic
The questions whether flm is a system of signs to be de-coded and
whether flm is a language ae interwined in psycho-semiotic flm
teory. Neverheless, there are several concers specifc to the frst ques­
tion that we shall consider. These have to do with the implications of
using key terms like ' code' and ' sign' in equivocal and ambiguous ways,
as well as with the prospects for substantial theoretcal results in the
mere classifcation of conventions. Ultimately, however, our goal in this
chapter is to see whether our understanding of te nature and value of
flm a is deepened or merely obfscated by the comparison to language
and other sign-systems.
Even though the flm-language analogy is most flly developed in
psycho-semiotic flm theor, it is frst made much earlier, during the era
of silent flm, by those Soviet flm-maers and theorists, paticulaly
Pudovkin and Sergei Eisenstein, who take editing or montage to be the
defning feature of flm a. Both Pudovkn and Eisenstein ae interested
in how the meaning of a flm qua work of a, and not just the depictive
meaning of its individual shots, is created in context through the process
of montage. In order to emphasize the imporance of montage for creat­
ing meaning and aristic value, Pudovkin makes a compason between
flm and literature. In a literay work, he suggests, it is not the literal
meaning of individual words that matters aesthetically but instead how
the writer combines those words to create rch images and associations.
The example Pudovkn uses involves the initial selection of the word,
'beech,' which, on its own, 'is only the raw skeleton of a meaning, so to
speak, a concept witout essence or precision' . When this word is com­
bined in a phase like, ' ' 'the tender green of a young beech,'" it is no
longer ' merely a bare suggestion' . Rather, it 'has become pa of a def­
nite, literar form. The dead word has been waked to life though a' .
Pudovkn goes on to suggest that the individual words of a literay work
ae equivalent to the individual shots of a flm. Thus in te sae way,
what those shots literally mean in ters of what they show is not aes­
thetically relevant, and the accumulated literal meaning of the shots is
not equivalent to the meaning of the flm. Rather, the creative combina­
tion of shots through editing yields the only meaning that matters for
evaluating and interpreting a flm.
One problem with this account is that in both the case of literature
and the case of flm, even though context is import for interretation,
literal meaning still matters. Thus to use Pudovkin's own example, it
matters what the word 'beech' literally meas in the phase, 'the tender
green of a young beech' , even though it is only the phase as a whole
that gives us a image of this paicula beech tree. It would make a
diference to the connotative meaning of the line if, for exaple, the
poet had written ' willow' or ' bicycle' instead of 'beech' . Tis problem
refects the limitations of the exclusive emphasis on editing among Soviet
flm-makers. Whereas it is undoubtedly tue that editing adds meaning
ad vaue to a flm, it is also true that a well-composed and expressive
shot can have the same knd of meaning and value.
Another problem with Pudovkin's account is that words ad shots
resist compaison. As Metz will later point out, even a close-up shot of a
single objecr- say a revolver - canot be translated as the single word,
'revolver' , 'but at the very least, and without speakng of the connota­
tions, it signifes "Here is a revolver! ' ' ' . 1 0 Metz goes on to suggest that
te shot is not even equivaent to a sentence, since it contains 'a quatity
of undefned information' . All the detail contained in a shot of a land­
scape, say, could be expressed in a multitude of ways, in a multitude
of sentences. Thus, Metz concludes, a shot is only equivalent to ' the
complex statement of undefned length' . 1 07 This is also because a shot,
lie a statement and unlike a sentence, is always assered - as indicated
by Metz' s tanslation of the close-up shot of a revolver as 'Here is a
revolver' . At this point one might be tempted to give up on te analogy
between the shot and any part of laguage. We will retur to this possi­
bility a little furher aead. For now, it is enough to note that the word-shot
analogy on which Pudovkin relies is problematic.
Like Pudovkin, Eisenstein is interested in the way that editing creates
connotative meaning. But he seeks to avoid some of Pudovkin's difcul­
ties by using a diferent language analogy: Instead of comparng flm
with literature, Eisenstein compaes flm, and specifcally shot combina­
tions, with a knd of chaacter in Japanese wrtng, derived from Chinese
wrting, tat he calls an ideogram. This kind of chaacter refers to an
abstact idea by means of combining and modifing pictogaphic charac­
ters that depict, in a stylized way, non-abstact objects associated with
the idea. The Chinese character for 'brght' , for exaple, combines the
pictograms tat represent the sun and the moon, and the Chinese charac­
ter for ' good' combines the pictograms that represent a woman ad a
child. As Eisenstein describes the way such chaacters function, ' [b]y the
combination of two "depictables" is achieved te representation of some­
ting that is graphically undepictable' . Then he claims tat this exact
same process occurs in flm. In virtue of editing, shots 'that are depictive,
single in meaning, neutal in content' are combined 'into intellectual
contexts and series' .
Eisenstein's own flms were the testing ground for his theory of 'intel­
lectual montage' . In October: Ten Days that Shook the World ( 1 927), a
6 1
comemorative dramatization of the Bolshevik Revolution of October
1 91 7, there is a famous sequence that juxtaposes a Baroque image of
Jesus with images of Hindu deities, the Buddha, Aztec gods, and fnally
a prmitive idol, in order to suggest the saeness of all religions. The
idol is then compared with military regalia to suggest the linking of patr­
otism and religious ferour. In another sequence, shots of Alexander
Kerensky, head of the pre-revolutionary Provisional Goverment, a
interspersed with shots of a preening mechanical peacock, to suggest,
quite clearly, the leader's vanity. ad decadence.
The fact that these shot sequences are so distinctive immediately raises
doubts about the utility of Eisenstein's theory as a general account of te
meaning of a flm. Sometimes - perhaps most of the time, there is no
conceptual meaning generated by a certain combination of images. Wen
a fght scene in a Kung Fu movie is edited so that we see only the most
salient moments of action, the meaning of the sequence depends upon
the literal meang of the individual shots, each depicting pa of the
action. In other words, the ordering of shots may not have ay meaning
beyond the meaning that it has for the action. Moreover, this case is not
unusual, since we are often meant to pay attention, frst ad foremost, to
what is shown, rather tan what is merely suggested, on screen.
Motivated by a commitment to Bazinian realism, the next generation of
theorists interested in flm language ae highly critical of Eisenstein's
account. Metz, in particula, seems t t that the montagists over­
emphasized symbolic meanng in flm at the expense of depictive meaning.
As a result, they failed to recognize the natural tendency of the flm
medium towads narative -just by joinng meaningfl images together,
a flm tells a stor. Thus according to Metz, ' [g]oing fom one image to
two images, is to go fom image to language' . 1 0
Christian Metz has been called the most imporant flm theorist since
Bazin. Indeed, Alfred Guzzeti chaacterizes Metz' s importance in terms
of a dialogue between Metz and Bazin: With the ttle of his main work of
flm teory, Bazin asks, Wat is Cinema?, and Metz replies, 'cinema is a
language' . Even if Metz is not the frst flm theorist to tin of flm in
terms of language, Guzzetti insists that Metz is te frst theorist to infor
this way of thinking with a sophisticated understanding of linguistics,
particulaly te linguistics of Saussure.
1 10
In fact, however, Metz' s answer to Bain' s question is not so simple,
precisely because he was working within the technical famework of
Saussurian linguistics. On Metz' s account, flm is not just a language but
a language without a system. This qualifcation refects Saussure' s dis­
tinction between langue and parole, or language and language system.
Lngue is te system of rles and conventions or codes that make up a
laguage, independently of its use on paicula occasions. Parole is an
instance of language use. Saussure was most interested in langue, or the
laguage system, and he mght have said that every individual flm is the
parle of an underlying flm langue. But Metz, based on his close study
of flm, agues that there is no underlying system of which individual
flms represent paricula applications. There is just the flm language
being created as it is used in ever flm.
The language analogy holds for Metz insofar as flm is fundamen­
tally comunicative - each flm has a message for its viewers. But the
analogy only goes so fa because a flm's message is available to us
diectly and naturally, and not by means of convention, given our ability
to recognize what an image depicts. At some points in Metz' s account,
this seems to mean that the message of a particula flm is not heavily
coded. But Metz reteats from this claim in suggesting that the standad
combination and organization of te natural signs of flm images, par­
ticulaly in the creation of a narative, creates a cinematic code tat it is
the task of flm semioticiansl l l to analyse.
Metz is most interested in codes that are specifc to the flm medium,
which include conventions of flm punctuation like the fade in and out,
ad the dissolve, as well as conventions of montage for presenting the
depictive material. According to Metz, a flm is made up of basic units of
meaning that he calls ' syntagmas' . In his analysis of narative flm, Metz
ar ives at a taxonomy of eight diferent kinds of syntagma. A series of
shots that alterate between two events might constitute a ' alterate
syntagma' insofa as it shows two events occuring simultaneously.
Moreover, a series of shots showing a landscape might constitute a
'descrptive syntagma' insofa as it shows what the landscape is like
rather than events unfolding over time. Te reason that the syntagmas
constitte codes is that they deterne meaning conventionally, accord­
ing to standad practices of montage. Thus, for example, even if the
depicted content of individual shots in an alterating sequence is natu­
rally available to us, we have to lea tat the alterating sequence itself
indicates simultaneity.
Metz assumes tat the codes of flm punctuation ad montage concer
denotation in flm, or what the images in a flm literally show us. But
there ae also codes that gover connotation in flm, or the symbolic and
expressive meaning that a shot or a series of shots acquires. For exaple,
an American Gangster flm might give an impression of foreboding in
the w;ay that it presents a scene of deserted wharves. While it is important
to fg�re out the conventions govering the creation of such an impres­
sion, Metz gives priority to fguring out conventions govering denotation.
In fact, however, it is not clear that Metz' s distinction between denota­
tion and connotation in a flm can be upheld, since, as Gilber Haa
suggests, we fgure out the plot of a flm by referng both to what is
shown and the way it is shown in context.
1 1 2
But even if the distinction
could be upheld, another flm semiotician, Peter Wollen, argues that
Metz has his priorities wrong. It is connotative meaning and not denota­
tive meaing that is the proper subject of flm aesthetics and criticism.
More generally, Wollen argues that Metz' s aalysis of cinematic codes
and meaning is limited by Metz's adherence to the language analogy. For
Wollen, since language is simply one kind of sign system among many,
we must analyse flm, not in terms of language, but just as another sign
For this knd of analysis, Wollen relies on a theory of signs given by
the philosopher Charles Pierce who, along wit Saussure, is considered
to be a co-founder of semiotics. According to Pierce, a sign can function
in three ways - as an icon, as an index, and as a symbol. According to
Wollen, the flm image can combine all thee aspects of the sign. Insofa
as a flm image resembles its subject, it fnctions as an icon; insofar as a
flm image is causally related to its subject, it functions as an index; and,
insofa as a flm image bears connotative meaning as a result of context
and convention, it functions as a symbol. A great flm-maer is someone
who can manipulate all three aspects of the sign, so tat his flms have
'pictorial beauty' (due to the iconic fnction), ' documentary trth' (due
to the indexical fnction), and 'conceptua meaning' (due to the sym­
bolic function). Wollen cites French New Wave director, Jean-Luc
Godad, as an example of a flm-maker who is great in this respect.
I 1 3
By appealing to a general theory of signs, Wollen moves away from
the linguistic analogy. Pa of the reason that he is not tied to this analogy
in the same way as Metz is that Wollen does not assume that flm is
essentially communicative and that every flm cares a message. As
works of ar, flms explore the implications of signs rather than simply
using them to communicate. However, there ae still codes for the flm
semiotician to study, which a the means by which we interpret the
signs mang up a flm. According to Wollen, works of a 'exploit and
call attention to varous codes. The greatest works "interogate" their
own codes by pitting them against each other'
. 1 1 4
Despite their diferences, then, Metz and Wollen agree that flm theory
ad criticism is largely concered wit identifying, organizing and deci­
pherng cinematic codes. But what, exactly, is a cinematic code? In
ordinary language use, ' code' can mean either a cipher - in the sense of
de-coding a message, or a convention or style - in the sense of codes of
dess and mlitay behaviour. The difculty, according to Haan, is that
flm semioticias fail to mark this distinction in teir use of the term.
Thus, it is unclea whether Metz' s syntagma taxonomy is an exercise in
de-coding in the sense of explicating how a flm has a meaing or in the
sense of identifying strctura features of a flm. Simlaly, when Wollen
refers to the imporance of iconography in flm, and the codes of sym­
bolic meaning on which it depends, he seems to be tang about style,
expression, thematic materal, ad symbolism all at the same time. This
leads Haman to accuse flm semoticians of 'cheating' with thei broad
and loose use of the term ' code' . This usage, he argues, ' disguises the
fact that much of aesthetics and criticism is properly concered with
somethng other than the signifcance of signs'
1 1 5
Remember, as well, that the ultimate purpose of bot the language
analogy and broader semiotic aalysis in flm teory is to help us account
for the way that flms acquire meaning, paticulaly story meaning, and
thus how we understand them. But it is not clea tat the work of flm
semioticians so far has contbuted to an account of flm interpretation.
Tae, once again, Metz' s syntagma taxonomy: It may be satisfying to be
able to identif and label the varous standard relationships between
edited shots. But being able to identif when I understand an alterating
sequence as indicating simultaneity hadly explains why I understand the
sequence in ths way or what signifcance this has for my understanding
of the stor or flm as a whole. And, in genera, taonomies of cinematc
codes do not substitute for a theory of meaning or close contextal anal­
ysis of a pacula flm's narative and nar ation.
These concers about the efcacy and breadth of semiotic and language­
based theoretical approaches to flm ae shaed by the philosopher
Gregory Curie. His interest in understanding how language works in
tens of communication leads him to confn many of Metz's own wor­
res about the weakness of te aalogy between language and flm.
Ultimately, in showing just how diferent fl is fom language, Cure
quashes Metz's hope of using linguistics to understand flm. Currie goes
further, however, to claim tat, even if flm could be compared to a lan­
guage, this would not provide a method for interpreting the story of a
flm. This is because, in order to fgure out what paticula images mean
for the story, we have to pay attention to context, or how te images ft
with oter images as well as with diaogue and other sound cues.
In setting up a comparison between flm and language, Curie focuses
on fve features of a natural laguage like English tat ae ' salient in
tens of comunication,' ad the logica relations between these fea­
tres. English is both ( 1 ) productive and (2) conventional, and as a result,
it is (3) recursive, (4) molecular and (5) acontextual. As we shall lea,
flm is aso productive and many flm language advocates have insisted
on its conventionality. But flm is not recursive, molecula and acontex­
tual. Since these tree features simply follow from the combination of
productvity ad conventionaity, the fact tat flm does not have them is
a decisive indicator of just how diferent flm must be from language.
English is productive because an unlimited number of English sen­
tences ca be uttered and understood. We can use and understad
sentences we have never head before or that have never even been used
before. As well, English is conventional because te meaning of its words
and sentences is deterned by how we use its words ad sentences in
cerain standad ways in order to communicate wit each other. Whereas,
in principle, any word could have been used to designate the anima that
we call a horse, we al stick to calling this animal a horse so that we ca
understand one another. As a result of its conventionality, language has
to be leat - we have to lea what members of a cerain language
community happen to call things, since there is no natural and universa
way that words and sentences have to mean. Since language is produc­
tive, however, it cannot be the case that we lea English sentence by
sentence. (Otherwise how could we understad new sentences as soon as
we encounter them?) Instead, we must lea Englsh recursively: We
acquire a set of conventions that assign meanings to a fnite stock of
words and a set of rles for the combination of these words into an inf­
nite number of sentences. In tum, this means that English is molecular
- its sentences ae built up fom independently meaningfl units - words
or ' meaing atoms' -by rules that make the meaing of sentences depend
on the meaning of their pas. And fnally, since te meaning of a word is
deterined by convention and the meaning of a sentence is determned
by the meanings of the words in it, literal meaning in our language is
acontextual1 l 6 - so the word ' horse' is always going to refer to the same
kind of four-legged animal even if it acquires frher connotations in
specifc linguistic contexts.
The next step is to consider whether flm images possess these fve
interconnected, comunicative features of a language. Since there is an
unlimited number of things that can be conveyed by flm images, advo­
cates of flm language are going to want to claim tat flm 'language' is
productive. Tey also tend to emphasize its conventionality, tough some
more than others. Whereas Metz recognized that the depictive meaning
of individual shots is natural and only insisted on denotative conventions
for narative meaning, another leading semiotician, Umberto Eco, insists
that the conventions go all the way down, so to speak, determining the
literal meaning of a single shot. 1 1 7 This is because even the most realistic
and characteristic shot of an object does not reproduce the object exactly
- thee-dimensionality is lost, for example, on the screen. Therefore, for
the viewer to recognize the object in the shot requires her to already
know which features of an object ae salient for its representation. And,
according to Eco, we can only know this according to some kind of inter­
nalized convention. At this point, however, a problem aises. If we insist
on the conventionality of flm language along with its productivity, we
a automatically committed to claiming tat flm language is recursive,
molecular and acontextual. Unfortunately, this claim is false, which
means that flm images just don' t work like a language.
There ae no atoms of meaning for flm images, since it is not the
case tat we understand a flm image by understanding its pats and the
rles according to which they are combined. A part of an image - say,
a unifory sandy pat of an image of the Sahara in Lwrence of Arabia
( 1 962) ¯ just has meaning as a pa and not independently of the whole.
Although one could take parts of several images and combine them to
make a new image, we surely do not want to say that every image is a
collage or a 'composite in this way. Moreover, the pars of a composite
image would still be pas of the original images fom which they were
extracted in the sense that they would not convey anything new in their
new context - if we included the sandy image-par fom Lwrnce
of Arabia in our composite image, that part of the new image would
still just show sand. Thus image-parts are clearly not meaning atoms.
But without meaning atoms, there is nothing to build up, according
to rles, into a whole that is therefore meaningful both recursively and
Insofar as we understand flm images, it cannot be a result of our hav­
ing acquired a lexicon of image-pars and having mastered rles for their
grammatica combination. Wile we may not understand the dialogue i
a Korean or a Trkish flm without subtitles, we have no touble making
out the image track (assumng the images are in-focus and properly lit)
even if what we make out includes unrecognizable objects and unfaml­
ia landscapes. Wat this suggests is that meaning in flm, despite the
standadization of flm practice, is not conventional ater all. For how
could flm language be productive, such that we recognze images we
have never seen before, and still be conventional, given that we do not
recognize new images as the result of their being built out of indepen­
dently meaningful units according to rules? Curie agues that the
productivity of flm images has to be explained, not in terms of conven­
tionality, but in ters of natural generativity. This notion is connected to
Currie' s agument for flm realism that we discussed in Chapter 2. Just as
with other knds of picture, flm images ae understood insofa as we
recognize what they depict in the same way as we would in rea life.
Thus we ca understand new images insofar as we automatically and
naturally recognze everything they show us.
Even if te meaning of flm images is not determined conventionally,
it can still be infuenced by convention. After all, a flmed scene will look
the way it does as a result of various conventions of dess, decor, and
decorum. Objects within the flmed scene may have symbolic meaning
- for instance, a crucifx, or a conventionally determined fction - for
instance, money. Butnone of these conventional infuences mean that the
image itself - like the word, 'horse,' is conventional. The problem, Cu­
rie suggests, with much of the semiotic theorizing about flm, is that its
conclusions about flm language depend on glossing over this distinction
and using the term 'convention' just as loosely and abiguously as the
term ' code' .
Given that flm images lack important communicative featues of lan­
guage, Curie concludes that there is no language of flm. This conclusion
seems to foreclose the possibility of using linguistic analysis to solve te
problem of flm interretation. However, it may be that this was never a
real possibility in the frst place. For, as Currie then goes on to suggest,
even if there were a language of flm, this would not wholly explain how
we understand te story told by a flm. Even though both words and images
may have an acontextal, literal meaning, te forer conventionally and
the later naturally, when either words or images are combined to mae a
naratve, context suddenly becomes highly important.
This is analogous to the importance of context in determining the
meaning of utterances as opposed to the litera meaning of the words
uttered. When you exclaim, 'I can' t believe that this is happening to me! '
or, 'My life sucks ! ' , 1 have to know more than the literal meaning of the
words you utter in order to know what you' re talng about. 1 have to
understand both the literal meaning of the words and the context in which
you utter them - so, for example, I have to know what 'this' is that is
happening to you or how your life can ' suck' but not in any literal way.
Aruably, this is still a matter of literal meaning, but then we need a dis­
tinction between the literal meaning of the utterance, which depends on
context, and the literal meaning of the stng of words uttered, which
does not. In literary works, context is even more imporant for under­
standing. For as well as understanding the context for particular utter­
ances made by the chaacters or the narator, we have to understand the
narative context of paricular events so that we can fgure out what is
happening to the characters. Say we have a literary descrption of the
heroine" sipping some soup prepaed by her jealous cousin immediately
followed by a description of the heroine falling ill. We have to fgure out
whether te soup is te cause of her illness by making reference to what
we already know about the story, especially the chaacters' motives.
The imporance of context for narative understanding is no less with
flm than with literature. If a shot of the heroine sipping her cousin's soup
were immediately followed by a shot or sequence of the heroine falling
ill, we would also have to fgure out the relation between the depicted
events by way of context. Recognizing what is depicted in the shots is
only par of the process by which I come to understand the stor. Thus
a teory about how I come to recognize what is depicted in the shots
will provide at best a partial explanation of narative interpretation. In
Chapter 6, we will examine more fully how flm interpretation works.
For now, it is enough to note that it is not just that there ae problems with
the analogy between flm and language; there are also problems with the
motivation for this analogy which is a solution to the problem of inter­
There may still be one fnal way to save the flm-language analogy,
however. What about the claim, not that flm images themselves have a
language-like structure, but that flm images combine with one another
in language-like ways? Is the meaning of certain stadad shot combina­
tions determined conventionally in just the same way as wit sentential
connectives? Take, as an example, the sentential connective, 'because' .
When we join together two sentences, P and Q, with 'because,' there i s a
convention detenning a literal meaning - that P is in some way a result
of Q. But there is no anaogous convention for detenning that the event
depicted in one shot is caused by the event depicted in the preceding
shot. Rather, as with the example of the poisoned soup, we have to infer
from the nar ative context whether a particula paing of shots indicates
a causal relationship.
However, you might reply that even if there is no convention for sig­
nalling causation in a flm, there are other conventions of editing that
fnction like conventions detenning the meaning of sentential connec­
tives. Take, for example, shotreverse-shot editing that combines a
face-on shot of a chaacter with a second shot from that chaacter' s point
of view.
l i S
Given the frequent and standard use of this combination, is
there a convention detenining that any shot folowing a face-on shot of
a chaacter is fom that chaacter' s point of view? No. In many instances,
face-on chaacter shots ae not followed by point-of-view shots. But
how, then, do we just seem to know when a shot following a face-on
character shot is from that chaacter's point of view?
It helps that shot/reverse-shot editing is used fequently in manstream
flm. But, ultimately, we have to be paying attention to contextual cues.
Perhaps, for example, te shot/reverse-shot combination is preceded by
a medium-long shot of the chaacter facing a certain scene, so that after
the face-on shot, we know that the next shot showing the same scene
fom the chaacter's perspective is a point-of-view shot. Or perhaps the
shotreverse-shot combination is pa of a sequence that depicts two
characters in conversaton such that it is the dialogue that prmarily cues
the viewer to recognize a point of view. Whatever the particular contex­
tal cues may be, the point is that we rely on such cues to understand the
literal signifcance of an edited sequence in a way that we do not to
understand the literal signifcance of sententia connectives.
Once again, therefore, flm fails to exhibit a key feature of language
- in this case, the acontextuality of the literal meaning of words. The
example of shotreverse-shot editing does not, however, suggest the com­
plete absence of conventions govering the meaning of shot combinations
in flm. Metz' s syntagmas and cases of flm punctuation, like the fade in
and out to indicate a signifcant passage of time, may still involve con­
ventions. But even if there are a few conventions in flm, they cannot
fnction without individual shots aleady having a literal, natral mean­
ing. Given the natura generativity of flm images, we just don' t need as
many conventions to detenine meaning in flm except at the level of the
ordering of images. At this level, conventions help constrain the interpre­
tation of the flm, which, Cure suggests, is the only way in which flm
conventions resemble language conventions. 1 I 9 But interpretation is a
subject for aother chapter.
In ths chapter, we have examined the dangerous move from a meta­
phorical to a literal use of language tens in relation to flm. Here is what
we have covered:
1 . Pudovkn' s analogy between the shot and te word in tens of con­
notative meaning.
2. Eisenstein's analogy between a shot combination and an ideogra in
tens of conceptal meaning.
3. Metz, Wollen and Eco' s reduction of flm analysis to a mysterious
process of de-coding.
4. Cure' s agument to show tat flm lacks the essential and intercon­
nected comunicative features of a natural language.
5. The furter claim that, even if flm were a language, this would not be
enough to explan interpretation.
At this point, we might just content ourselves with saying that flm is
lik a language in several ways - for exaple, in the way that some con­
ventions play a role in constaining interretation or in the way that flm
images ae productive. But flm is probably like a lot of other thngs in
this same kind of loose and paial way. Ultimately, then, this kind of
likeness is not enough to give us insight into the nature of flm art.
7 1
As you may have noticed, the knd of flm which has received the most
attention from philosophers is nar ative fction flm rather than, say,
expermental non-narative flm or documentary. This is not surrising,
perhaps, considerng that most of the flms most of us watch are narative
fction flms. In the simplest terms, a narative flm is a flm that tells a
stor and a narative fction flm is a flm that tells a made-up story. This
of course introduces the question of what counts as a made-up story and
why, since not everything in a made-up story is made-up. But leaving
this question aside, given that our interest is in a paticula narative ar
form, we need to think about what it really means to use moving images
and sound to represent something that we then recognize as a story. The
leading question for this chapter is, How does a flm tell a story?
It is worth noting that in the history of flm theory, this question has
not been paticularly central. The study of narrative is a well-established
branch of literary theory, and as we shall see, some literary theorsts
extend their accounts of nar ative to flm. This indicates that the literary
paadigm which infuences discussions of flm authorship and flm lan­
guage also infuences discussions of flm naration. In additon, some
cognitive flm theorsts ae interested in the psychological process of
story comprehension. But the conceptual issues raised by the question
of flm naration have been of independent interest to philosophers. Thus
the approach in this chapter is somewhat diferent than in previous
chapters: We wl not begin by trying to make sense of the flm theor
that deals with a particula philosophical issue but instead we will simply
jump right to the curent philosophical debate about paicular aspects
of narative flm.
There are many other kinds of narative at besides flm. There are
novels, plays, epic poetry, ballet, opera, program music, comic strps
and narative painting and sculpture. Moreover, narative is limited nei­
ther to ar nor to fction, since there are naratives of historcal events ad
scientifc discover. On a minimal defnition, a narative is a chain of
events, usually involving intentional action, which ae ordered either
temporaly or causally. On a fuller defnition, a nar ative is also purpose­
flly designed by someone to have a certan stucture - a beginning, a
mddle, and an end - or a certan point - for example, entertainment. 1 2
Note that the defnition of narative maes no reference either to
medium or to content. As long as an aistic medium can represent a
chain of events, it is a narative medium. And regadless of what these
events ae, as long as they are ordered in a certain way, they are pat of a
nar ative. This explans why the same story can be told in many fors -
for example, the story of Dido and Aeneas told in an epic poem by Virgil,
in a play by Chstopher Malowe and in an opera by Purcell; or the
adaptation of Jae Austen' s Pride and Prejudice for a television series
ad for the Bollywood flm, Bride and Prejudice (2004).
Even though narative is independent of paricular aistic media, it
seems likely tat tere ae diferences in nar ative comunication across
media. In other words, the story of Dido and Aeneas is going to come
over a little diferently as an opera and as a epic poem. There is thus an
interesting question about the distinctive featres of nar ation in flm,
whether in the traditional medium of flm or in video or digital media.
In order to come to an understanding of flm nar ation, we will con­
sider the following three questions:
1 . Must nar ative flms always have narators?
2. Wat is it for a flm's naration to be unreliable?
3. How do flms support narative comprehension?
These ae the main questions about flm nar ation that concer phi­
losophers and cognitive flm theorists. The frst two questions ae often
motivated by an interest in te possibility of stuctural similarities
between naration in flm and in literatre. Thus the frst question is
often taen as shorthand for a more focused question: Given that literar
fctions always or almost always have narators, must fction flms, too,
always have narators? Similarly, the second question is ofen under­
stood as a question about whether te same fors of narative uneli­
ability are found in flm and in literature. Even when theorsts approach
the frst two questions with reference to literature, however, they end up
drawing attention to the remarkable vaety of stategies available to a
flm-maer for keeping the audience interested in how the stor goes.
The third question ten reminds us that whether a flm' s nar ative strat­
egy is efective greatly depends on the viewer.
To a large extent, we will consider the second and third questions
in relation to the frst. Our aswer to the frst question will be that nar a­
tive flms may have but need not have nar ators. Given ths answer, and
given that nar ative unreliability is ofen accounted for in terms of the
activity of a narator, pa of what will concer us in tang up the second
question is whether there can be uneliable naration without a nar ator.
Wen it comes to the third question, we will focus on how we should
divide up the work involved in following a stor between the nar ator, if
there is one, and the viewer. A broader and more detaled analysis of
the activity of narative comprehension will be reserved for the next
chapter, since nar ative comprehension is one aspect of our cognitive­
perceptual engagement with flm. Overall, most of our attention in this
chapter will be focused on the elusive fgure of te cinematic narator.
In general terms, a nar ator is a fctional agent who tells a story, be it
in a flm, a novel, or a play. The narrator is pa of the work of fction, and
most commonly, she reports or presents the story events that unfold in
the work as though they actualy happened. Sometimes, however, a na­
rator tells a fctional stor, in which case, since the nar ator is still pat of
the work, it is fctional that she is actually tellng us a fction. In either
case, however, the task of the narator, fctionally at least, is to report or
present story events to the audience.
One motivation for aguing that flms and other kinds of nar ative
aworks always have narators is the observation that a story is always
told in a cerain way. This accounts for te tone of a work, or the set of
attitudes manifest in the way that chaacters ad events ae descrbed or
depicted. The ton

of a flm is the result of a wide varety of stylistic
choices concering lighting, cinematography, mise en scene, and editing
of both the image and sound tacks. Wen we pick up on the attitudes
manifest in a flm's style, we naturally want to assign these attitudes to
someone. And if we want to assign them to someone inside the fction,
we assign them to a narator.
There are in fact several knds of narator in flm ad when theorists
argue about the necessity of a nar ator, they ae aguing about one pa­
ticular knd. This knd of nar ator is usually referred to as the 'cinematic
nar ator' and can be distinguished from chaacter nar ators and voice­
over nar ators. As an audio-visual medium, flm is lie theatre and unlie
literature in being able both to tell and to show a story. 1 21 This means that
i flm tere ca be both verbal and visual nar ators. A verbal nar ator is
most commonly a voice-over nar ator who intoduces and explains past
events in the fctional world of the flm as we ae fctionally shown those
events on screen. This kind of narator can be one of te characters, as
in Fight Club ( 1 999) or Sunset Boulevard ( 1 950), or a impersonal and
omscient fgure as in Ma
nolia ( 1 999) or The Nakd Cit ( 1 948).
In flms that use voice-over nar ation, it is often the case that we are
simultaneously 'shown what is described by te verbal nar ator. But in a
few more unusual cases, verbal naration ad visua nar ation come
apat. Wat the flm shows us may contadict or complicate te story we
ae being told in voice-over. We are forced to decide which version of the
story to tst - the one we ae shown or the one we ae told, tus indicat­
ing a for of narative unreliability unique to two-track media lie fl
media. We will examne this and oter forms of nar ative unreliability in
detal frther ahead. Before we do, however, it is imporant to note the
way in which the coming-apa of visual and verba naration mght sere
to revea te cinematic nar ator. When we ae shown someting oter
tan what we are told by a flm, it is as though someone else besides the
voice-over nar ator is in control of the image track. But who? Some
teorists say it must be a implicit and exclusively visual nar ator. In
other words, someone whose presence is not explicitly signaled in the
flm but rather implied by the way tat stor events ae shown.
This someone is the cinematc nar ator, the fctional agent tat certain
theorsts tae to be part of the very stcture and fabric of the narative
process. The most comon conception of te cinematic nar ator, ad the
conception that is assumed in most aguments against te cinematic
nar ator, is of te narator as a guide to the events depicted on screen.
Oter more contoversial conceptions include the narator as witnessl 22
ad the nar ator as a 'image-maker' .
1 23 The diference between the con­
ception of the nar ator as a guide ad the conception of the nar ator as an
image-maker is primarily a diference in what te narator is taen to be
presenting, whether the fctional events themselves or recorded images
of those fctional events (instead of recorded images of the actual, staged
events representing te fctiona events). 1
24 The appea of the conception
of te nar ator as guide is that it accommodates the way in which the
fame of the shot seems to be used to point to the depicted events. How­
ever, some teorsts question whether it is really a narator, ad not just
a implied author, who is doing the pointng.
I Chapter 3, we leat tat an implied author is the person to whom
we imaginatively attbute a work of at in order to make sense of that
work in interretation. A sufciently unifed flm that was in fact made
by a diverse group of atists and aisans could nevertheless be assigned
a single, implied author in interpretation. It is imporant to distinguish
between the implied autor and the nar ator since they stand in diferent
relations to the story. The implied author is the agent to whom we att­
bute the whole work which happens to tell a fctional stor. The narator
is pat of this work and connected to the story in a particula way. The
events of the story are always fctional for the implied author whereas
they are usually real for the narator.
If we ae to conceive of the cinematic narator as a guide, we need to
ask whether his role can be distinguished from that of the implied autor.
The literay theorist Seymour Chatman is one of the leading defenders of
the necessary role of a guiding narator in narative flm. l
5 He assumes
that when we watch a flm we imagine a narrator presenting the fctional
events that ae shown in the implied author' s selected images. Further
ahead, we shall see that other theorists question Chatman' s assumption
that selection and presentation are cared out by diferent agents. But
frst we need to consider Chatman's agument for the necessity of the
cinematic nar ator which has become standard in the literatre. This
argument has been called the a prori agument1
because it is presented
as simply elucidating how we use the concepts of naration and the na­
rator without relying on any empirical assumptions about the narative
arts. The basic idea is that there cannot be nar ation without a narator
because the one is conceptualy dependent on the other. Here' s one way
the agument could go:
1. Narration is the activity of telling or showing a story.
2. Every activity has a agent.
3. The agent of �� narration of a fctional story is the narator.
4. al narrative fctions, including all nar ative fction flms, have
The most common objection to te a prior agument is that it doesn' t
prove what it aims to prove because te thid premise involves an unwa­
ranted assumption. It follows from the frst and second premises of the
argument that there must be a narating agent. But this agent cannot be
fctional, as the third premise clams, because the naration is a real thing:
A nar ative flm, for example, is an actual story-telling and -showing.
Te people responsible for making the flm ae the ones who are neces­
say for there to be an activity of naration, even if it is the naration of
fctional events.
Another objection to the a priori agument concers te assumption
behind the second premise, that every activity has an agent by defnition.
Is this really true? Sometimes when we t about an activity, we do not
mean to imply the necessary prsence of an agent - as when an astrono­
mer remaks, 'there's a lot of activity in the sky tonight' . Why couldn' t
nar ation be this knd of agent-free activity? Clealy, the defender of the
a priori argument has her work cut out for her.
The a pror argument is by no means the only argument for the neces­
sity of the cinematic narator. There is also an agument concering
how we come to know about the fctional world of a flm, or more spe­
cifcally, who is responsible for giving us a fctional view into this
world. 1 27
is agument begins wit the assumption that it is reasonable
for the flm viewer to ask about te means of perceptual access to a flm's
world. By way of an answer, the argument then claims that whoever
gives us' access to the fctional world of the flm has to be part of that
world. This is because only someone on the same ontological level as the
chaacters and events of the story could present those chaacters and
events as real . 1 28 Whereas the flm-maer (or implied author) presents
the story as a fction, the narator on the inside presents the events and
characters of the story just as they are - in a novel ts is achieved with
te use of declarative sentences and in a flm this is achieved just by
showing te represented events and characters directly on screen. Thus
in answer to the viewer's starting question about means of access, the
argument purpors to show that, since it cannot be the flm-maker that
provides access, it has to be the cinematic nar ator.
In responding to this agument, we need to ask two questions: ( 1 ) Is it
really that reasonable to expect an answer to the question of who is
giving us a view into the fctional world of the flm? And even if it is
reasonable, (2) is it really the case that positing a nar ator solves the
problem of access? In answer to the frst question, it might be unreason­
able to ask about means of access because it is one of the conventions of
fction that we overlook inconsistent or paradoxical implications of the
origins of the fctional naraton. Take the example of te flm, American
Beaut ( 1999), which begins with voice-over naration by a dead charac­
ter. If we felt compelled, in watching this flm, to fgure out how we
ae being told about the fctional events of the story, we would no doubt
fnd ourselves confused or distacted. Simlaly, if we stopped to fgure
out how we ae being shown Charles Foster Kane alone on his deatbed
in the dramatic opening scene of Citizen Kane ( 1 941 ), we would aso get
nowhere. The point is tat we don' t have to stat fguring this out and, in
fact, many works of fction require us not to in order to engage properly
with the story. How we come to be told about or shown fctional events
simply remains indeterminate in the fction.
In answer to the second question about whether positing a narator
solves the problem of access, it is not clea how a fctiona agent can
present te fctiona world of a flm to a rea audience. The agument
suggests that the flm-maer cannot give us access to the fctiona world
of the flm because he is on a diferent ontological level fom the flm
characters and events. But te narator is on a diferent ontological level
fom us. So the same difculty faces the narator as the flm-maker: How
is the nar ator supposed to brdge the ontological gap, assuming there
is one, between the real world and the fctional world of the flm in order
to provide access?1 29
The question of the relation between the real and the fctional is also
raised in an independent argument against the necessity of te cinematic
narator as guide. One way of constring this argument is as a chalenge
to a crucial assumption behind the a priori agument. Remember that
te a pror agument claims tat there must be an agent caring out te
activity of naration. The reason that this agent must be the narator
is because the events being nar ated ae fctional. Tus te assumption is
that only a fctional agent can show us fctional events. The argument
against fctional showing, as we shall call it, challenges this assumption:
Just because you have the naration of fctiona events doesn' t mean that
you have to have a fctional naration of those events by a fctional agent
(the nar ator).
1 30
While it is tre that only a fctional agent can show us the story events
as actual, this does not rle out the possibility of the real author showing
us the stor events as fctional. What is more, the predominance of
third-person, omniscient nar ative style in varous atistic media and
genres might show how this possibility is realized. When someone tells
a bedtime story to a child or repeats a ghost stor round the campfre, he
is the only one nar ating - unless, of course, he is telling a story of te
telling of story events. Similarly, when I decide to show you a fctiona
story in pictures, comic-stip style, unless I draw the pictures so as to
represent, say, the eye-witess record of a fctional agent, I a the only
one doing the showing. Nicholas Wolsterstorf makes the point in the
following way.
Presumably not al huma nar ation consists of narating what
someone narated. Why then must fctional nar ation have tis
stctre? Why can' t te novelist just staightforardly tell us a tale
of love �d deat, birt and wa, jeaousy and endeavour? Why must
his tale aways be the tale of a narrtion of love and death, birt and
wa, jealousy and endeavour?1 31
This may confse the issue, however. Someone who believes that
every work of narative fction is narated need not hold that the nara­
ton forms a part of or the subject of the story being narated. The
nar ation is not norally about the naration. An additional wory is
tat the argument against fctiona showing canot account for the way
in which a work of fction almost always presents te fctional as acta.
Wereas te author can present the fctiona as fctional - as te agu­
ment suggests - only a fctional agent lie the nar ator can present the
fctional ' as actual. This is an important point and much contested by
philosophers who ae interested in te natre of fction.
32 The debate
about the nature of fction is complex and separate fom our principa
concers in this chapter. But it is important to note that the point just
raised could suggest a qualifcation to the scope of the agument against
fctional showing.
At the end of the day, although there is disagreement about the neces­
sit of te cinematic narator, hardly ayone denies the possibilit of the
cinematic narator. This is parly because some flms give us a stong
sense of being guided in our view of fctional events by an invisible and
efaced agent. Sometimes a flm may deliberately signal the presence
of such an agent by breaking with cerain narative conventions. For
example, there is a scene in Last Year at Marienbad ( 1 961 ), where te
voice-over narator instucts someone to show us a diferent location
for the event being related. Who could it be that the voice-over narator
instructs? must be someone inside the fction who is charged with
showing us what happens, this sounds a lot like te cinematic nar ator.
Examples lie this show us that we need to thin about when it is usefl
for our making sense of flm naration to posit a cinematic narator. This
will depend on the paicula naration and whether it is constcted so as
to iply the mediation of an implicit visual guide.
Another reason why it is imporant to decide whether flms must always
have narators is so that we can understand narrative unreliability in flm.
If, when watching a flm, we have the sense that we're not being told the
whole or the tre story of events, say by a chaacter narator, then we are
most likely identifying narative unreliability. The standard defnition of
narative unreliability in literay theory invokes a narrator who intends to
have us believe something other than what the (implied) author intends
to be te in the fction. Given the discussion we have just had about
whether every flm has a narator, the interesting question is whether you
can have narative unreliability without a nar ator. If you can, then the
possibility of narative unreliability cannot be assumed to suppor any
necessay claims about the narator.
Even if it is te that in literatre unreliable nar ation is always the
work of an unreliable narator, this does not automatically account for
the case of flm. This is because flm, in its capacity both to tell and to
show a story, can be unreliable in unique and distinctive ways. To-track
unreliability can take diferent fors, depending on how a paricula flm
sets up a tension between verbal and visual naration. In more unusual
cases, even though verbal naration and visual nar ation are uniforly
consistent, a flm may exhibit a subtle form of global unreliability. This
kind of unreliability is revealed by ambiguity in the interpretation of a
flm and can be described without reference to a narator.
Two-track Unrel i abi l ity
To-track unreliability usually involves a voice-over nar ator telling us a
diferent story fom the one being shown. It depends on the flm as to
whether it is the verbal or visual version of te story that is the tue (or
'ter' ) version. But it is most common for the visual version to be true,
most likely because we have a tendency to trst what we see over what
we are told. Films narated in voice-over by a child or a slow-witted
adult like Forest Gump will ofen show more than te verbal narator
understands to tell. More unusual cases where we are supposed to trust
what we are told over what we ae shown include flms which blur or
obscure our fctional view of events for paicula emotional efect and
then help us follow those events with voice-over explanation.
Sometimes verbal and visual naration only come apat at certain
points in a flm or part-way though a flm. This shift serves to draw
atention to the possibility of unreliable naration and to prompt the
viewer to pay careful attention to how the story unfolds. Although most
commonly the point at which visual and verbal naration come apa is
te point at which unreliability begins, more unusually, it can also be the
point at �hich pror unreliability is revealed and comes to an end. This
involves the visual naration, up to a certain point in the flm, having
coroborated an unreliable verbal naration, or vice-versa. Recent flms
tat employ this stategy include A Beautil Mind (2001 ) and Sith
Sense ( 1 999). But the most famous example is Sta
e Fright ( 1 950).
In this flm, Johnny is suspected by the police of killing his lover's
husband. His fend Eve ofers to hide him and Johnny explains to her
tat his lover is the real murderer. Eve decides to investigate for herself,
however, and soon leas that Johnny has been lying to her. But when
Johnny is telling Eve his version of the murder, the image-track corro­
borates by showing his lover commtting the murder. This makes it a
suprse to lea along with Eve that Johnny was lying. It is only when
Eve starts investigating for herself that we lea that the naration of the
murder fashback was unreliable. Interestingly, the flm received a great
deal of crticism for its use of unreliable naration. This was because the
flm brngs the audience to tust Johnny' s version of events by providing
a visual exposition of his verbal naration. There is something paicu­
laly unsetling about discovering that one has been deceived by what
one sees as well as by what one is told.
Gl obal Unrel i abi l it
Global unreliability is found only in flms with a particula kind of com­
plex narative stctre. It is not based on a tension between showing
ad telling, or on a maked shift away from a faithful showing of a cha­
acter' s voice-over naration. Instead, this kind of unreliability chaacter­
izes the entire nar ative process of a flm and only becomes evident
tough questions of story interpretation. George Wilson claims that You
Only Live Once ( 1 937) exhibits global unreliability. In the flm, Eddie, a
forer criminal, is accused of carying out a fata robbery and sentenced
to death on circumstantial evidence. At the last minute, however, follow­
ing the apprehension of a new suspect, he is granted a padon. Unfortu­
nately, Eddie is aleady involved in a prison break, and interpreting his
8 1
padon as a tick, kills the prison chaplan. The delivery of Eddie's pa­
don, combined with the sympathetic teatment of his character, make it
clear to most viewers that Eddie is innocent of the fatal robbery. How­
ever, Wison suggests that on closer inspection the flm makes Eddie's
innocence an open question for the viewer that ultimately cannot be
answered. This is because the robbery is shown in a highly ambiguous
way so that we cannot say for sure who is involved. The flm's nar ation
is uneliable because it initially encourages us to accept Eddie's inno­
cence but not without intoducing vaious incongrities that undermine
any fnal judgment. In other words, the naration itself suggests an initial
interpretation of the story which obscures the corect interpretation.
Unlike the frst knd of narative unreliability, global unrelability need
not involve a explicit, verbal narator. Of course it could be claimed
that this knd of unreliability still depends on the work of a narator;
just an implicit one. It is not easy to support this claim, however, because
of the fact that to judge an agent to be uneliable requires knowing
something about her personaity and we don' t know anything of the per­
sonality of an implicit nar ator just because he is implicit. Perhaps, then,
it would be easier to try to account for this knd of uneliability without
reference to a nar ator. Gregory Cure does just ths when he gives an
account of complex narative uneliability exclusively in ters of the
intentons of the implied author. 1 33 According to Cure, the implied
author can intend for the images and sounds of te flm to be taken one
way on a supercial interpretation and anoter way on a deeper, more
refective interpretation. In the case of You Only Live Once, since it is the
flm as a whole that suppors a second, conficting interretation of the
events porayed, and the flm is seen as the work of the implied author,
only the implied author - and not someone inside the fction - can be
responsible for its uneliability .
Where nara
ve uneliability is the result of tension between a chaac­
ter's telling ad the showing of story events, we should not assume that
this amounts to a confict between two narators, one explicit and verbal,
the other implicit and visual. As we have seen, some theorists think that
visual nar ation, even when it is a nar ation of fctional events, can be
directly attibuted to the implied author. Where naratve unreliability
does not involve te activity of an explicit, verba nar ator, we also
cannot assume tat some other knd of narator must be involved, spe­
cifcally, an implicit, visual nar ator. As Currie's account suggests, such
cases may be best accounted for in terms of the complex intentions of the
implied author. Thus if, as suggested ealier, we should be interested
primaily in the interpretive fnction of positing the cinematic nar ator,
when we are tying to understand the subtle forms of nar ative uneli­
ability that ae distinct to flm, we may not wish to mention the cinematic
nar ator at all.
Cases like You Only Live Once draw our attention to the fact that
nar ative unreliability takes more or less complex fors. Moreover, it
may be �at only complex forms of nar ative uneliability ae of real
theoretical interest. This is Cure' s suspicion based on the fact that we
don' t assume that narators ae always reliable in the frst place. It is thus
paticulaly when we a led to tst in the reliability of a flm's nar ation
- whether or not it involves a narator - only then to discover that the
nar ation is unreliable in some way, that we become awae of new pos­
sibilities for narration in flm.
The case of unreliable naration draws our attention t o the work of the
viewer in mang sense of the stor because the function of unreliability
very much depends on how and when the viewer is supposed to pick up
on it. Ths brngs us to our last question about nar ative comprehension.
The way in which this question has been framed in the literature is in
ters of how much work the nar ation leaves up to the viewer. In par­
ticula, theorists a interested in whether a flm's naration diects te
viewer to constct the full story herself rather than having it provided
entiely by the sounds and images on screen.
Tere ae important links between naration, the narator, ad nar a­
tve comprehension. Depending on how one conceives of nar ation, or
the narative process as a whole, one will be more or less inclined to
insist on the necessity of the nar ator. And depending on whether one
insists on te necessity of the nar ator, one will be more or less comfort­
able with assigning a role for the viewer in nar ative constction.
Remember that Chatan comes to the conclusion tat every work of nar­
rtive fction; including ever nar ative fction flm, has a narator because
nar ation is an activity cared out by an agent. More specifcally. Chat­
ma conceives of nar ation as a complex communicative activity whereby
the nar ator as comunicator conveys a story to his audience. This is not
the only way of conceiving of nar ation, however. The leading cognitive
fl theorist, David Bordwell, rejects Chatman' s conclusions about the
role of the narator in part because he rejects Chatman' s conception of
naration as comunication.
34 Moreover, Bordwell conceives of nar ation
in a way that implies an active role for the viewer in shaping te narative.
I order for the narator to play te role that some theorists assign
to him, the whole story has to be present in the flm - only then can the
narator serve to tell or show the story from the inside. Bordwell doesn't
think that the whole story is i the flm. In fact, he doesn' t think that any
of the story is in the flm because he thinks that the viewer constructs the
stor herself. First, Bordwell marks a distinction between plot ad story,
for which he borows the terms syuzhet ad fabula fom te Russia
forists, a group of literary theorists active between 1 91 4 and 1 930
who Bordwell claims were the frst to fully theorize the distinction
between the narative and naration. 1 3S The syuzhet, or plot, as the actual
presentation and arangement of story events in te flm cues the viewer
to constrct the fabula, or story. 1 3
In other words, the viewer makes
inferences fom what she is shown on screen, together with her back­
ground kowledge of the world and flm cultre, in order to arrve at a
coherent picture of 'what really happened' in the flm - a chronological,
causal chain of events. 1 37
This may sound more mysterous a process tha it really is. With his
extensive knowledge of the psychology of flm-viewing and flm form,
Bordwell is just building on the observation that no flm gives us ever­
thing in terms of a story. We have to assume that characters don' t cease
to exist in the fctional world of the flm when the actors walk of screen.
We have to assume that a whole night has passed when we ae shown
two consecutive shots of a character lying down to sleep in dakess
and awakening in daylight. And we have to assume the corect order of
events in a character' s life even when we are shown his childhood in
fashback from his deathbed. These ae just three examples of the multi­
tude of assu

ptions about stor events tat we mae in watching a flm.
Such examples suggest that it is false to think of the flm viewer as a
passive recipient of a fully worked-out and comprehensible story fom
an active and contolling narator. If story comprehension is constructive
as Bordwell claims, then an advocate of the cinematic narator would
have to rethink the narator' s role. Perhaps the narator gives us the na­
rative inforation out of which to constct a story; perhaps the narator
and the viewer enter into a collaboration for deterning the whole stor;
or perhaps the viewer takes over the role of the narator in this process.
The question of whether the viewer constucts the story is a much
more complicated one than we have made it sound here. This is paly
because, for Bordwell at least, the question is attached to a much broader
constctivist theory which applies to every level of engagement with
flm - fom the most basic level of image recognition to the highest levels
of interretation and criticism. We will retur to consider Bordwell' s
constctivist theory more closely in the next chapter. For now, let us
keep in mind that we need to consider the activity of the viewer in
following a flm' s story in order to understand more flly the nature of
flm narati
As a fnal reminder, here ae the main points we have covered in this
1 . The status of the cinematic narator; her relation to the story, to the
implied author, and to the viewer.
2. Two standard arguments for the necessity of the cinematic nar ator -
that naration just is the activity of a narrator, and tat only a narator
can give us access to fctional events.
3. Some compelling objections to these arguments concering the status
of the agent responsible for naration, and the indeterminacy of means
of access to the fction.
4. A further agument against the necessity of the cinematic narator that
challenges the assumption that fctiona events ca only be shown by
a fctional agent.
5. Complex fors of two-track narative unreliability only possible in
flm and not in literature.
6. The possibility of global narative unreliability without a narator.
7. The implications of positing a nar ator for our understanding of the
process of nar ative comprehension - in paricula, whether the viewer
passively receives the story from a narrator or actively constcts the
In this chapter we have raised doubts about the necessity of the cine­
matic narator. As we discovered, this has considerable implications for
our understanding of narative unreliability and nar ative comprehen­
sion. We leat that whereas uneliability in verba naration may best be
explained in terms of the intentions of a narator, unreliability in visual
naration may not be. This highlights the fact that, with both an image
tk and a sound tack to manipulate though editing, flmmakers have
a great dea of feedom in how they tell a stor, perhaps more tan in any
other a form.
Given this feedom, we mght expect that flm naratives would be
particulaly had to follow or engage with. This does not seem to be the
case, however. If anyting, flm nar atives are the most absorbing and
accessible knd. The furer question is what exactly is going on when
we engage with a narative flm. Ou work in this chapter suggests that it
need not be due to the work of a narator in supplying a pre-digested
story that we ae able to follow a flm narative. In the next chapter, we
will see the extent to which it is up to us to brng the story together, and
how much of the work assigned to a narator in guiding us though fc­
tiona events could actually be assigned to the viewer.
The experence of watching a flm, especially a popula, mainstream
flm, can feel eforless and absorbing. However, in some ways, ths is
deceptive, since the ease of the experence is facilitated by our perfor­
ing a lage number of complex cognitive ativities. If we ae watching a
nar ative fction flm, for example, we have to fgure out what' s going on
in the story. Ths sounds simple enough and yet requires us to perceive
movement, recognize what is being depicted on screen, interret charac­
ters' behaviour, mae causal connections and more generally fll gaps in
the nar ative, for expectations about what will happen next, and keep
tack of spatial and temporal confgurations witn the flm's fctional
world. Moreover, the viewer's cognitive activity does not end when the
flm ends.
After we have left te cinema or tued of the television, we may
start wondering about the flm's deeper meaing, its temes, moral, or
message - for example, what te flm tells us about life and death, love,
power, sexuality or even te nature of flm itself. And even if we don' t
look for deeper meaing, we almost cernly form judgments about
whether the flm was any good, judgments that go beyond mere prefer­
ence and aspire to objectivit. Tis suggests that there are three principal
activites performed by the thinking flm viewer: comprehension, inter­
pretation, and evaluation. In this chapter, we will analyse each of these
activities and tereby come to see that, although our cognitive engage­
ment with flm is rarely sel-conscious, it is complex ad vaed. In
other words, contar to the familia dispaagement, tere is nothing
' mindless' about the entertainment flms provide.
Comprehension involves our fguring out what is going on in a flm as
we watch it. The most flly developed account of comprehension in rela­
tion to narative flm is the constctivist account given by the leading
cognitive flm theorist, David Bordwell. 1 38 On his account, comprehen­
sion is the activity of coming to grasp the literal meaning of a flm. There
ae two knds of literal meaning: referential meaning consists in the
content of the fctional world - everthing depicted, including
spatial and causal relations that structure the narative, and explicit mean­
ing consists in a moral, message or theme explicitly signalled by the
flm. 1 39 The reason Bordwell labels his account ' constructivist' is that
he takes the literal meaning of a flm to be constructed by or projected
onto the flm by the viewer in response to the flm's structural ' cues' .
Bordwell's constructivism is opposed to te traditional view that the
viewer simply fnds the meaning that is already there in a flm. As we
shall see, it is not just literal meaning that Bordwell takes to be con­
stucted, but also the less obvious meanings tat we grasp through inter­
pretation. At the
level of interpretation, there are, again, two kinds of
meaning: Both implicit and symptomatic meaning are like explicit mean­
ing in being general and abstract - indeed, the same moral, message, or
theme could be any of the three kinds of meaning. The diference is in
the relationship between the moral, message, or teme and the 'text' -
whether the meaning is signalled explicitly or implicitly, or not inten­
tionally signalled at all but rather symptomatic of, say, the dominant
ideology or the flm-maer's hang-ups. l 40 Further ahead, we will exam­
ine Bordwell's account of interpretation as well as George Wilson's
critique of Bordwell's distinction between comprehension and interre­
tation. 1 41 But frst, we will see how Bordwell maps out a hierachy of
cognitive activities involved in comprehension, activities which are
no less constctive for being primaily automatic, instantaneous, and
At the most basic level, the very act of perceiving a flm is constc­
tive. Bordwell embraces the dominant view in cognitive psychology that
perceiving an object does not just require passively receiving visual,
aural or tactile data but also making an inference from the data in order
to reach a perceptual judgment. Bordwell refers to such judgments as
'hypotheses' to suggest that perception is always open to revision in
the face of new data or a new application of relevant background know­
ledge. Perceptual inferences can be made in two directions, either from
the 'bottom up,' as in colour recognition where we draw a conclusion
fom the perceptual input, or from the 'top-own,' as in facial recogni­
tion where the perceptual data is largely organized by our expectations,
background kowledge, and other cognitive activities. ' 42 In either
case, however, whether the 'premises' of the inference derive fom the
perceptual input or fom pror knowledge, it is up to the perceiver to
reach a conclusion. In this sense, then, the perceiver makes what she
. Interestingly, flm depends on our mang the wrong infer­
ences as a result of two physiological defciencies in our visual systems. 143
In order to see a flm as a sequence of continuously lit, moving images,
we must make the wrong inferences from the data which consists in
rapidly nashing light and a rapid display of static images. It is only
because our eyes cannot keep up with the rapidity of changes, both in
light intensity and in the image display, that a flm is what it is experi­
enced as; namely, a motion picture.
For most viewers famlia with the relevant representational, narative,
ad cinematc conventions, once they have constructively perceived the
two-dimensional image sequences projected on screen, it is just as auto­
matic for them to constuct a three-dimensional, fctional world fom
these images. This is not to say, however, that the second level of con­
stction is straightforwad. The viewer must draw on considerable
background knowledge to deterine the right confguration of three­
diensional objects shown in an inherently ambiguous two-dimensional
image. This ambiguity derives fom the fact that the same two-dimen­
sional image could be an image of many arangements of objects. For
example, when we recognize an image of a 'costumed man' we have
aeady disambiguated the image by imagining its object as a paticular
three-dimensional confguration - that of a costumed man. But instead of
a costumed man, we could imagine 'a scatter of gaents fung up and
fozen, with a huge head miles of that happens to coincide, on our view,
with the top edge of the colla' . 1 4 Indeed, if the image is a photograph, it
could have been caused by either three-dimensional confguration.
Thus the process of disabiguating the image, while instantaneous
ad automatic, is still highly informed. But this is not te only process
involved in constructing the three-dimensional world of the flm. In
ad�ition, the viewer must draw on considerable background knowledge
to determne the fctional state of af airs represented by the image. Pa
of constructing the fctiona state of afairs is constructing its spatial and
temporal paameters which extend beyond the image. When an actor
walks out of the fame of the shot, for example, we do not think, without
special reason, that her character has ceased to exist; rather we think that
the character has moved into pat of the fctional world that is not pres­
ently shown to us.
Insofa as the viewer is constcting fctional events and chaacters
according to the order of images, she is constcting what Bordwell,
following the Russian Fonalists, calls the 'suzhet,' or 'plot' . As we
leat in the previous chapter, the syuzhet is te nar ative just as it is
shown on screen - as a incomplete and often out-of-order sequence of
events. By flling in mssing nar ative infonation and re-ordering the
depicted events into a causal sequence, te viewer then constructs the
'abula,' or ' story,' from the syuzhet. The diference between tabula
and syuzhet can be seen when someone asks you to tell tem what hap­
pens in a flm.
In aswering tis question, you would not just describe what was
depicted on screen in the order it was depicted, since te listener would
quickly become lost. If, for exaple, your description began, 'All of
Ralphie's family and fiends are gathered by his graveside, then Ralphie
steps on to a tra, ten he orders lunch . . .' , the listener would likely
interpt with a string of questions: ' How could Ralphie be dead one
mnute ad then boading a tram the next, and how could he be on the
tram and then instantaneously at lunch? Surely you'r not saying his
ghost ordered lunch on the tra?' The listener is expecting you to have
already done te work of re-orderng and flling in the stor so that you
can tell him that Raphie rode a tam to his lunch destination, which
led to some other event, which led to still others, which fnally lead to
Ralphie's death. indicated right at the beginning of the flm but lef unex­
plained until the very end. Thus what happens in a flm is not just what
happens on screen (and on the soundtack). Rather, what happens i the
flm is an imaginatively flled in, re-ordered, ad expanded version of
what happens on screen.
The role of imagination is crcial in the flm viewer's construction of
the fctional world of the nar ative and the nar ative itself. We imagine
at when we are looking at fat images o�-screen, we ae watching fc­
tional events occur in a three-dimensional world. We also imagine, but
without imagining seeing, the of-screen space into which a chaacter
steps when the actor moves beyond the caera's view. And we imagine
what happens between te shot of Ralphie stepping onto the tra and te
shot of his ordering lunch; namely, that he takes the tram to his lunch
destination. All of this suggests that although Bordwell takes himself to
be giving a unifed constctivist account of flm-viewing, constction
at the level of perception may be a diferent thing than construction at
the level of tabula apprehension. l 45 Perception is constctive because it
involves the application of concepts to perceptua data in the makng
of inferences, not because we imagine what we perceive. It is important
to keep in mind this distinction between constuction as an inferential
activity and constction as an imaginative activity because, when
Bordwell proceeds to claim that interpretation, and not just comprehen­
sion, is constctive, he revers to the perceptual model.
Once the viewer has fgured out what is going on in a flm by constuct­
ing te fabula, she may wish to continue generating more abstact
tematic and symptomatic meanings for the flm. However, this is ofen
a activity that only flm scholars pursue. Indeed, this is ofen the only
activity that flm scholars pursue. As Bordwell ad others point out,
ever since its entace into the academy, flm stdies has been heavily
dominated by interpretation at te expense of analysis and evauation.
Moreover, the process by which academc critics interpret flms has
become hghly regulated such tat it is govered by institutional nors.
But te interpretive process is still constctive just by being actively
On Bordwell' s analysis, the critic maps concepts, which ae stctred
as ' semantic felds,' onto those cues in a flm that the critical tadition
considers efective in viewers' comprehension and capable of bearing
meanng. For example, crtics inspired by psychoanaysis mght set out
to interpret a flm in ters of the dua theme of voyeurism ad fetishism,
a theme that is recognized as signifcant within the community of crtics.
Such a interretation would involve a highly selective teatment of the
flm in terms of paricula suggestive features - say, caerawork that
can be descrbed as voyeuristic, or a contast in characters' behaviour
tat ca be described as symbolizing the voyeusmfetishism duaity.
A certain theme becomes institutionaly entenched when it is fexible
enough to be used in many contexts. For exaple, the auteurists favoured
the theme of conession since it alows the critic to tie together dispaate
scenes across a director' s body of work involving confession of any
knd - legal, religious, or personal, and then, perhaps, to tie these scenes
of confe
sion to the 'confessional' style of the director. The process of
mapping semantic felds onto cues is aided by 'socially implanted
hypotheses about how texts mean,' in paricula te hypotheses that the
text is unifed and tat it is related to a exteral world. l 4 In other words,
te critic assumes that he can draw on background kowledge about the
world aound him to make sense of the flm as something with a single,
overall meaning.
The mapping process is achieved by the employment of heuristics,
standad rules of thumb which have proved usefl in generating novel
and plausible interpretations. One popular heuristic is the punning heu­
ristic, which involves, for example, tang the depiction of passageways
in a flm to suggest that the flm is about the 'passage' of life, or tang
the faming of cerain shots to indicate that the flm is about the 'fam­
ing' of innocents. Once the critic has settled on and refned her semantic
felds, and organized thei application to suggest an overall meaning for
the flm, she must write up her interretation in a way that is recognized
as suitably persuasive by the interpretive instittion. Not surrisingly,
perhaps, Bordwell' s analysis of interpretation as highly stadadized has
been very unpopular among academic critics who do not like to think of
temselves as assembly-line workers in ' Interpretation, Inc' . However, a
deeper criticism of Bordwell' s analysis can be found by examining his
prior commitment to constrctivism.
Although, as we have seen, ther may be more than one version of
constctivism at work in Bordwell 's account, one basic idea is that the
flm viewer's activities are constrctive insofar as they ae inferential ad
involve the application of concepts. Since perception, narative compre­
hension, and interpretation are all constuctive processes, Bordwell
assumes that their objects ae constcts. Hence his famous statement,
'Meanings are not found but made' . 147 Unforunately, however, as Berys
Gaut points out, Bordwell is maing a false assumption. The fact that
perception is an active, inferential process does not mean that te objects
of perception - namely the things we perceive around us - are con­
strcted. After all, one can tink that the exteral world we perceive is
really there and yet still think that th
mind' s activities ae constctive.
, a crtic who thinks, contra Bordwell, that meanings are
not made but found in the flm can still allow that the process of fnding
them is active and inferential.
This same critic can also allow that interretation is govered by
instittional norms, since many forms of detection, including criminal
and scientifc investigation, occur within institutions ad yet retain inde­
pendent objects of inquiy. Further evidence against interpretive con­
structivism is suggested by the fact that interpretation is constrained, not
by the nors of the interretive institution as Bordwell suggests, but
by the norms of the flm and its generative context. A critic who fails to
understand how certain techniques were standadly used in a particular
flmmang tadition is likely to misinterpret their signifcace in a flm
which is pa of that tradition. Wereas, for example, a critic would be
right to interpret low-angle shots in Citizen Kane as connoting the power
or threat of a dominant character (or, in certain scenes, as connoting lost
power and pathetic isolationl 4
), he would be wrong to interpret low­
agle shots in Rashomon ( 1 950) in the same way. This is because, in
classical Japanese flmakng, the low-angle shot simply captures te
perspective of someone seated on a tatami mat and has no special expres­
sive force. 1 49 If the critic were the one constructing signifcance, there
would be no possibility of her being wrong about the meaning of low­
angle shots in Rashomon. The fact that she can be wrong indicates that
the nors govering interretation ae not the nors interal to critical
practice but the exteral norms of artistic and flmmaking taditions,
and social practices.
Thus it seems that interretation is not constructive in the way that
Bordwell suggests. Interestingly, however, this crticism may not stick,
since it is unclea whether Bordwell actually endorses the knd of inter­
pretation he analyses. He may simply be descrbing the way things ae
done in flm studies while recognizing tat constructive interpretation is
incoherent - a critic may in fact proceed as if it is entirely up to him what
a flm means when in fact this is not the case. Bordwell may wish to
reveal what is really going on in academic criticism in the hope of under­
mining curent practice. This reading of Bordwell's aims is supported
by the fact that he also calls for a redress of the imbalance in favour of
interretation that has chaacterized contemporay flm studies.
As an alterative to nar owly interretive flm analysis, Bordwell rec­
ommends a ' historical poetics' of flm which is in fact sensitive to the
knd of historical and social norms that properly constrain interpreta­
tion. This approach involves analysing how flms work - how they are
put together technically and forally to achieve cerain ends (say, to tell
a story or embody certain meanings) and to prompt a certain knd of
active engagement on the pat of the viewer. The reason that Bordwell
considers his poetics to be ' historca' is because it involves recognition
of the fact that flms work diferently in diferent contexts - as part of
diferent flmmaking and broader artistic traditions, or when viewed by
diferent kinds of audience at diferent times.
The ambiguity of Bordwell 's aims in giving an account of flm inter­
pretation afects the force of another line of crticism, George Wilson
agues that interpretation is more than Bordwell claims, since there is
a knd of interpretation that is continuous with comprehension. But if
Bordwell's account of interpretation is merely descrptive rather than
prescrptive, Bordwell can respond simply by arguing that this other knd
of interpretation which Wilson identifes happens not to have been part
of the academic tradition to which Bordwell 's stipulative distinction
between comprehension and interpretation applies. Wilson in fact ack­
nowledges that Bordwell may be happy to prescribe this other, neglected
kind of interpretation as pa of a poetics of flm. But even if Wilson does
not underne Bordwell's account of the curent state of interpretation,
he draws our attention to the fact that taking interpretation to be con­
cered with messages ad themes may involve a problematic linguistic
conception of narative meaning to which there is a aterative; namely,
nar ative meaing as the causal signifcance of events.
According to Wilson, there is a knd of interpretation concered, not
with abstract themes and messages, but with a more nuanced grasp of
the referental meaning of a flm. This knd of interretation does not rely
on any specialized theoretical skills ad remains open to rational criti­
cism by those outside particula critical schools. This is because it draws
on the ordinay cognitve sklls, and the nors govering thei applica­
tion, that we employ when tying to mae sense of puzzling events in our
lives. We fnd meaning in, or give meaning to, such events by connecting
tem to other events so that they have a signifcant place in a nar ative
famework or causal patter. Similaly, when we are watching narative
fction flms, we are ofen lef with questions as to why the story took the
tur it did at a paricular point, or why a chaacter behaved a cerain way
in a cern scene. I order to answer these questions, we may have to
look more closely at the details of the narative, and in so doing, we may
make diferent connections between events, or weigh events difer­
5 1 Ks a result, we may fnd that the events of the flm tell a slightly
diferent, and perhaps better, story than the one that we understood on
frst viewing. Alteratively, we mght fnd that the flm does not tell a
consistent story at al and thus exhibits one of the fors of narative
unreliability discussed i. the previous chapter. This kind of interpreta­
tion is essentially refective comprehension that draws on more complex
evidence and pays more attention to the broader narative context than
ordinay comprehension, but nevertheless relies on the same skills in
making causal judgments.
The norms that deterine the reasonableness of our judgments about
real-life causes and efects cary over to our judgments about causes
and efects in a flm' s fctiona world - that is, unless it is clear that cau­
sality works diferntly in a particulaly strange fctional world. Given
the universality of these norms, interpretations involving causal explana­
tion need not be relativized to a pacular academic practice involving
te use of chaacteristic semantic felds and heuristics. We do not need
to be par of such a practice, faliar with all its acceptable moves and
theoretical assumptions, to give tis kind of interpretation. We do, how­
ever, need to be sklled at noticing anomalous nar ative elements and
scrtinizing the flm in order to integrate these elements into a signif­
cant causal pater. This taes practice, but not at the tcks of the
academic interpretative tade. Rather, it takes practice simply at watch­
ing flms wel.
In closing our discussion of interpretation, it is worh notng that whle
flm scholars may be centaly, or even exclusively, concered with the
deeper meaning of flms, ordinary viewers tend to skip from ordina
comprehension to evaluation. As long as we ae able to follow a flm, or
sometimes even precisely because we are unable to follow it, we usually
emerge fom the cinema eager to shae our opinions about wheter te
flm in question was any good.
Have you notced how much we talk about the flms we see and how
much of our ta involves evaluating these flms? If you tell a friend tat
you have just seen a new flm, there is a good chance that her frst ques­
tion will be, 'Was it any good?' Even if your fiend does not kow what
te flm is about, she may still ask for your evauation before or witout
askng for a plot or thematc synopsis. It is, moreover, revealing to think
about ways in which you might answer your friend' s question. You could
say, 'Well, no, the flm wasn' t tat good, but I still lied it' . This implies
that flm evaluation is not just te registering of preferences, likes and
dislikes. Similarly, if you had answered just by saying, 'Well, I liked it,'
te fact tat your fend could then repeat her question ¯ ' OK, but was it
any good?' - indicates that flm evaluation, lie a evaluation in general,
ams at objectivity. This is also supported by the way you could follow
up your initial answer by giving reasons for your evaluaton, reasons
which are grounded in intersubjectively verifable facts. If you had
aswered your frend by saying, 'Yes, te flm was brilliat,' or 'No, the
flm was bad,' it is also liely that she would not be satisfed until you
gave reasons for such an evaluation.
But suppose you went to see the flm with your cousin, and when your
fend asks whether the flm was ay good, you and your cousin give
opposite answers. While you say the flm was good, your cousin says it
was terrible, and neither of you tae yourself merely to be stating a pref­
erence. Upon discovering that you and your cousin have reached diferent
conclusions about the flm, you need not simply shrug and say, ' oh well,
he and I often like diferent things' . Instead, you could proceed to argue
with your cousin in the hope of convincing him, through the presentation
of reasons, to change his mind. But is tis a false hope? Can we really
expect to resolve disagreements about whether a flm is good or bad?
According to Noel Caroll, a theory of flm evaluation addresses just
these questions, thereby aiming to account for the possibility of the ratio­
nal resolution of disagreement.
Before we examine the details of Caroll's own theory of evaluation,
we should note that even though all of us like to evaluate the flms we
see, professional critics still serve a number of useful functions. They
can help us select which flms to watch, either by identifying flms which
match our current tastes or by makng a case for the merits of a flm that
we would otherwise have skipped. In addition, critics can prepae us to
defend our evaluations more efectively, including when we defend our
evaluations against the critics themselves in imaginar conversations
sparked by reading flm reviews. 1 53 Finally, Carroll suggests a way in
which crtics can help us resolve a certain kind of evaluative disagree­
ment; the kind involving a category eror. Critics can introduce us to new
kinds of flms, thereby enabling us to place a flm in the corect category
ad evaluate it according to the standards of that category. This rases
questions about what it means to categorze a flm, how we know we
have categorized a flm corectly, and why a corect categorization might
support an evaluation that we can agree on. To answer these questions,
we need to examine Caroll' s 'Pluraistic Category Approach' to flm
Although there are vaious kinds of evaluative disagreement, Carroll
suspects many disagreements could be resolved if the flm in question
were placed in the right categor or categories. A category can be a genre,
a movement, a school, or a style, which is not necessarily specifc to flm.
To tae a simple example, suppose that your cousin complains about the
lack of explosions in Four Weddings and a Funerl ( 1 994). By pointing
out that the flm is a romantic comedy, you undermine the criticism and
open the way for a discussion of the merits of the flm relative to, say,
The Wedding Crashers (2005) or Annie Hall ( 1 977). Precisely because
these other flms are of the same kind as Four Weddings and a Funeral,
they shae some of the same general aims - for example, the ams of
being fnny and making us root for the principal romantic couple. It is
simply unfar to condemn Four Weddings and a Funeral for a lack of
explosions, since having explosions is not essential to meeting the ams
of this paicula sub genre. Of course, a romantic comedy could have a
explosion, which could either sere a comedic or romantic purpose, or
sere some other purose not specifc to the genre. This indicates that we
ae not limited to evaluating flms by categor, and even when we do
evaluate flms in this way, we ae not thereby ignoring the unique ways
in which particular flms flfl the purposes of their categories.
It is not just a matter of personal opinion as to how a flm is to be
categorized. Caroll names thee kinds of objective reason for placing a
flm in a cerain category: ( 1 ) stctural reasons, (2) intentional reasons
and (3) contextual reasons. I
4 Tere a stuctural reasons for categoriz­
ing Four Weddings and a Funeral as a romantic comedy insofa as the
flm has many chaacterstic features of the sub genre, including witty
dialogue, a sympathetic yet amusingly incompetent romantic hero, ad
a happy ending involving romantic reconciliation. In addition, there is
good evidence, both maifest in the work - say, in the dialogue and
choice of actors - and outside the work - say, in interiews, that the ma­
ers of Four Weddings and a Funeral intended the flm to be a romantic
comedy. And fnally, there is some contextual evidence for the cate­
gOrzation, since romantic comedy is the favoured subgenre of Richard
Curis, the writer of Four Weddings and a Funeral, as well as a staple of
mainsteam, nar ative flm.
As our example suggests, stuctural, intentional, and contextual rea­
sons often converge in supporting a paricular categorzation. The fact
that Richad Curis is the writer of Four Weddings and a Funeral sug­
gests an intentional reason for the categorization, and in general, context
can be a clue to intentions, since flmmakers tend to want to mae flms
in categories that are falia to the audience. Simlaly, stctural con­
siderations can be a clue to intentions - as above, where the dialogue
seems deliberately crafted for romance and comedy. Despite this conver­
gence, however, one kind of reason alone may be sufcient to justify a
particular categorzation, particulaly when there ae no good reasons for
an alterate categorzation.
In sum, Car oll aims to show that the rational resolution of evaluative
disagreements about flms is possible, at least when the disagreements in
question ae the result of a categor eror which we ca work to corect
by the citing of structura, intentional, and contextua evidence. In recom­
mending categorical evaluation as a prmary for of flm crticism, Car oll
aso aticipates and responds to a number of potential objections. l s
Categorical evaluation, Caroll insists, is neither nar owly aestetic nor
inherently conservative. It is not narowly aesthetic because some cate­
gories have cognitive, political, or ethical aims, or aims that depend on
cognitive, political, or ethical crteria. And it is not inherently conserva­
tive, or insensitive to the historical progression of flmmang, because
there are varous ways in which we can understand new categores amost
as soon as tey emere. Caroll also mantains that a flm that cannot
be categorzed is highly unlikely, and indeed, such a flm would be unin­
telligible. This is not to say, however, that ever flm can be easily
categorzed ad indeed, as noted above, one task for te professional
crtic is to introduce us to the mor obscure and unfailiar categores
into which unusual flms ft. Finally, Car oll recognizes that categorcal
evaluation will be complicated, since many flms belong to multiple cat­
egores and meet the ams of each category with varying success.
It is impornt to note that Car oll is not claimng that corect catego­
rzation will end all disagreement about the value of a flm. We may
be able to agree on which categores a flm belongs to and yet be unable
to agree on the standads, and weighting of those standards, for te
categories in question. I 57 Furerore, even if we successfully make the
move from corect categorization to shared evaluation, still other kinds
of evaluative disageement may persist. For example, we might disagree
about whether a flm' s twisted plot structure is efective or simply con­
fsedly -contved, whether a cerain actor was the rght choice for te
protagonist, wheter the flm's special efects are convincing, wheter a
flm' s style of camer�work or lighting, or its colour palette, ae aresting
or simply annoying, or whether a flm is a successfl adaptation of a play
or a novel. Whereas some of these disagreements might invoke catego­
ries, clealy not all of them need to invoke categories. But remember that
Car oll is not comtted to explaining every kind of evauative disagree­
ment about flm. Neverheless, it seems he must explain when it is
sensible to compar flms from diferent categores, according to thei
categories. Car oll suggests tat ths is sometimes a matter of compang
how flms ran in tei respective categores.
Other times, we may be
engaged in a much broader kind of criticism whereby we compare flms
according to the cultural imporance of the diferent categorcal aims that
tey meet. 159 However, Car oll insists tat it will not always mae sense
to compae very diferent fls, perhaps because the aims of the catego­
ries to which they belong ae of equal or unknown value. l 6
Whereas Car oll is caefl to specify the limts of his account, he is
frmly committed to the idea that categorcal evaluation is a central,
perhaps the cental, kind of flm evaluation. This may stke you as ques­
tionable, however, since it is ofen the case that when we say that a flm
is a good flm of its kind, we imply that it is not very good on some
broader scale. But what might this brader scale be? Perhaps it is a gen­
eral aistic scale, as when we assess the expressiveness or composition
of a flm, or when we assess a flm in ters of a imporant element that
it shar�s with another at for, like the actors' performances. Altera­
tively, perhaps the broader scale of evaluation is a scale for flm per se,
as an art for. Citi
en Kane ( 1 941 ) has been caled a ' supremely cine­
matic' flm for the way it exploits what the flm medium can ad canot
show us directly. The flm appears to mae plain everyting that happens
in Kane' s life and yet al it gives us ae surface appeaances, since Kane's
inner life is lef up to the viewers' imagination. The flm is rch with
interpretive possibilities and al because it exploits an appaent limitation
of the medium as a visual record. 1 61 This is just one example, however,
and it remains to be seen whether every flm can be judged, not just as a
cerain kind of flm, but just as a flm.
Victor Perkins' Film as Film: Understanding and Judging Movies, as
its tite suggests, is concered with just this question. This puts Perkins
in the mnorty of flm scholas who have looked beyond interpretation to
criticism ad the role of criticism in our enjoyment of flms. With a focus
on nar ative fction flms, Perkins attempts to derve general evaluative
criteria from an understanding of the foral possibilities inherent in the
flm medium. A flm contains many elements and the key is to bring
those elements into ' signifcant relationship' so as to create a unity which
conveys the flm-maker's vision. This is not easy to do because the
medium pulls in opposite directions, bot towads 'credibility' or real­
ism and towards ' signifcance' or expression and communication. There
is aways a theat of ' collapse,' either the loss of illusion in the pursuit
of expression or the loss of signifcance in te pursuit of realism. 1
2 The
way in which a flm avoids such a collapse, 'absorbs its tensions' , 163 and
achieves unity is by constraining expression with realism. The fewer
'one-fnction elements' a flm has - in other words, the fewer elements
it has which serve only to establish either credibility or signifcance - the
better that flm is.
To illustrate this point, Perkins compaes a scene in Rope ( 1 948) with
a scene in The Loudest Whisper ( 1 961 ) . 1 6 Both scenes use the device of
the 'controlled viewpoint' to create suspense. However, Perkins thins
that the former case involves a superor use of the expressive device
which is less contived because the device is motivated realistically.
In the scene from Rope, our view is restcted seemingly just by the
paeters of the depicted apartment wit which we are already familia.
In other words, suspense is generated by a restiction within the fctional
world of the flm. In the scene from The Loudest Whisper, on the oter
hand, we are suddenly prevented from heaing a crcial pat of the
ensuing conversation by the camera moving, in a cut, behind a glass
partition. There is no reason within the story for our suddenly being
cut of fom the conversation - it is not as if the glass partition actually
slides across our view, in the fction, as we listen to the conversation.
Instead, the suddenness and appaent abitainess of the change in
camera position draws attention to itself as a device for creating suspense
that breaks with te realism of the scene.
Perkns' s measure for cinematic achievement is a knd of unity to be
achieved through realistically motivated stylization. Thus te question
we need to consider is whether this is really a measure for flm as flm
rather than just a particular kind of flm. In considering this question, it
is signifcant
hat te knds of flms which ae of the highest quality on
Perkns's measure ae Hollywood flms from te late 1 940s to the
early 1 960s. This suggests tat Perkins' s measure is approprate to his
examples just- because it refects te aims of the relevant tradition of
flmang. Films in other taditions - for example, the. tadition of
moderism - that do not aim at realistic stylization ae going to be judged
poorly on Perkins' s account. But it is clealy not the case that all
moderst flms are bad. To be fair to Perkins, however, despite te way
the title of his book implies tat he is interested in how we evaluate any
kind of flm just as flm, he belatedly admits that he has ' deliberately
restcted the feld of this inquiy in order to examne sources of value
within a paicula for' .
This suggests tat Perkins might just be
doing the kind of evaluation that Car oll describes - identifing realistic
stylization as a central aim of the category of classical Hollywood flms
and then assessing paricula flms according to the standards of that
Wheter tere can be a single evaluative standard for flm as flm
remains an open question. Caroll is skeptical of this possibility, since
attempts to identify such a standad often refect a bias towards certain
stylistic tendencies in the history of flm. Neverheless, te fact remains
that flm evaluation operates on multiple levels: When we evaluate flms
by category, the kinds of category we invoke vary in teir breadth - from,
saY,flm noir to melodrama to nar ative fction. But we can also praise a
flm for being 'pure cinema,' or even great a, which may involve a
1 00
appeal to the place occupied by tat flm in the culture at large and the
role it has played in people's lives.
Here is what we have covered in this chapter:
1 . Bordwell's constrctivist account of story comprehension, including
his account of the actively inferential process of perceiving the
images and sounds that make up the flm, and his account of the imag­
inative process by which we follow a flm's story and navigate the
2. Bordwell's analysis, if not his defence of, academic interretation as
a self-justifying and replicative practice.
3. The objection tat Bordwell ' s constructivism does not support the
claim that a flm's meaning is made by the viewer.
4. Wilson's agument for a diferent knd of interretation than the one
Bordwell describes, a kind of explanatory interpretation not reserved
for specialist critics.
5. Caroll's account of objective categorical fl m evaluation.
6. Perkins' critera for judging ' flm as flm' which tum out to be criteria
for judging one kind of flm as that kind of flm.
In this chapter we have leat that the nar ative flm viewer is a think­
ing viewer, both during and after the viewing experience. This is not to
say that the flm viewer is cold, calculating, or distant. As we have seen,
much of the viewer's cognitive activity is uncalculated just because it is
unconscious and automatic. And, as we shall see in the next chapter,
there are a great many ways in which a viewer can rspond with feeling
to a flm. These responses often rely on our understanding of the flm and
may, in tm, infuence how we interpret and evaluate it. Thus the division
set up by this chapter and the next between the thinking and the feeling
viewer is an artifcial one simply for the purposes of analysis. The flm
viewer is in fact both thining and feeling, and thinking may be a com­
ponent of her feelings just as feeling may infuence her thinking.
To retur to the details of this chapter, Bordwell is surely right that
comprehension is a complex constructive process even if that does not
mean it is entirely up to the viewer what a flm literally means. Similarly,
if we take Bordwell to be providing a crtical expose of the curent state
1 0 1
of interrtation, we ca reconcile hs account with the possibility of
other kinds of interpretation, including the explanatory kind described
by Wilson. Whereas an ordinay viewer may be unlikely to attempt a
psychoaaytic, auteurist, or otherwise theoretically laden 'reading' of a
flm, he is highly likely, especially when confonted by a paicularly
complex flm, to engage in the interretve process of refective integra­
tion required to mae the most sense of the flm's narative.
When it comes to evaluation, Caroll and Perkins have given us a lot
to think about, paticularly in relation to the remakably vaed and
numerous criteria by which we judge diferent flms. There is clealy a
great deal of work still to be done on the rational foundations of flm
evaluation. At this stage, we must commend te flm scholars who have
begun this work. By breang away fom the mainstream of academic
interpretation, they are able to refect on the fact that, when it comes to
flm, everone's a crtic.
1 02
The ways in which we respond with feeling to flms a rich and
vaed - arguably as rch and vaed as the ways we respond in rea life.
We may pity the plight of the doomed heroine or feel temendous anxiety
as the shak approaches the unsuspecting swimmer, perhaps resisting the
temptation to yell at the screen, ' Get out of the water! ' We may auto­
maticaly jump in our seats at a sudden explosion or recoil fom the sight
of a gruesome crme scene. We may feel indignant. sad, joyful, excited,
apprehensive, dejected. fstrated, hopefl or jubilant, either aong with
te chaacters, on the behalf of the chaacters, or about some tur of
events in te story. And lastly, we may be left feeling a cerain a way by
the f as a whole - perhaps vaguely anxious, buoyed up, exhilaated or
sweetly melancholic.
All of tese responses involve ' afect' or feeling, even if we do not
consider all of them to involve emotions. You might not think that being
starled by a sudden explosion counts as an emotion, since it is an auto­
matic refex response. Alteratively, you might not think that the anx­
ious or melancholic mood that a flm puts you in counts as an emotion
because it is an amorphous state without an object ¯ I a not anxious
or melancholic about anything in paicula. The distinction between
emotions and oter kinds of feeling is in fact contested even though phi­
losophers have taditionally focused on a narow class of emotions. For
our puroses, it is import to recognize that such a distinction is often
made even if. in the end, it may not be essential to uphold it.
Broadly speag, a emotional response ca be understood a a rs­
ponse to a event of signfcace to the subject involving bodly changes
and typic111y characteristic motivating behaviour. Given the impor­
tance of emotions for the quality and meaing of life, it is not surprsing
that most of the great clasical philosophers fom Arstote to Descaes
to Hume have somethig to say about te natur of emoton. What is
1 03
surprsing, however, is that in the twentieth century, philosophers stopped
paying attention to emotions. And it is only quite recently, partly in
response to work in psychology, neurology, and evolutionary biology,
that the philosophy of emotion has become a respectable feld of study.
Until even more recently, this feld was dominated by a particular knd
of theory that 1 ) only applies to the typical instaces of paradigmatic
emotions -for example, fea, anger and j oy -but not to atypical instances
of these emotions or to other kinds of feeling like moods and refex
responses; and, 2) distinguishes between diferent emotions by te cog­
nitions they involve. Most commonly, the kind of cognition tat an
emotion is thought to involve is a belief or evaluation. Thus you can' t be
angry with someone unless you believe that she has slighted you; you
can' t be afraid of something unless you believe that it represents a threat.
Even though both anger and fear are felt, bodily responses, what makes
them anger and fear respectively, and not some oter emotion, is te knd
of belief they involve.
As we shall see, the assumption that every emotion involves a belief
has led to some difculties in accounting for our emotional responses to
works of fction, including fction flms. Recently, however, philosophers
have begun t
question whether a belief is absolutely required for having
an emotion. This is paly a result of recognizing vaous situations in
which emotions persist without the requisite belief - for example, when
someone is afraid of fying despite knowing that it is safe. In response to
the complexity of real-life cases, alteratives to the belief-theory have
emerged. One alterative is to claim that emotions still necessarily
involve a cognition but not a belief. Given the way that emotions focus
our attention on those aspects of our situation which ae most signifcant
for us, tis kind of theory suggests that the cognitive element of emotions
is something like a perception, a construal or a way of seeing. 1
alterative is to claim that an emotion does not necessarily involve any
kind of cognition. The crucial focus that an emotion provides is achieved
by bodily means, and though emotions ae infuenced by cognition, they
can be characterized in terms of their unique neurphysiological pro­
fles. 1
This suggests that all felt responses could be emotions. While
there ae compelling arguments for both alteratives, as well as for the
original belief-theory, we will not be defending a paticula theory of
emotion in this chapter. For our purposes, we need only take note of a
growing philosophica appreciation of the complexity and heterogeneity
of felt experience.
In te frst two of the following tee sections, we will be deaing
almost exclusively with those feelings taditionally thought to be emotons
because tey take an object and involve cognition. Both of these sections
dea with our emotional responses to chaacters or the nar ative content
of the flm more generally. The frst section examnes two puzzles about
our emotional responses to fctions. The second section exanes one of
the prmary ways in which we engage with chaacters which is empa­
thetic id
ntifcation. Once we reach the thid section, our perspective
broadens to include the full range of feelings elicited, not just by what
the flm is about, but by the flm itself in its stylistic and technical aspects.
In this way, we will begin to see the remakable extent to which flm is a
matter of feeling.
Since most of the flms that we watch ae works of narative fction,
our ana1ysis of flm engagement canot proceed without facing an appa­
ent paradox concering the intelligibility of having emotional responses
to fctions. Fornately for us, there is aleady an extensive literature
on the so-called paradox of fction. Since flms ae not the only knd of
fction, this paradox does not pertain exclusively to flm. But if we wish
to examine the ways in which various flm stctures support our felt
engagement, we frst need to mae sure that we understad the natre of
ou engagement. Afer al, if it ts out tat we cannot in fact respond with
genuine feeling or emotion to fction in genera, we canot proceed to a
anaysis of te ways in which fcton flm i paicula suppors feeling.
A paadox is a piece of reasoning that appears perfectly sound but
neverheless leads to a false conclusion. To solve a paadox, you must
show that the reasoning it involves, despite appeaaces to the contrary,
really is unsound. This does not mae te paradox disappear, since there
is still a piece of reasoning that appeas to be sound when it is not. But in
solving the paradox, you show that the appearance of soundness is merely
that - a appeaance. You may even go on to explain why the piece of
reasoning under consideration has ths appeaance, thereby preparing
others not to be fooled. Alteratively, you could simply accept the
contadiction implied by te paradox and lea to live with it. Bot of
these approaches have been taken to the paadox of fction.
l OS
The Paradox of Ficion
The paradox of fction is often expressed in ters of three propositions,
all of which seem te even though they cannot all be tre without yield­
ing a contadicton.
1 . We only respond emotionally to what we believe to be real.
2. We do not believe that fction is rea.
From these two premises, it seems to follow that we do not respond emo­
tionally to fction. And yet it seems that
3. we respond emotionally to fction all the time.
Those who ae committed to solving the paradox of fction aim to
show that one of these propositions is false. Others, though admittedly a
minority, have accepted the paadox and concluded that our responses
to fction ae simply irational. 1
69 If, however, you fnd it implausible to
think that we lose all sense of reason as soon as we become engaged with
fction, �� is worth pursuing a solution to the paadox.
bate about the paradox of fction has evolved to the point where
a signifcant number of those involved agree that the best route to a solu­
tion i�volves a denial of the frst proposition - that we only respond
emotionally to what we believe to be real. Before embarking on this
route, however, we need at least to consider solutions tat involve deny­
ing either the second or the third proposition.
In order to solve te paradox by denying the second proposition, one
has to give an account of how, in the course of engaging with fction,
one' s belief that the chaacters and events of the fction ae not real is
temporaily suspended. 170 It turs out that giving such an account is
difcult to do. In real life, we might suspend a particula belief before
discading it completely in the face of new evidence that challenges the
belief in question. But this does not seem to be what's going on when we
engage with fction, since there is no new evidence to challenge my belief
that, for example, the zombies depicted on screen are not real. You might
reply, however, that the fction itself provides the evidence - the incred­
ibly life-like images of the zombies compel me at least to suspend my
disbelief, if not quite adopt a new belief, in real-life zombies. But there
is a problem with this reply. If it is the fction that supplies the evidence
tat challenges my belief, then I must have aleady suspended my belief
that the fction is not rea in order to have te evidence I need for
suspending my belief that the fction is not real. If I have already sus­
pended my belief in order to treat the fction as evidence, it canot be
te fction that provides evidence supporing a suspension of belief.
Even if it were possible to suspend our belief tat te fction is not
r, t� way that we experience fctions suggests that this is not what's
going on. If I no longer believed that the vicious zombies on screen wer
fctiona, I would surely not remain seated in the cinema for long. Pre­
sumably, upon discovering that there are kller-zombies on the loose,
I would alert the authorites and tae measures to protect myself and my
loved ones. But I do not do ths. I other words, I do not act as though
I have suspended my belief that the fction is not real. Moreover, it is not
just that I stay seated in te cinema, but I seem to be enjoying myself
as I watch te zombies approaching on screen. I a able to appreciate
the vivid depiction of an army of zombies surging forwad with as
outstretched, the use of special efects or highly emotive music, the
imporance of this scene for the narative, ad so on. Surely, if I had
suspended my belief that the zombies a fctional, I would be too fright­
ened to appreciate te flm in ths way.
So much for solving the paradox by denying the second proposition.
Wat about a solution that involves denying the thrd proposition - that
we respond emotionaly to fctions? Philosophers who attempt a solution
of tis knd do not deny that we respond to fction at all . Rater, tey
deny that our responses consist in genuine emotions.
71 Instead our
responses to fction consist in fctional, mae-believe, or pretend emo­
tions. These responses may have a genuine feel to tem - while watching
te zombies approach, I feel my hear pounding and my hands turing
clamy as I grip the ams of my seat. But assumng that ou responses
cannot be reduced to brte feeling, insofa as they only involve beliefs
about what is fctionally the case, they cannot be real emotions.
This knd of solution immediately raises some questions: What exactly
is a fctional emotiona response? How can a real person have a fctional
response? And if none of the components of the response are fctiona,
why is the response itself still only fctiona? A belief about what is
fctionally the case is still a genuine belief, after all. And the feelings
involved in my response ae most certainly genuine insofa as I a expe­
rencing changes in my body that can be objectively measured. The
only component of my response tat is fctional is its object, what my
1 07
response is directed towards - for example, the zombies on screen. And
while the status of an emotion's object is imporat for determining the
appropriateness of the response, it does not determine that the response
is a genuine emotion. If my fea is directed at a ktten, we might wat to
say tat my fea is inappropriate. But it is still fea. Moreover, some of
our emotions seem not to take an object at al - for example, when I feel
ang but do not kow what about - and yet they a no less ral for that.
Solutions to the paradox of fction that involve denying the third
proposition thus leave us wondering why beliefs about what is fctionally
the case - that is, that there ae zombies on the loose - canot inspire
genuine emotions. This brings us to the tird kind of solution to the
paradox of fction which denies the frst proposition - that we only
respond emotionally to what we believe to be real. The most comon
form of this solution consists in a denial, not just of te necessit of a
belief in the reality of the object of our emotion, but of the necessity of
any kind of belief for a genuine emotional response. In other
ords, this
solution engages with the curent debate in the philosophy of emotion
about what knd of cognition, if any, is essential to emotion.
The philosopher who wishes to solve te paradox of fction by deny­
ing te frst proposition has two options: She can accept tat emotions
always involve cognition but deny that emotions always involve beliefs.
Then she ca suggest that emotions involve another knd of cognition
like a thought or a perception which, unlike belief, is supported by
fction. Alteratively, she can refse to accept that emotions always
involve ay knd of cognition. Then she can suggest that flm media ae
ideally suited for the spontaeous excitation of emotion or feeling with­
out ay cognitive processing. Films ae pariculaly good at exploiting
our phobias, for example, in an automatic, visua way - swinging aeral
camerawork or the image of a wrthing mass of snaes may be enough to
trigger a viewer's vertigo or ophidiophobia, respectively. A potential dif­
fculty with this second option, however, is that those felt rsponses to
flms that ae automatic and visceral tend to be triggered by stylistic and
technical means rather than by the flm's content. But the paadox of fc­
tion concers just those responses we have to characters and events
which ae pa of content. For this reason, we will leave aside the possi­
bility of emotions without cognition until we reach our discussion, in the
fnal section, of the general emotional appeal of a flm.
Several philosophers who tae the frst option claim that a fction
prescribes paricula imaginings or thoughts about its content and tese
imaginings or thoughts in themselves inspire an emotional response. l
1 08
In rea life, we may shudder at the thought of something horible or
amuse ourselves with whimsical imaginings. This, some claim, is all that
is going on with fction but in a more systematic way. Wen I am shown
a zombie on the flm screen in all its horble detail, I tin about or
imagine just what I am shown. In other words, the depiction of the zom­
bie constitutes the content of my thought or imagining. But I can think
about o

imagine the depicted zombie as a combination of just those
properes specifed in the depiction witout believing in the zombie.
I can think about a creature that looks human except for its dead eyes
ad pallid skn, its tance-like lumbering movements, its unstoppable
dve to kll, and so on, for all the featues of the zombie that I recognize,
independently fom each other, fom real life. Moreover, it is very easy
for me to do this given the vivid moving image of the zombie projected
rght in font of me.
One objection to this account is that it fails to preserve the intuition
tat the zombie itself is the object of my fea. 1
3 If you ask me what I' m
aad of, I' ll say the zombie. I cernly won' t say that I' m afaid of my
thought about the zombie. However, this objection fails to recognize that
tere ae diferent kinds of object tat a emotion can take, and that te
cause and object of a emotion need not be the same. Thus while I a
fightened by te thought of the zombie - the thought causes my emo­
tion, I am fghtened of the content of the thought; namely, the zombie
itself. My thinng about the zombie is what makes me aaid but it is
what I a tinkng about tat is the taget of my fear.
Anoter objection to the thought or imagining account is that it is
limited in scope. Even if it is the case that I could be frightened by the
thought of the zombie, it is also possible that I a fghtened simply by
what I see on screen, witout thinkng much about it at all. Just the look
of te zombie - those dead eyes and that cadaverous complexion, the
reaching ams ad unnatural movements - is enough to scae me. My
response is more imediate tan what one would expect if it involved
a lot of tought. Moreover, te possibility of this knd of responsive
immediacy is suppored by recent experiments in neurophysiology.
These experiments show that the emotional processing of perceptual
data can occur independently of cogntive processing because there is a
direct route fom te perceptual apparatus to those parts of the brain
that ae crucial in generating emotion. Perhaps, then, the most promising
solution to the paadox of fction involves pointing to a rage of emo­
tional tggers - thoughts, imaginings, ad perceptions - that can
substitute for the belief that the fction is real.
1 09
The Paradox of Horror
Once we expand our understanding of emotions, it seems that the para­
dox of fction is eminently solvable. Unfortunately, however, we are not
yet fee of complications surounding our felt engagement with fction
flm. Even if we can respond with genuine emotons t fction, a fer
paadox emerges when those responses a negative ad thus intnsically
unpleasant - emotons le fea, pit, anxiety, ad sadness. I rea lfe,
these a the sorts of emotons that we do not want to have. But i the con­
text of fction, some people appea to seek them out and even enjoy them.
This paadox ases with the enjoyment of any negative emotion in the
context of fction. However, there are paicular versions of the paradox
for paicula genres of fction that ae defned in ters of the knds of
response that they aim to evoke. The paradox of tragedy has a long and
distinguished philosophica history: Aristotle and later Hume were both
puzzled by the way that tagedy aims and ofen succeeds in creating a
enjoyable experence of fea and pity. More recently, philosophers have
tured to te paadox of horor, which will be our focus here. The reason
for this focus is that by fa the largest portion of horor fctions are
flms. Indeed, tere seems to be something about flm media which make
them pariculaly well�suited both to the depiction of horor and to the
evocation of the emotions of fea and disgust that defne the genre}74
Just like the paadox of fction, the paradox of horor can be expressed
in ter
of tree propositions, all of which seem te but together yield
a contadiction.
1 . Negative emotions lie fear and disgust a intnsicaly unpleasant.
2. Horor fctions aim to evoke fear and disgust.
3. Some people enjoy horor fctions
If what it means for a emotion to be negative is just tat it is experi­
enced as unpleasat, and if tis aounts to the same thing as our not
enjoying te emotion, given that we cannot simultaneously enjoy and not
enjoy the exact same thing, it seems to follow fom tese two premses
that we cannot enjoy horor fctons, or at least not the emotive aspect of
horor fctions. And yet, it certanly appears as though some people enjoy
horror fctions, and what they enjoy about horor fctions is just that the
fctions a frghtening and disgusting.
One approach to solving the paadox involves claiming that what some
of us enjoy about horor is not the fea and disgust evoked but something
1 10
else - for example, the aousal and satisfaction of our curiosity about
te impossible creaturs we encounter on screen, 175 or the sense had at
the end of a flm that we have worked unpleasant emotions out of our
system. 1 7
Perhaps one could also claim that we enjoy a paicula com­
ponent of fea and disgust - for example, the 'rsh' we feel as our body
releases a burst of adrenalin and our heart rate increases - but that this
does nQt amount to enjoying the emotions themselves. Finally, one could
clam that we enjoy a sense of our own courage or resolve in being
able to face unpleasant emotions at least in te context of fction.
The touble with all the solutions of this knd, however, is that they
essentially mss te point. There is only a paadox to solve in the frst
place if we grant that some of us et0Y fea and disgust. But this is
precisely what solutions of this kind refse to grant. The challenge is
thus to explain how we can enjoy negative emotions given that they a
negative . and thus unpleasant. One way to meet tis challenge is to
ague that we can enjoy negative emotions because negative emotions
ae only contingently unpleasant. Some theorsts claim tat we can enjoy
negative emotions when tey ae under our contol - for example, when
I can direct my attention away from the gor bits in a horor flm.
Other theorists claim that we can enjoy negative emotions as long as
they do not take an unpleasat object. Fea is only unpleasant in re
life insofa as it is directed at something unpleasant - specifcally some­
tng that constitutes a theat. Wen I fea a fctional monster on screen,
I am not responding to a genuine threat and thus my fea need not be
There are varous problems specifc to eah of these suggestions, but
one problem they have in comon is that i mang negative emotions
only contingently unpleasant, they seem to go against our understanding
of negative emotions. If someone told you that the emoton he exper­
enced on the death of his beloved great aunt was highly enjoyable, you
would surely hesitate to call hs emotion grief. 1 7
This suggests that te
connection between the negativity and unpleasantness of an emotion is
necessay rather than contingent. But now we ae back at square one,
facing on�e again the paadox of enjoying what, by defnition, cannot be
There may still be a way out of ths bind, however. Bers Gaut sug­
gests that only in the typical case ae negative emotions experienced as
unpleasant. 1 79 This means that, necessarily typically, negative emotions
a unpleasant. In the atypical case, however, this need not be so. Atypi­
cal cases ae not limited to fction but include cases like the enjoyment of
exteme spors and roller coasters. But the experence of horor is te
atypical case in which we ae interested. In the atypical case, what makes
a negative emotion still count as a negative emotion when it is experi­
enced as pleasant is just the background of typical cases in which human
beings experience te emotion as unpleasant. Just as what the masochist
enjoys only counts as pan against a backgound of noral aversive
reactions, so what the horor buf enjoys only counts as fea against a
background of unpleasant experiences.
In his own caeful and detailed presentation of the agument, Gaut
spells out the varous philosophical presuppositions that ae required for
this solution to work. If there is something about the way that Gaut's
solution appeals to the typical case that doesn' t seem quite right to you,
the next step would be to examne closely the presuppositions that
ground this appeal. Then again, you may still be inclined to tink that
there is no paradox needing to be solved in the frst place. That is, you
may suspect tat it is never the case that what we enjoy about horror
fctions is the fear and disgust that they evoke. Even when a, horor
buf complains that a particula horor flm wasn' t frightening enough to
be enjoyable, tis need not indicate that when the horor buf does
enjoy a horor flm, what she enjoys is being fghtened (and disgusted).
Perhaps instead, by complaining tat a paicular horor flm wasn' t
{rghtening enough, the horor buf is signalling that the fl m failed to be
convincing as a result of poor nar ative stcture, editing, or special
efects. Moreover, the horor buf may describe a good horor flm as
fightening without being frghtened herself. Probably she wouldn' t be
such a buf if she was being frghtened all te time. Rather, when she
describes a good horor flm as frightening it is because she recognizes
that, in general, a good horor flm can fghten people even though te
ones who ae frghtened tend not to enjoy te flm.
Before leaving the paadox of horor, it is worth noting one other way
in which our fear of horor is disputed. Is our fear, fea for te endan­
gered characters or fear for ourselves? Am I fightened of the zombies,
along with the characters, or am I frightened for the characters, whether
or not they're frightened themselves, because they a endangered by the
zombies? Some theorists deny that it ever makes sense to fear for our­
selves in response to fction because we canot actually be teatened by
something that is only fctionaly dangerous. And yet, we often tal as
though we're afaid of the same horors as the chaacters. Perhaps what
feels like genuine fear in these cases is really only pretend or imagined
fear. But perhaps not. In the next section, we will consider how we may
1 1 2
come to shae genuine emotions like fea with chaacters through an act
of identifcation.
When o.rdinary flm-goers talk about their experience with a particular
chaacter - usually a protagonist - in a flm, it is quite common to hear
remarks of the following kind:
'I could really identify with her' ;
'I think the scene moved me so deeply because I identifed with him' ;
'There was nothng about her with which I could identify' .
It i s highly likely that the frst two remarks ae part of a positive evalua­
tion, and the third remak par of a negative evaluation, not just of the
way a particular chaacter is developed, but of the flm as a whole. Indeed,
remakng on whether or not a central chaacter with whom we ae
clealy meant to identify is one with whom we really can identify may
make the diference between a good ad a bad flm - or a highly engag­
ing and a not-so-engaging flm. Given this, ad our everyday ways of
talking about whether or not we like a flm, identifcation seems to be
a phenomenon that we cannot ignore in this chapter.
But what exactly is identifcation and why is it so important? It is this
question that will concer us here. In what follows, we will lin what we
ordinarily mean when we talk about identifcation with the vaous ways
in which theorists have handled the notion. First we will consider what
psychoanalytic flm theorists have to say, then what cognitive flm theo­
rists and philosophers have to say, about identifcation. Ultimately, we
will conclude that identifcation is a valuable imaginative process that
supports empathetic responses to chaacters.
This is not meant to suggest that identifcation represents the only
way in which we engage wit flm chaacters. When we identify with a
chaacter we seem to share her emotional response - for exaple, when
the zombies scae me almost as much as they do the flm's protagonist.
But as well, we can have responses to chaacters that ae not shared with
them - for example, when I pity a chaacter that is in no way self-pitying.
Such responses ae usually described as sympathetic whereas responses
involving identifcation a usually descrbed as empathetic. Even though
1 1 3
we will also explore this distinction in what follows, our focus will
remain on identifcation. 1 8o This is because identifcation is perhaps the
most mysterious and misunderstood, but also the most powerfully inti­
mate, aspect of our felt engagement with flm characters.
Psychoanalytic I dentifcation
One of the defning characteristics of 'psycho-semiotic' flm theory is
its use of psychoanalytic theor - particulaly of the Lacanian vaety -
to understand our experience of flm. Given this, it is no surprise that a
standard notion of cinematic identifcation in flm theor is modelled on
a particula subconscious phenomenon that psychoanalytic theorists take
to be centa to our psychosexual development. In fact, the reason that
psychoanalytic flm theorists call cinematic identifcation, ' identifca­
tion,' is just because this is what Lacan calls te analogous developmen­
tal phenomenon.
According to Lacan, one of the most signifcant stages in our psycho­
sexual development is the so-called miror stage. This is the stage we
enter at around eighteen monts of age when we develop a powerfl
sense of self by identifying with our own miror images. Seeing our
refected selves for the frst time entails recognizing ourselves as discrete
individuas for the frst time. This recognition is paly mistaken, how­
ever, because the infant assumes that his miror-image self has the motor
capacities that the infat himself has yet to develop but which match his
curent mental capacity. Thus Laca implies tat identifcation, whether
with one' s own image in infancy or with others later on, involves the
projection of an ' ego ideal' such tat one sees an enhanced version of
oneself in the object of identifcation. In flm, Lacaia theorsts use this
idea to explain the identifcation of viewers with the larger-than-life
heroes portayed on screen.
It is worth noting that this notion of hero identifcation in te psycho­
analytic tadition has inspired a feminist critique of our engagement with
mainstam flm. In a highly infuential aticle,
1 8 1
British flm-maker,
critic, and theorist, Laura Mulvey, claims that taditional movie heroes
exhibit masculine traits and tus only male (or 'masculinized' ) viewers
identif with them. In t, this gives the male viewer a sense of contol
over the events of the story. Such control is crcial for counteracting the
efect of the female screen icon, which, by demanding erotic contempla­
tion, efectively stalls the nar ative. In this way, the sense of empowerment
1 1 4
derived fom identifcation in general taes a particula and patiarchal
form in the context of nar ative flm. Interestingly, Mulvey' s crtique,
while provocative and revealing, does not question the psychoanalytic
account of identifcation. Rather it assumes the trth of tis account in
order to claim that mainsteam flm seres patriachal ends. Thus, for the
sake of comng to a crtical understanding of identifcation, we can tr
our attention back to the psychoanalytic account itself.
According to Chistian Metz, the most infuentia of the psychoana­
lytic flm theorists, pa of what explans why we watch flms and how
we make sense of them is tat flm reactivates infantile identifcation.
The crucial similaity is what Metz calls, 'the play of presence and
absence' - te simultaneous presence of an image and absence of the
referent on the screen or in the mr or. It is because our infantile experi­
ence with the miror and our experience with flm are both experiences
of the play of presence and absence that Metz thins we can understand
tem in terms of the same subconscious process of identifcation.
1 82
The pleasure of identifcation explans why we enjoy watching flms.
More importantly for Metz, however, the role that Lacan assigns identi­
fcation in comunication explains how we make sense of flms.
According to Lacan, any knd of comunication - and social life in gen­
eral - is impossible without the alteration of the subject position
facilitated by identifcation. The fact that I have to identify with someone
in order to understand what she is telling me suggests to Metz tat I also
have to identify with the source of telling (and showing) in a flm in order
to understand the flm. The source of a flm's telling and showing is, in a
literal sense, the camera. Thus te fundamental role of identifcation in
comunication leads Metz to clam that identifcation with the camera is
the primary form of identifcation in flm. Identifcation with chaacters
is secondary because the intelligibility of flm in general does not depend
on it. After all, there are intelligible flms that do not have chaacters.
There is clearly much to be questioned in Metz' s theor. Is the psycho­
analytic account of identifcation even plausible and is Metz really
justifed in applying this account to flm? Morover, does the way in
which Metz applies this account, assuming te account makes sense,
really illuminate the cinematic experence? What does it mean, exactly,
to identify with the camera? All these questions and more have been
taen up by critics of the psychoanalytic approach to flm. While it is
crcial to do tis, what is most imporant for our purposes is to consider
whether the psychoanalytic account of flm identifcation helps us make
1 1 5
sense of the kinds of things we say about identifcation when we leave
the cinema or tum of the television.
Whatever conclusion one reaches about the plausibility of the psycho­
analytic account, the fact remains that this account does not seem to
apply to the kind of identifcation we actually talk about. Personally,
I have never heard anyone emerge from the cinema saying, ' gosh, I could
really identify with the camera in that flm! ' Indeed, insofar as psycho­
analytic identifcation is subconscious, we wouldn' t be in a position
to comment spontaneously on our having identifed wit the camera.
Of course, if we were well-versed in psychoanalytic flm theory, we
might infer from our comprehension of the flm that identifcation had
occured. But if this kind of inference grounded any remark about iden­
tifcation, the evaluative weight of these remarks would be lost. Wen we
say something about the way a · flm suppors identifcation, we mean
to single out that flm - not just as one of many flms which ae intelli­
gible, but as a good flm among intelligible flms. According to te
psychoanalytic flm theorist, primar identifcation occurs in every case
of our understanding a flm and thus doesn' t seem to be something
worth remaking on. The fact that we do remark on identifcation sug­
gests that we need to look elsewhere for an explanation of the conscious
phenomenon of character identifcation. l s3
So where mght we look? The obvious place would be in anoter
branch of flm theory - cognitive flm theory. Unforunately, some cogni­
tive flm theorists and philosophers of flm ae so suspicious of the
historcal link between identifcation and psychoanalysis that tey wish
to abandon the notion altogether. The highly circumscribed and idiosyn­
cratic use of the term 'identifcation' by psychoanalytic flm theorists
just highlights for them the instability and heterogeneity of the notion in
general. Noel Car oll, for instance, gives us a sample list of all the difer­
ent things we can mean by 'identifcati
n' with characters:
. . . that we like the protagonist; that we recognize the circumstances
of the protagonist to be signifcantly like those we have found or
fnd ourselves in; that we sympathize with the protagonist; that we
are one in interest, or feeling, or prnciple, or all of these with the
protagonist; that we see the action unfolding in te fction fom te
protagonist's point of view; that we share the protagonist's values;
that, for the duration of our intercourse with the fction, we ae
under the illusion that each of us somehow regards herself to be the
protagonist. 1 8
1 1 6
If 'identifcation' is being used to cover all of these phenomena, plus
the subconscious phenomenon that psychoanalytic flm theorists ae
interested in, how usefl can this ter really be? This is only the frst of
Car oll's worres, however. He also agues that even if we can isolate a
core meaning of the ter, 'identifcation,' the phenomenon it refers to is
not pat - or not a signifcant par - of our cinematic experience. Car oll
suggests that te core meaning of 'identifcation' concers the viewer's
duplication of a chaacter's emotion though empathy but he denies that
tis reg
laly - perhaps ever - occus. This is because, automatically, an
actual person watching a flm and a fctional chaacter in the flm stand
in very diferent relations to the events of the story. For the character,
those events ae rea and directly afect her fate. For the viewer, on the
oter hand, those events ae fctional and the chaacter' s fate is pat of a
lager constcted whole. This means tat even if the chaacter and the
viewer have the same knd of emotion, the viewer's response does not
duplicate the chaacter's because it takes a diferent object. So, for exa­
ple, if both the viewer and the character experience fea, the viewer's fea
is fea for the chaacter whereas the character's fear is fea for himself.
As well, the character's fea is fea of a personal threat whereas the view­
er's fear is a fea of a threat to the chaacter and may be mixed with other
emotions that the chaacter does not have like enjoyment or curiosity.
Car ol's objections desere our attention and further aead we will
consider how to respond to them. First, however, we can use these objec­
tions to guide our analysis of identifcation. Car oll shows us that if we
want to explain what we mean when we say things like, 'I could really
identif with Ern Brockovitch,' or ' . . . with Donnie Darko,' or '. . .
with Atanauat the Fast Runner,' we will have to meet several require­
ments. We have to ( 1 ) locate the core meaning of the term 'identifcation' ;
(2) give a plausible account of this core meaning; ad, (3) show that
identifcation on this account is an imporant aspect of our engagement
with flm.
I mainative I dentifcati on
In fact, the frst of these requirements is easy to fulfl because we can
agree with Caroll that the core sense of 'identifcation' has to do with
emotiona duplication or empathy. All of the philosophical accounts of
identifcation that we will draw on in this chapter link identifcation and
empathy. On its own, the idea of empathy as emotional duplication
1 1 7
sounds mysterious and unlikely. It is therefore unsurprising that Car oll
would approach the whole question of identifcation with such scepti­
cism. But with the right account of how empathy comes about, we
can begin to appreciate the short-sightedness of Caroll's skepticism.
Indeed, the rest of this section will show tat the most substantial work
to be done in giving a philosophical account of identifcation is in
explaining the process behind empathy rather than empathy itself.
Among those cognitive flm theorsts and philosophers who refse
to dismiss the notion, Rchard Wollheim's account of central imagining
is regaded as the best resource for understanding chaacter identifca­
tion. 1 8
According to Wollheim, there are two vaeties of representa­
tional imagining that can occur in any mode of perception - central ad
acental imagining. In the visual mode, both cental and acental imagin­
ing involve, not just imagining that you see someting, but imagining
seeing - or visualizing - it. The diference between cental and acental
imagining is that central imagining involves imagining seeing from a
certain point of view, whereas acentral imagining involves imagining
seeing from no pacula point of view - in other words, without the
flter of a paicular person' s perspective. Thus, to use Wollheim's own
example, suppose that while reading Gibbon' s history of the Roman
Empire you conjure up an image of the celebrated entry of Sultan
Mahomet II into Constantinople. Either you can visualize te event
acentally - ' as sttched out, feze-like, the fa side of the invisible
chasm of history' 1 86
- or you ca visualize it centrally, from the perspec­
tive of one of the paticipants - say, the Sulta himself. In the latter case,
you a representing, for your own inner viewing pleasure, what the
Sultan saw of the event. As long as you kow enough about the Sultan
(say, fom Gibbon' S descrption) to infer how he would have experi­
enced the event, you can centrally imagine the Sultan' s (visual) experi­
ence as though it were your own.
Central imagining is thus imagining what another person is going
through as though you a the other person. It is not that you ae deluded
about really being this person. It is just that, in imagination, you shar
the other person' s experience. This helps us see why some philosophers
and flm theorists use Wollheim' s account of central imagining to explain
cinematic identifcaton. Skeptics who argue that identifcation is not pa
of flm engagement ofen take 'identifcation' literally and then point to
difculties wit the idea of imagining oneself identical with aother. 1 87
Wollheim's account of centa imagining lines up more closely with how
we actually use the ter 'identifcation' . By 'identifcation' we mean,
1 1 8
not ' melding' with a chaacter, but living wit her: shang her cares
ad coming to understand her in a paticulaly intimate way.
1 88
There ae
surely many things about another that we could only begin to understand
by fnding out, through an act of imagination, what it is like to have his
point of view. In tur, the understanding that natrally accompanies
cental imagining breeds care and concer.
Woll�eim's account of cental imagining provides the beginnings of
an explanation of the imaginative route to empathy. Having completed
his descrption of central imagining, Wollheim sums up by listing what
he taes to be the thee essential features of this phenomenon: point of
view, plenitude, and cogency. 1
9 The frst of these features has already
been mentioned. But what of plenitude and cogency? Plenitude descrbes
the way that my imaginative project tends to develop a life of its own so
that, by imagining doing and saying what the protagonist does and says,
I naturally end up imagining thinng and feeling as she most likely
thinks and feels. In tur, the cogency of cental imaginings depends on
their plentude and descrbes their special psychic power: By imagining
thinng and feeling as the chaacter thins and feels, there is every
chace that I will be left in the condition ¯ cognitive, conative, but
paticularly afective - in which the imagined mental states, were I actu­
ally to have them, would leave me. This suggests that a particula kind of
emotional state could follow from identifcation. More is needed, how­
ever, to confr that this state is empathetic.
In his aspectual' account of empathetic identifcation, Berys Gaut
provides this confrmation. 1
Gaut's account might be seen to follow
from an explication of Wolleim' s idea that cental imagining involves
imagining diferent components of another' s experence - percepts,
actions, thoughts, desires and feelings. Given the limits on just how
much we can imagine in one instance, Gaut argues that identifcation is
necessarly partia and can tae several forms. Insofa as I can imagine
perceiving, believing, desiring, or feeling what another perceives,
believes, desires, or feels, I can achieve perceptual, epistemic, motiva­
tional, or afective identifcation+ The vaety of forms that identifcation
can take is limited only by the variety of aspects of another's experience
I can imagine having. Moreover, diferent fors of identifcation can
combine simultaneously as well as foster one another.
Since some of our beliefs ae perceptual, just by imagining perceiving
what a character perceives, I achieve a degree of epistemic identifca­
tion. As well, however, just like actua perceiving supports beliefs that
ae not just perceptual, so imagined perceiving suppors imagined beliefs
1 1 9
that are not just perceptua. Once I have imagined seeing and believing
what the chaacter sees and believes, and given that the chaacter's
desires and emotions are based, in large pat, on what he sees and
believes, tere is then every chance that I will also achieve some degree
of motivational and af ective identifcation with the character. But, as we
know from Wollheim's emphasis on the feature of cogency, the course
of my engagement need not end here. Gaut argues from common experi­
ence that imagining feeling what the chaacter feels can foster actually
feeling what the chaacter feels. Any emotion ca be experienced empa­
thetically just as long as it is someone else' s in the frst place.
Empathetic I dentifcation
If empathy fundamentally involves imaginative identifcation, then we
can appreciate tat not every incidence of concurring emotion is going to
count as empathy. I mght happen to react in the same way to the same
event as someone else - as is often the case with fans at a sporting event
or members of a cinema audience. But this does not mean I am empatiz­
ing with the other person. For empathy to occur, I have to respond in the
same way as another because she responds that way. 1 91 There is thus a
special relationship between my response and the response of whomever
it is with whom I empathize. Forunately, we already have the resources
to explan this special relationship; namely, an account of imaginative
Insofar as empaty is not just the coincidence but the sharng of emo­
tion, it requires the imaginative adoption of the grounds of another' s
emotion, or the particular experience and perspective which gives rise
to just that emotion. For me to shae Mara's fear of the bear, it is not that
I have to be in the presence of the bear and judging the bear to be danger­
ous. Then, presumably, I would be having my own fear but not sharing
Maria' s. Rather, I need to imaginatively adopt Maria's experience of
te situation with the bear ¯ her belief about te dangerousness of the
bear; her desire to rn from the bear, and, indeed, her feelings about
the bear. If I can imaginatively adopt Maria's perspective in this way, and
if afective identifcation tends to foster empathy, then the relationship
between my response and Maria's would seem to be one of identity.
This brings us to another imporat distinction: The distinction
between empathy and sympathy. One handy way of explicating the dis­
tinction is to say that empathy is feeling with a chaacter and sympathy
1 20
is feeling for a chaacter. Thus I can have the sae kind of emotion both
empathetic ally and sympathetically - I can be happy with you and for
you. There is more to the distinction than this, however. A number of
theorists have pointed out tat empaty is a success term in a way that
sympathy is not.
1 92
An empathetic emotion can only really be empathetic
if the person with whom one empathizes is really feeling that emotion -
if you take someone to be feeling fear ad thus feel fea yourself when,
in fact, the other person is not feeling fear, then you have not suc­
ceeded in empathizing with him. In contast, whether or not I am
mstaen about what the other is feeling - and even if the other is not
feeling anyting - I can have a sympathetic response to him. A sympa­
thetic response does not depend on the other's actual emoton but on a
concered judgment of her situation. You fea for someone because you
judge him to be in danger ad this concers you whether or not the
person has realized the danger himself.
Thus even tough the terms, 'empathy' and ' sympathy' a often used
loosely and interchangeably in everyday life, there is still a substantial
distinction to be made between two categores of response to other, real
ad fctional, people. In fact, however, the tendency to use the terms
rather loosely may be taken to refect the way empathy can foster sympa­
ty and sympathy can foster empathy. Gaut maes the point that, insofa
as most people are concered for temselves, by empathizing with some­
one you may come to shae her concer. 1 93 And as we have seen,
sympathy depends on a concered judgment of another. In addition, Gaut
notes that by sympathizing with someone you may align yourself emo­
tionally with him to the point that you end up empathizing as well . 1 94
Now that we have an account of imaginative and empathetic identif­
cation, we can respond to Noel Caroll' s objections to the possibility
or likelihood of this phenomenon. Recal that Caroll agues that identi­
fcation cannot occur because emotional duplication canot occur. And
emotional duplication cannot occur because the viewer and the charac­
ter stand in such diferent relations to the events of the story that thei
emotions, . even when they ae of the same kind, are going to have a
diferent scope and direction.
What can we say to Caroll? First of all, we can say that his objec­
tions have failed to engage with or even acknowledge the imaginative
component of identifcation. Whatever Car oll has to say about emo­
tional duplication or empathy does not touch the plausibility of centra
imagining. As a result of ts oversight, however, Caroll' s objections
to empathy itself are undermined. Once you tae into account the
1 2 1
imaginative activity behind empathy, emotional duplication suddenly
makes sense. It is not, in fact, the case that the viewer's and the chaac­
ter's responses have a diferent direction and scope. When the viewer
centally imagines experiencing the story events as the chaacter does,
then the viewer's response is just like the character' s - it is based on the
same experence imaginatively realized. 1 95
Of course, this response assumes that the emotional duplication that
Caroll is objecting to is the kind based on imaginative identifcation.
In the literature on empathy and identifcation, there are at least two
other kinds of account of emotional duplication. The frst deals with
what is called afective mimicr. The second deals wit a diferent imag­
inative process than the one described above called simulation, which
is thought to suppor shared responses. Afective mimicry involves an
automatic response to the registering of emotional expression in others.
Tus when I see a close-up on screen of someone crying, I may involun­
taly and unconsciously adopt a simla expression - my facial muscles
tense in a certain way, my mouth turs downward, ad so on. My expres­
sion then has a 'feedback' efect: Suddenly I a feeling how I look. In
other words, I end up mimicking in feeling what I have subconsciously
taken to be expressed by the other person.
For this to occur, however,
we don' t need to know why the other person is feeling what they express.
But many philosophers think tat a feeling is not a fll-fedged emotion
unless it is felt for a reason. In other words, many philosophers thin tat
emotions involve a feeling that is caused by having a certain conception
of, or thought or belief about, one' s situation. Thus, although my mim­
icked state might feel like sadness, it is not sadness proper. Moreover, if
this is right, we cannot really refer to af ective mimicry as full emotional
duplication or empathy.
Simulation is a process of imaginatively 'tying on' another' s beliefs,
desires, or attitudes in order either to reach an understanding of her
curent behaviour or to predict her fture behaviour. There are close pa­
allels between this kind of account and our account of central imagining,
since, on both accounts, the imaginative adoption of another' s perspec­
tive leads to emotional engagement. Unfortunately, however, recent work
in the philosophy of mind has cast signifcant doubt on whether simula­
tion is really how we come to understand other people. Furhermore,
whether in the context of fction or real life, simulation only results in
imagined and not real emotions, 1 97
ad no explaation is given of the
possibility of moving fom imagined feeling to the actual feelings
involved in a empathetic response.
1 22
If imagiative identifcation is the sole or primary route to empathetic
engagement with flm chaacters, then its signifcance is secured. How­
ever, for some theorists, there is a furher reason for atempting to resolve
all the complications surounding the notion of identifcation, and this
is that an account can be given of the unique way that identifcation
contbutes to our understanding of chaacters and our own emotional
respo�ses to characters. This is because imaginative identifcation allows
us to adopt someone else' s perspective and respond from that perspec­
tive, thereby giving us some refective distance on our own perspective
and typical responses. 198 Although we do not have the space to flly
explore this idea here, it is important to mention because it may help
explain the observation we made at the beginning of the section - that the
possibility of identifcation can make the diference between a good ad
a bad flm.
Not all good narative flms suppor identifcation, however. Indeed,
some narative flms fail to support ay kind of emotonal engagement
with chaacters. This may be because a flm has no chaacters - as in the
case of Koyaanisqatsi ( 1 983). It may also be because the chaacters
a flm does have ae inaccessible and dispassionate - as in the case of
The Man Who Wasn't There (2001 ) . The fact that such flms can never­
theless be intensely moving suggests that there is more to our felt
engagement with flm than our engagement with flm chaacters.
It is not surrsing that so far most of the work on our felt engagement
with flm has focused on our responses to flm chaacters and narative
content more generally. After all, Hollywood-style flms tend to privilege
character and narative development. It is usualy the actions of charac­
ters that drve the plot forward, and our expectations about whether
the characters will meet thei goals deterine how we mae sense of
the plot. These expectations also gover our apprciation of the flm­
makers' stylistic choices in terms of narative signifcance. Nevertheless,
while the focus on chaacters and content may be understandable, it is
still unduly limiting. We not only respond to content but also to form or
style, and sometimes we respond to for regadless of content.
A flm's camerawork, lighting, editing, mise en scene, music, and
sound efects can be used to trigger emotional responses that, in t,
reinforce the signifcance of narative events. But even when what is
1 23
depicted has little or no narative signifcance, the way it is depicted
can have an emotiona efect - for example, in the way that rapidly
changing light gradients can trigger an immediate response of fea. I
general terms, however, responses to form alone still support responses
to content because the consistency of our responses to characters can
best be explaned in terms of the emotive function of form.
This brings us to what we shall call the consistency puzzle: How
does a flm reliably elicit the same responses from a diverse audience? I
real life we do not expect this to happen. If, for exaple, a couple is
arguing on a street comer, passers-by are likely to respond in any number
of ways. Some of them may be annoyed at the disturbance; others may
feel embar assed for the couple; still others may feel sor for one or
both members of the couple. Each kind of response rests on a diferent
interpretation of the event, infuenced by the personalityg background
assumptions and life experience of the person passing by. 19 But imagine
that you are watching the couple ague, not on the street corer but in a
flm. Now it seems that there is only one right way for everyone to
respond and this is in fact the way that most of us do responda How do
we explain this diference?
One way to solve the consistency puzzle involves an analysis of the
way that flms - paiculaly though their formal features - exploit or
appropriate the fnction of emotions. Noel Caroll agues that in rea
life, emotional responses fnction to focus our attention on or mae
salient those aspects of our situation that ae signifcat for our goals and
interests. Films can tae over this focusing function and thereby secure
our attention in just the way requied for a paicula intended emotional
response.20 Caroll is not alone in claiming that emotions have a focus­
ing fnction. This claim is also made by theorists who emphasize the
analogy between emotion and perception,201 and theorists who tn of
emotion as a physiological alerting mechanism.202 Moreover, fom frst
refection on our own emotional experience, it is not had to see why this
claim is commonly made. When I a araid, I am automatically focused
on what is fghtening and therefore threatening about my situation.
When I am angry, I am automatically focused on what is angering and
therefore unjust about my situation. This suggests to Car oll that if a flm
can establish the rght focus for a paricular emotion, then most viewers
will respond with just that emotion.
In order to illustrate Car oll' s claim, let's retur to the example of
the aguing couple. In the flm version of the agument, let's say we
a meant to feel indignant on behalf of one member of the couple.
1 24
The flm-makers will ten have set tings up so that our attention is
drawn to the character's decency and how she is being unfaily treated.
How the flm-makers do this depends on a myriad choices about how to
tell the story, who to choose for the role - perhaps an actor with an a of
vulnerability or one who tends to play decent and sympatetic charac­
ters, how to constct the dialogue, and how to shoot and edit the scene
of the agument so as to favour one chaacter - perhaps though the use
lose-ups, gentler lighting, and musical punctuation. Whereas in real
life, passers-by respond according to their own interpretations of the
situation, in te flm, viewers respond according to the flm's inter­
prettion of the situation. This interpretation is manifest in the flm's
formal features; in te way that the flm works according to the stylistic
choices of its makers.
If we al have the same interpretation of the aguing couple, as a
result of the way tat our att
ntion has been focused on cerain aspects
of the narative, Car oll claims it is not surprsing that we all tend to
respond in the way tat the flm intends. Thus the consistency puzzle is
solved. While this account is highly appealing, it seems to contain a
lacuna. This is because no explanation is given of why it is that, just
because emotions provide focus, when tis role is taen over by a flm,
the focus provided causes an emotion. An emotion' s cause ad its fnc­
tion a not one and the same, after all. But perhaps Caroll is simply
relying on the fact that tey are closely linked. The reason that emotions
function as they do is in pat because of the stadard way that they come
about. Thus, my anger is tiggered by a perceived slight, but once it has
been tiggered, my anger seres to focus my attention on features of my
situation which are relevant to having been slighted. If we ae already
focused on features of a flm scene that a relevat to the standad
tigger for a cerain emotion, perhaps tat emotion naturally tends to
follow. Once completed, this kind of explanation could, in tum, complete
Car oll' s account.
Another version of this solution to the consistency puzzle focuses, not
on the function of episodic, object-oriented emotional responses, but on
te function of moods that pervade a flm or signifcant portions of a flm.
Greg Smth agues that it is only by creating a mood that a flm can man­
tan a consistent emotional appeal and reliably evoke the same emotional
responses towads chaacters and events from diferent viewers. 203 Mood,
on Smith' s account, is a long-lasting and low-level feeling-state orient­
ing us towads stimuli for paicula emotional responses that in tum
sustain the mood. If I am in a fearl mood as I walk home on a dak
1 25
night, I am likely to notice frghtening aspects of my environment - say,
looming shadows and mysterious noises. In this way, the orientation of
my mood makes me prone to respond with fea to a certain object, which
in tum feeds back into my feal mood and sustans it well past the
subsidence of my particular fearfl rsponse. When I a in a fearful
mood, I am not afrad of anything in particula fom which I a moti­
vated to fee or hide. But once I am interpreting my environment fearflly,
I am more likely to have goa- and object-oriented responses of fea.
Even if, as Smit implies, every emotionally appealing flm must
sustain a mood, diferent flms or flm genes ca accomplish tis in
more or less nuanced ways. Most action-adventure flms, for example,
do not aim at emotional subtlety. The mood of a typical flm in this genre
tends to be overshadowed by dense, often redundat, and hghly fore­
grounded triggers or 'cues' for brief and focused emotional responses.
At stategic moments in the flm, tere will be a cluster of emotional cues
for, say, excitement and suspense - cues such as musical ' stingers' (sud­
den bursts of loud music), rapid cross-cutting or tacking shots, expressive
close-up shots, an increase in the intensity or contast of light, and an
increase in the pace of action or the animation of the chaacters. The
moments at whch these cue-clusters occur tend to be strategic for nara­
tivereasons. Indeed, one way in which emotiona cues ca beforegrounded
is by tying them to narative and chaacter development. Thus if we are
anxious to see the action hero escape fom some deathly predicament,
our emotional involvement in the moment can be heightened by the use
of paicula stylistic devices.
With his 'mood-cue approach' to solving the consistency puzzle,
Smith also provides an explanation of how we come to sha the overall
feel of a flm. Perhaps we ae ' stirred up' by an action-adventure flm
such that we leave the cinema in an excitable mood. On Smith' s account,
this is likely to be the result of our having been cued at stategic points in
the flm to have paicular emotional responses, which feed back into and
sustain a paticula mood. However, this leaves unanswered the question
of how we catch the mood of a flm in the frst place for it then to be
sustained by paicular emotional responses. More imporatly, it also
leaves unanswered the question of why it is that subtle flms with spaser
and less foregrounded emotional cues - for example, Toko Stor ( 1 953)
- ae ofen the ones most likely to leave us in a certain mood. Part of a
explanation could be that when we are not being bombarded by densely
foregrounded emotional cues, we ae beter able to appreciate the mood
1 26
behind the cues. But this does not explain why noticing the mood of a
flm makes it more likely for us to tae up that mood.
Perhaps the mechanism involved is simila to the one involved in
aspectual identifcation with chaacters: Insofa as the mood of a flm is
patly a matter of an attitde or point of view expressed in the overall
look (and sound) of te flm, we ca take up this attitude in imagination
and thus come to shae the feeling that it involves. On the other hand,
perhaps the mechanism involved is more visceral and automatic, involv­
ing less cognitive processing than identifcation. We will not decide tis
issue here, but it is certainly worh further tought and discussion as to
how flms not only stir us up in the moment but also have more lasting
emotional efects.
This brings us to the end of our critical surey of philosophica issues to
do with flm and feeling. Here is what we have covered along the way:
1 . A sketch of the traditional philosophical debate about the nature of
emotions, and the recent move to more unifed study of the full range
of felt responses.
2. Solutions to the paradox of fction, or how we can respond with genu­
ine emotion to what we know is not rea.
3. Solutions to the paradox of horor, or how we can enjoy the un-enjoy­
able emotions of fea and disgust in the context of horor fction.
4. An acknowledgment of the fact that, despite theoretical suspicion of
the notion, identifcation comes up regularly in ordinay discussions
about the flms we watch.
5. Metz's psychoanalytic account of primay identifcation with the
camera which fails to captre the notion of identifcation invoked in
ordinary discussions.
6. An aterative account of identifcation as a paicular kind of imagi­
native activity which may also provide a unique form of insight into
our own ad others' responses.
7. The possibility of empathy, and not just sympathy, with flm chaac­
ters through imaginative identifcation.
8. The implications of our responsiveness to flm form for the consist­
ency of our responses to content.
1 27
We were able to proceed with our survey of flm and feeling once we
had deterned that a solution to the paradox of fction will likely follow
from a deeper understanding of te nature of emotion. A solution to the
paadox of horor also seems to depend on such an understanding, and in
particula, on an understanding of what maes negative emotions nega­
tive such that they can still be enjoyed. There may be less reason,
however, for thinking that the paradox of horor is realy a paadox, given
that it is unclea even for the greatest horor buf what he is really enjoy­
ing. Since horor fls seem to be a case where our emotional responses
duplcate the chaacters' , ths led us into a discussion of identifcation
and empathy. We saw how the psychoaaytic notion of identifcation is
distinct from the commonsense notion, but that if we account for identi­
fcation as an imaginatve activity, the possibility of duplicating a
character' s emotion becomes intelligible. Finaly, by moving beyond the
emotional appeal of characters, we discovered that this appeal is best
explained in ters of the overall emotive stcture of a flm.
Perhaps this brngs us as fa as we can go with general theor, and
what remains is close analysis of the efects of paicular flms. Such a
analysis would surely give us an even greater appreciation of the extent
to which flm media are media of feeling.
1 28
1 Tese principles refect the genera anti-mimetic view that came to dominate
the a world ater the advent of non-objectve panting. At that time, may
crtics ad artists saw photography and flm as liberating pantng fom te
task of imtation. By completing ts task mechaically, however, flm and
photography tae the artisty out of the taditional aims of ar.
2 Scrton ( 1 995).
3 Ibid. , 89.
4 King ( 1 995).
5 This point is developed by Gaut (2002) in an extemely usefl crtque of
6 King ( 1 995: 1 1 8-20).
7 Aheim ( 1 957: 1 2).
8 Ibid. , 26.
9 Iid. , 14.
10 Iid. , 1 61 7.
1 1 Ibid. , 27-28
12 See Caroll ( 1 988a: 58-70) for an assessment of Aeim's ideas about ar
ad expression.
13 Aeim ( 1 957: 35).
14 Iid. , 58.
15 For mor examples of ts kind, see Ibid. , 6064.
16 Aeim's general point about the expressive use of the fame is illustrated in
ts way by Car oll in hs ( 1 988a: 41-42).
17 Aeim's contemporares, the Soviet-montage teorists, were paticulaly
interested in the way that editing, or montage, ca create atistic meaning and
expression. For more on this, see Chapter 4.
1 8 Again, see Caroll ( 1 988a: 58-75) for a critique of Aeim' s views on
19 Aeim ( 1 957: 1 06).
20 See Sesonske ( 1 973), ( 1 974), ad ( 1 980).
21 Sesonske ( 1 974: 52-53).
22 Ibid. , 54.
1 29
23 Talk of a flm 'rpresenting space and time' is really just shorthand for saying
that a flm represents the spatial and temporal relations between objects ad
events. As well, talk of motion in flm may or may not be literal, since there
is an ongoing debate about whether the motion we see on screen is real or
illusory. To le about this debate, see Chapter 2.
24 Sesonske ( 1 974: 54).
25 Ibid. , 54-56.
26 Ibid. , 56-57.
27 Ibid. , 57.
28 This point is also made by Caroll in hs ( 1 988a: 3 1-32).
29 For example, Car oll, in his ( 1 988a: 94), assigns singular importance to
Bazin. Andrew Sarris, whose version of auteur teor we will consider in
Chapter 3, also described Bazin as ' te greatest flm critic who ever lived' .
3 0 Walton ( 1 984).
3 1 Curie ( 1 995a).
32 In his Philosophy ofMotion Pictures (2008), Caroll gives a defnition of flm
as an art for parly in ters of the technical possibility of the 'promotion of
the impression of motion' (73).
33 Bazin ( 1 967).
34 This usefl sumar list is provided by Caroll in his ( 1 988a: 1 05-6).
35 Bazin ( 1 967: 1 3-14).
36 Ibid., 14.
37 Ibid. , 1 4-1 5.
38 Ibid. , 1 5.
39 Ibid., 1 4.
40 In Car oll ( 1 988a: 1 47-8).
41 Cure ( 1 995a: 75-76).
42 Walton ( 1 984: 250-5 1 ) .
43 Bazin ( 1 967: 14).
4 Walton ( 1 984: 252).
45 Ibid. , 261 .
46 Curie ( l 995a: 55).
47 Walton ( 1 984: 270).
48 Ibid o • 270-71 .
49 In paricular. Curie ( 1 995a) and Main ( 1 986).
50 Curie ( 1 995a). chapter 3.
51 Frday ( 1 996).
52 A theory according to which our everyday thought in some aa has ben taint­
ed by a msten philosophical view such that it involves systematic erors.
1 30
53 The image that is made visible on a flm screen or on photographic paper
is a real image whereas a image in a (fat) miror is a virual image. In the
photographic imageg te rays of light ae brought to a focus at te position of
te image. The mr or-image, on the other had, is made by rays of light tat
do not come from where the image seems to be. If I am stading two feet in
font of a mror, my image will appea 4 feet rater than 2 feet away from
me - that is, 2 feet behind or inside the mror.
54 Waburton ( 1 988).
55 Walton ( 1 997).
56 Frday ( 1 996: 36).
57 Ibid. , 39-40.
58 Many of the key features of Cure's account, including the notion of our
'recognitional capacities,' ae dawn from the work of Flint Shier. See Shier
( 1 986), paiculaly section 9. 3.
59 Cure ( 1 995a: 82-83).
60 Ibid. , 83.
61 Ibid. , 85.
62 Ibid. , 100.
63 Ibid. , 107.
64 This may be a over-simplifed account of projection, since each image is
often projected multiple times to reduce the ficker efect.
65 In what follows, I will refer to a paricular aricle - Cure ( 1 996) ¯ whch
most fully presents Curie's account. Curre also discusses his account in the
frst chapter of his ( 1 995a), and in the second section of his ( 1 997).
66 Caroll (2008: 87-93).
67 In Adelson' s checker shadow illusion, for example, we ae shown an image
of what appeas to be a black and white checker-boad with a green cylinder
resting on it that casts a shadow diagonally across the middle of the board.
The black ad white squares ae actually diferent shades of grey. The image
has been constructed so that ' white' squaes in te shadow, one of which is
labelled 'B,' ae actally te exact same grey value as 'black' squaes out­
side te shadow, one of which is labelled 'A' . The illusion created is that te
squaes A and B are diferent colours (or shades).
1 3 1
68 Cure ( 1 996: 336).
69 Ibid. , 248.
70 Kania (2002: 248).
71 Ibid. , 249.
72 Ibid.
73 Ibid. , 25 1 o
74 Ibid.
75 Car oll (2008: 92).
76 Currie ( 1 996: 326).
77 Quoted by St (2000: 85).
78 Livingston (2005: 68-69). Practically the same defnition is given by
Livingston in a earlier essay, ( 1 997).
79 Livingston (2005: 73).
80 Livingston ( 1 997: 1 36).
81 Ibid. , 1 35.
82 Livingston discusses i n detail the conditions on joint authorship i n his
(2005: 75-89).
83 Livingston ( 1 997: 1 35).
84 Nehamas ( 1 986: 689).
85 Livingston ( 1 997: 1 37). Given the vaiety of flm-screening technologies,
the curent form of this defnition may be too limited. Not all screened
flms ae projected - flms shown on television usually ae not, for example.
Neverheless, ther seems no reason why Livingston's defnition could not
be modifed to accomodate technological variation.
86 Livingston ( 1 997: 143-4).
87 Livingston (2005: 88).
88 Gaut ( 1 997: 1 57).
89 Gaut ( 1 997: 1 57-58).
90 Wollen ( 1 972a).
91 Iid. , 1 71-72.
92 Livingston ( 1 997: 145-6).
93 Sars ( 1 962/63: 1 05).
94 Ibid. , 1 07.
95 Kael ( 1 965).
96 Ibid. , 298.
97 Ibid. , 304.
1 32
98 As with the following examples: Spottiswoode ( 1 950); and, Arijon ( 1 976).
99 As with the following examples: Monaco (2000); and, McDonald (2005) .
1 00 Goodman ( 1 976).
1 01 Metz ( 1 974b: 92).
1 02 Ibid. , 1 07.
r03 Cure ( 1 995a: 1 1 7-1 8) .
1 04 Pudovkn ( 1 958: 1 21 ) .
1 05 Ibid. , 24.
1 06 Metz ( 1 974a: 67).
1 07 Metz ( 1 974c: 1 1 5).
1 08 Eisenstein ( 1 957: 29-30).
1 09 Metz ( 1 974a: 46).
1 1 0 Guzzetti ( 1 985: 1 77-93).
I I I Semiotic teorists can be called semologists, semioticists, or semioticians.
Given that 'semiologist' is often reserved for theorists workng stctly in
the tradition of Saussure, and given that 'semioticist' is the least common of
the thee terms, I prefer 'semiotician' .
1 1 2 Haan ( 1 999: 92-93).
1 1 3 Wollen ( 1 972b: 1 55).
1 14 Quoted by Haan in hi s ( 1 999: 95).
1 1 5 Ibid., 96.
1 1 6 Currie ( 1 995a: 1 21 -22).
1 17 See, for example, Eco ( 1 985).
1 1 8 The comparison between sentential connectives and shot/reverse-shot edit­
ing is made by Cure in hs (2006: 97-98).
1 1 9 Cure ( 1 995a: 1 35-36).
1 20 For a useful overview of the nature and value of narative, see Livingston
(2001 ) .
1 21 Arguably, stories a not so much told i n theate perforance as enacted.
The fact remains, however, that nar ative communication in theate, as in
flm, is both visual and verbal.
1 22 See Kawin ( 1 978).
1 23 See Wilson ( 1 997a: 295-3 1 8).
1 24 Ibid. , 31 0-1 1 .
1 25 See Chatman ( 1 990).
1 33
1 26 See Gaut (2004). It is also called the analyical agument by Kania in his
1 27 A version of this agument is put forard by Levinson in his ( 1 996).
1 28 In fact, it is not clea that the narator is always on the same ontological
level as the fctional characters and events. An omscient nar ator may be
considered to be 'outside' the story she is narating even though she is still
considered to be a fctional agent, par of the work of fction.
1 29 These objections to the agument from means of access are made by Kania
in his (2005).
1 30 See Wilson ( 1 997a: 298-300).
1 31 Wolterstorf( 1 980: 1 72).
1 32 For seminal work in the debate about fction, see Lewis ( 1 978) and Seale
( 1 979).
1 33 Cure ( 1 995b) .
1 34 Bordwell ( 1 985: 62).
1 35 For a detaled introduction to the histor and views of the group, see Erlich
( 1 98 1 ).
1 36 Bordwell ( 1 985: 49-51 ) .
1 37 Ibid. , 33-7.
1 38 Bordwell ( 1 985).
1 39 Bordwell ( 1 989: 8).
140 Ibid. , 8-9.
141 Wilson ( 1 997b).
142 Bordwell ( 1 985: 3 1 ).
143 Ibid. , 32.
14 Ibid. , 1 02.
1 45 This is suggested by Gaut in his ( 1 995).
146 Bordwell ( 1 989: 1 33).
147 Ibid., 3.
148 Bordwell and Thompson ( 1 979: 1 1 9).
149 Gaut uses ths example in his ( 1 995: 1 8).
1 50 See Bordwell 2008.
1 51 The idea that the flm viewer's seach for narative signifcance involves
the faming ad answering of questions is elaborated in a diferent context
by Noel Car oll. According to Car oll, the basic strcture of the most com­
mon form of flm narative is best conceived of as a network of questions
and answers. Thus we can understand why a certain scene in a flm follows
another scene if we think of the ealier scene generating a question about
1 34
the nar ative action that the later scene goes on to answer. For a -detaled
discussion of Caroll' s 'erotetc' model of movie naration, see his (1988b)
ad his (2008).
1 52 Car oll frst develops his theory of flm evaluation in hs (2003) and ten
refnes hs argument for te theory in chapter 7 of hs (2008).
1 53 Caroll (2003: 1 47-8).
1 54 Here Car oll is drawing on Walton ( 1 970) in which Walton arues that a
work's aestetic properties a to be found in a work when we perceive the
work in te corect categor. Caroll appropriates three of the four kinds of
reasons that Walton gives for a corect categorization. The fourth kind of
reason given by Walton tat Car oll ignores has t do with a work providing
the best knd of experience when perceived in a certain category.
1 55 Caroll (2003: 1 59-60).
1 56 See Frelad (2006) for some objectons that Caroll does not anticipate in
his earlier account of flm evaluation, (2003), but which he then addresses
in hs most recent account, (2008: 21 3· 1 7).
1 57 Carroll (2008: 21 3).
158 Ibid. , 21 9.
1 59 Ibid. , 220-21 .
1 60 Ibid. , 223.
1 61 Gaut ( 1 995: 21 -22).
1 62 Perkins ( 1 993: 1 20).
1 63 Ibid. , 1 3 1 .
1 64 Ibid. , 1 24-27.
1 65 Ibid. , 1 90.
1 66 For a recent perceptual theory of emoton, see Prinz (2004). Robers, in his
(2003), chaacterizes emotions in terms of constal, whch is analogous to
the phenomenon of seeing-as. De Sousa, in his ( 1 987), characterizes emo­
tons as ways of seeing that precede the foration of beliefs and desires.
1 67 See, for example, Robinson (2005).
1 68 Alteratvely, the paradox can be expressed, not in ters of how genuine
emotons can be evoked by fction, but in ters of whether rsponses t
fction really count as cases of genuine emotion. See Walton ( 1 990: 1 97),
Matavers ( 1 998), ad Neill ( 1 995).
1 69 For example, Radford has a whole seres of aricles on the iratonality of
our responses to fctions, beginning with his ( 1 975).
1 70 The idea tat we somehow suspend belief in te context of fction can be
attbuted to Coleridge, although he talked about suspending disbelief,
1 35
meaning a belief in someting that is not the case - that is, the reality of
fction. See his ( 1 951 ). Car oll discusses the suspension of (dis)belief in
his ( 1 990).
1 71 See Walton ( 1 990: 1 95-204, 241-55) for a famous version of this solution
that appeals to te notion of make-believe.
1 72 See Lamarque ( 1 981 ) as well as Car oll ( 1 990: 79-88).
1 73 Tis and other objections ae made by Walton in his ( 1 990: 202-3).
1 74 See Caroll ( 1 990) for a defnition of the horor genre in terms of fea and
disgust felt towards monsters.
1 75 This is the solution proposed by Car oll in his ( 1 990: 1 58-95).
1 76 This expressivist solution is descrbed though not endorsed by Gaut in his
( 1 993).
1 77 Versions of te contol soluton are developed in Eaton ( 1 982) and in
Morreall ( 1 985).
178 This example is adapted fom Gaut ( 1 993: 339).
1 79 See Gaut ( 1 993).
1 80 See Caroll (2008) for an account of our engagement with flm chaac­
ters that takes a diferent focus. In chapter 6, Carroll argues that sympaty
'constitutes the major emotive cement between audiences and the perinent
movie characters' ( 1 78).
1 81 Mulvey (2000).
1 82 Metz's account of cinematc identifcation is developed in hs ( 1 982). For
a helpful explication and crtique of this account, see chapter 1 of Carroll
( 1 988a).
1 83 Another reason to worry about the psychoanalytic account is tat it implies
tat cinematic identifcation is pathological and thus undesirable or even
dangerous. This implication sprngs fom the fact that te viewer who iden-
tifes, according to Lacanian-inspired flm theorsts, is also a fetishist and a
voyeur. According to Metz, the viewer fetishizes te material design of the
flm to compensate for the absence of what she is in fact most interested
in; namely, what is shown on screen. Given this absence, there is no pos­
sibility of involvement with or detection by the flm characters: The viewer
becomes a passive and rather sneay spectator - in other words, a voyeur.
1 84 Caroll ( 1 990: 89).
1 85 See Wollheim ( 1 984).
1 86 Iid. , 72.
1 87 See, for example, Carll ( 1 990: 88-96).
1 88 This knd of 'interal' understanding is discussed in Keran ( 1 996) and in
Neill ( 1 996).
1 89 Wolleim ( 1 984: 79-80).
1 90 See Gaut ( 1 999), paticularly 20616.
1 91 This point is made by Neill i n hi s ( 1 996).
1 36
192 Tese teorists include Feagin ( 1 988) and Neill ( 1 996).
1 93 Gaut ( 1 999: 207).
1 94 Ibid. , 21 1 .
1 95 Te same kind of rsponse to Caroll is given by Gaut in his ( 1999: 207-8).
1 96 See Smth ( 1 995: 98-1 01 ).
1 97 See Feagin ( 1 988).
198 Both Neill, in his ( 1 996), and Gaut, in his ( 1 999), discuss te lin between
identifcation and understanding.
1 99 This example is adapted from Car oll (2008: 1 56-7).
200 See, once again, chapter 6 of Caroll (2008).
201 See, again, de Sousa ( 1 987) ad Roberts (2003).
202 This is a comon view aong psychologists and evolutiona biologists
but it is also shaed by Robinson (2005).
203 See Smith ( 1 999) and (2003).
1 37
Ajon, D. ( 1 976), Grmmar of the Film Lnguage (London and Boston: Foca
Amheim, R. ( 1 933), Film, t. L. M. Sieveking and I. F D. Morow (London:
Faber and Faber).
Aheim, R. ( 1 957), Film as Art (Berkeley: University of Califoria Press).
Bazin, A. ( 1 967), What is Cinema? ed. ad tr. H. Grat (Berkeley ad Los
Angeles: University of Califoria Press).
Bordwell, D. (2008), Poetics ofCinema (New York: Taylor ad Francis).
Bordwell, D. ( 1 985), Narration in the Fiction Film (Madison: University of
Wisconsin Press).
Bordwell, D. ( 1 989), Making Meaning: Infernce and Rhetoric in the Interreta­
tion ofCinema (Cambridge, MA: Havad University Press).
Bordwell D. and Thompson, K. ( 1 979), Film Art: An Intrduction (New York:
Alfred A. Knopf).
Caroll, N. ( 1 988a), Philosophical Prblems of Classical Film Theor (Princeton:
Princeton University Press).
Carroll, N. ( 1 988b), Mystiing Movies: Fad and Fallacies in Contemporar
Film Theor (New York: Columbia University Press).
Caroll, N. ( 1 990), The Philosophy ofHorror or Paradoxes ofthe Heart (London
and New York: Routledge).
Car oll, N. (2003), 'Introducing Film Evaluation' , in Engaging the Movie Image
(pp. 147-64) (New Haven and London: Yale University Press).
Car oll, N. (2008), Philosophy ofMotion Picturs (Malden, MA: Blackwell).
Chatman, S. ( 1 990), Coming to Ters: Te Rhetoric of Narrative in Fiction and
Film (Itaca, NY: Corell University Press).
Coleridge, S. T. ( 1 95 1 ), Biographia Literaria in D. Staufer (ed. ), Selected Poetr
and Prose ofColeridge (New York: The Moder Libray, Random House).
Cure, G. ( 1 995a), Image and Mind: Film, Philosophy, and Cognitive Science
(New York: Cambrdge University Press).
Currie, G. ( 1 995b), 'Unreliability Refgured: Nar ative in Literature and Film' ,
Joural ofAesthetics and Art Crticism 53. 1 , 1 9-29.
Cure, G. ( 1 996), 'Film, Reality, and Illusion' , in D. Bordwell and N. Caroll
(eds), Post Theor: Reconstrcting Film Studies (pp. 325-4) (Madison:
University of Wisconsin Press).
Cure, G. ( 1 997), 'The Film Theory that Never Was: A Nerous Maifesto' ,
in R. Allen and M. Smth (eds), Film Theor and Philosophy (pp. 42-59)
(Oxford: Oxford University Press.
1 38
Currie, G. (2006), 'The Long Goodbye: The Imaginar Laguage of Film' , in
N. Caroll and 1. Choi (eds), Philosophy of Film and Motion Pictres: An
Anthology (pp. 91 -99) (Malden, MA: Blackwell).
De Sousa, R. ( 1 987), The Rationalit of Emotion (Cambridge, MA: MIT
Eaton, M. ( 1 982), 'A Stange Knd of Sadness' , Joural ofAesthetics and An
Criticism 41 , 5 1-64.
Eco, U. ( 1 985), ' On the Contribution of Film to Semiotcs' , in G. Mast and
M. Cohen (eds), Film Theor and Criticism: Intrductor Readings, Third
edition (pp. 1 94-21 4) (New York and Oxford: Oxford University Press).
Eisenstein, S. ( 1 957), 'The Cinematic Prnciple ad the Ideogram' , Film For
and Film Sense, ed. and t. 1. Leyda (pp. 28-4) (New York: Meridia).
Erlich, V ( 1 981 ), Russian Foralism: Histor, Doctrine, T editon (New
Haven: Yale University Press).
Feagin, S. ( 1 988), 'Imagining Emotions and Appreciating Ficton' , Canadian
Jourl of Philosophy 1 8. 3, 485-500.
Freeland, C. (2006), 'Evaluating Film' , Film Studies 8, 1 5460.
Friday, 1. ( 1 996), 'Transpaency and te Photogaphic Image' , British Joural
ofAesthetics 36. 1 , 30-2.
Gaut, B. ( 1 993), 'The Paadox of Horor' , British Joural ofAesthetics 33.4,
Gaut, B. ( 1 995), 'Makng Sense of Films: Neoformalism and its Limits' , Forum
for Modem Lnguage Studies XXXI. l , 8-23.
Gaut, B. ( 1 997), 'Film Authorship ad Collaboration' , in R. Allen and M. Smith
(eds), Flm Teor and Philosophy (pp. 1 49-72) (Oxford and New York:
Oxford University Press).
Gaut, B. ( 1 999), 'Identifcation and Emotion in Narative Film' , in C. Plantnga
and G. M. Smth (eds), Passionate Vews: Film, Cognition, and Emotion
(pp. 200-1 6) (Baltimore ad London: Johns Hopkins University Press).
Gaut, B. (2002), 'Cinematic A' , Jourl of Aesthetics and An Criticism
60.4, 299-3 1 2.
Gaut, B. (2004), 'The Philosophy of the Movies: Cinematic Nar aton' , in
P. Kvy (ed. ), The Blackell Guide to Aesthetics (pp. 230-53) (Malden,
MA: Blackwell).
Goodma, N. ( 1 976), Lnguages ofAn (Indianapolis: Hackett).
Guzzet, A. ( 1 985), 'Chstian Metz and the Semiology of the Cinema' , in
G. Mast and M. Cohen, Film Teor and Criticism: Intrductor Read­
ings, Third edition (pp. 177-93) (New York ad Oxford: Oxford University
Haman, G. ( 1 999), ' Semiotics and the Cinema: Metz and Wollen' , in L. Braudy
and M. Cohen (eds), Film Theor and Criticism: Intrductor Readings,
Fifth edition (pp. 90-8) (New York and Oxford: Oxford University Press).
1 39
Kael, P. ( 1 965), 'Circles and Squares: Joys and Saris' , in I Lst It at the Movies

. 292-3 1 9) (Boston: Little, Brown, and Company).
Kania, A. (2002), 'The Illusion of Realism in Film' , British Joural ofAesthetics
42. 3, 243-58.
Kania, A. (2005), 'Against the Ubiquity of Fictional Narators' , Joural of
Aesthetics andArt Criticism 63. 1 , 47-54.
Kawin, B. ( 1 978), Mindcreen: Berman, Godard, and First-Person Film
(Pnceton, NJ: Prnceton University Press).
Keran, M. ( 1 996), 'A, Imaginaton, and the Cultivation of Morals' , Joural of
Aesthetics and Art Criticism 54.4, 337-5 1 .
King, W. ( 1 995), ' Scrton and Reasons for Looking at Photographs' , in A.
Neill and A. Ridley (eds), Arguing about Art: Contemporar Philosophical
Debates (pp. 1 1421 ) (New York: McGraw-Hill).
Lamaque, P. ( 1 981 ), 'How Can We Pity and Fea Fictons?' British Joural of
Aesthetics 21 .4, 291-304.
Levinson, J. ( 1 996), ' Film Music and Narative Agency' , in D. Bordwell and
N. Caroll (eds), Post-Teor: Reconstructing Film Studies (pp. 248-82)
(Madison: University of Wisconsin Press).
Lewis, D. ( 1 978), 'Trth in Fiction' , Amercan Philosophical Quarerly 1 5,
Livingston, P. ( 1 997), 'Cinematic Autorship' , in R. Allen and M. Smith (eds),
Film Theor and Philosophy (pp. 1 32-8) (Oxford and New York: Oxford
University Press).
Livingston, P. (2001 ) , ' Nar atve' , in B. Gaut and D. M. Lopes (eds), The
Routledge Companion to Aesthetics (pp. 275-84) (London and New York:
Livingston, P (2005), Art and Intention: A Philosophical Study (Oxford,
Clarendon Press).
Marin, E. ( 1 986), 'On Seeing Walton' s Great-Grandfather' , Crtical Inquir 1 2,
Matravers, D. ( 1 998), Art and Emotion (Oxford: Claendon Press).
McDonald, K. (2005), Reading A Japanese Film: Cinema in Contet (Honolulu:
University of Hawaii Press).
Metz, C. ( 1 974a), 'The Cinema: Language or Language System?' , in Film
Lnguage: A Semiotics of the Cinema, t. M. Taylor (pp. 31 -91 ) (New York:
Oxford University Press).
Metz, C. ( 1 974b), ' Some Poits in the Semiotics of the Cinema; Film Lnguage:
A Semiotics ofthe Cinema, t. M. Taylor (pp. 92-1 07) (New York: Oxford
University Press).
Metz, C. ( 1 974c), 'Problems of Denotation in the Fiction Film' , in Film Ln­
guage: A Semiotics of the Cinema, t. M. Taylor (pp. 108-6) (New York:
Oxford University Press).
Metz, C ( 1 982), T Imaginar Signier: Pschoanalysis and the Cinem, t.
C. Brtton, A. Williams, B. Brewster, and A. Guzzett (Bloomington: Indi­
ana University Press).
Monaco, 1. (2000), How to Read A Film: The World ofMovies, Media, Multime­
dia: Lnguage, Histor, Theor, Thrd edition (New York: Oxford Unver­
sity Press)
Moreall, 1. ( 1 985), 'Enjoying Negatve Emotions in Fictions' , Philosophy and
Literature 1 9, 95-1 03.
Mulvey, L. (2000), 'Visual Pleasure and Nar atve Cinema' , in E. A. Kapla
(ed. ), Feminism and Film (pp. 347) (Oxford University Press).
Nehamas, A. ( 1 986), 'What an Author Is" Joural of Philosophy 83. 1 1 ,
685-91 .
Neill, A. ( 1 995), 'Feelings and Fictions' , in A. Neill and A. Rdley (eds), Aruing
about Art: Contemporar Philosophical Debates (pp. 1 75-93) (New York:
Neill, A. ( 1 996), 'Empathy and (Film) Fiction' , in D. Bordwell and N. Caroll
(eds), Post-Theor: Reconstrcting Film Studies (pp. 1 75-94) (University
of Wisconsin Press).
Perkns, V. F ( 1 993), Film as Film: Understanding and Judging Movies (New
York: De Capo Prss).
Prnz, J. (2004), Gut Reactions: A Pereptual Theor of Emotion (New York:
Oxford University Press).
Pdovkn, V. I. ( 1 958), Film Technique and Film Acting, t. and ed. I. Montagu
(New York, Grover Press).
Radford, C. ( 1 975), 'How Can We Be Moved by the Fate of Anna Karenina?
Pa 1 ' , Aristotelian Societ: Supplementar Volume 49, 67-80.
Roberts, R. C (2003), Emotions: An Essay in Aid of Moral Psychology
(Cambridge and New York: Cambridge University Press).
Robinson, R. (2005), Deeper than Reason: Emotion and its Role in Literature,
Music, and Ar (Oxford: Clarendon Press).
Sars, A. ( 1 962163), Notes on the 'Auteur Theor in 1 962' , Film Culture 27,
Scher, F ( 1 986), Deeper into Picturs: An Essay on Pictorial Representation
(Cambrdge: Cabridge University Press).
Scrton, R. ( 1 995), 'Photography ad Representation' , in A. Neill and A. Ridley
(eds), Arguing about Ar: Contemporar Philosophical Debates (pp. 89-1 13)
(New York: McGraw-Hill).
Searle, 1. ( 1 979), 'The Logical Status of Fictonal Discourse' , in Expression and
Meaning: Studies in the Theor of Speech Acts (pp. 58-75) (Cambridge:
Cambrdge University Press).
Sesonske, A. ( 1 973), 'Cinema Space' , in D. Carand E. Casey (eds), Eplora­
tions in Phenomenology (pp. 399-409) (The Hague: Marnus Nijhof.
1 4 1
Sesonske, A. ( 1 974), 'Aesthetics of Film, or A Funny Ting Happened on the
Way to the Movies' , Joural of Aesthetics and Ar Criticism 33. 1 , 51 -7.
Sesonske, A. ( 1 980), 'Time and Tense i n Cinema' , Joural of Aesthetics and
Ar Criticism 38.4, 41 9-26.
Smith, O. M. ( 1 999), 'Local Emotions, Global Moods, and Film Stcture' , in
C. Plantinga ad O. M. Smith (eds), Passionate Vews: Film, Cognition, and
Emotion (pp. 1 03-26) (Batimore: Johns Hopkins Universit Pess).
Smth, O. M. (2003), Film Structure and the Emotion System (Cambridge and
New York: Cabrdge University Press).
Smith, M. ( 1 995), Engaging Characters: Fiction, Emotion, and the Cinema
(Oxford: Claendon Press).
Spottiswoode, R. ( 1 950), A Grammar of the Film: An Analysis ofFilm Technique
(Berkeley: University of Califoria Press).
Stam, R. (2000), Film Teor: An Introduction (Malden, MA: Blackwell).
Walton, K. ( 1 970), 'Categores of A' , Philosophical Review 79, 334-67.
Waton, K. ( 1 984), 'Transpaent Pictres: On the Nat of Photographc Realsm'
Crtical Inquir 1 1 , 246-77.
Walton, K. ( 1 990), Mimesis as Make-Believe: On the Foundations of the Repre­
sentational Arts (Cambridge, MA: Haard University Press).
Walton, K. ( 1 997), 'On Pictres and Photographs: Objections Answered' , in
R. Allen ad M. Smt (eds), Film Teor and Philosophy (pp. 60-75)
(Oxford: Claendon Press).
Warburon, N. ( 1 988), ' Seeing Through "Seeing Tough" Photographs' , Ratio
1 , 62-74.
Wilson, O. ( 1 997a), 'L Grand Imagier Steps Out: The Primitive Basis of Film
Naration' , Philosophical Topics 25, 295-31 8.
Wilson, O. ( 1 997b), 'On Film Narative ad Nar ative Meaing' , in R. Allen
and M. Smth (eds), Film Theor and Philosophy (pp. 221 -38) (Oxford:
Claendon Press).
Wollen, P ( 1 972a), 'The Auteur Teor' , in Signs and Meaning in the Cinema
(pp. 74-1 1 5) (Bloomington: Indiaa Unversity Press).
Wollen, P. ( 1 972b), 'The Semology of the Cinema' , in Signs and Meaning in
the Cinema (pp. 1 1 6-55) (Bloomington: Indiana University Press).
Wollheim, R. ( 1 984), Te Tread ofLife (Cambridge,
A: Harad University
Wolterstorf , N. ( 1 980), Works and Worlds of Art (Oxford: Oxford University
1 42
A Beautil Mind 8 1
and authorship 49-50
Adams, A.
'Moon and Half Dome' 7
afect - see 'feeling'
afective mimcry 1 22
Aldrch, R. 41
American Beaut 77
Antonioni, M. 54
Apocalypse Now 56
Arijon, D. 57n
Aheim, R. 2, 3, 20
against sound flm 1 1-1 2
on a as expression 1 0
on flm as an a 8-1 2
auteur 40-2, 4, 502, 53-5
auteur-strcturalism 41 , 50-1
auteur theory 53-5
auteurist critics 41 , 53, 91
defnition of 45
authorship in achitectre 49
authorship in flm
three knds of claim 42
two philosophical conceptions 47
authorship in paintng 49
Barthes, R. 42
Bazin, A. 3
and Cahiers du Cinema 40-1
on the nature of a 21
on realism 1 6-17, 1 8-23, 24
on the realist style 1 8-20, 21 -2,
Bergman, I.
Winter Light 48, 52-3
and the historical poetics of
flm 93-
on interretation 91-, 1 01-2
on nar ative comprehension 83-5,
87-91 , 1 01
Bresson, R. 41
Burbage, R.
as Othello 49
Cahiers du Cinema 40-1 , 53
Caroll, N. 1 8n
against identifcation 1 1 5n,
1 1 6-17, 1 1 8n, 1 21-2
on Areim l On, l I n, 1 6n
on Bazin and fction flm 22-3
on evaluation 96-9, 1 02
on motion in flm 345, 37-8
on naration 94n
on suspense 1 1 n
on sympathy 1 1 4n
on the consistency puzzle 1 245
on the paadox of fction 1 06n,
1 08n
on the paadox of horor l I On,
I 1 1 n
CGI (computer-generated
imagery) 1 5
Chaplin, C. 3
Chatma, S.
on flm naration 76, 83-
common usage of the ter 1
Citizen Kane 16, 78, 93, 99
in flm 59, 63-5
Coleridge, S. T. 1 06n
1 43
collaboration in flm-making
implications for authorship 40,
42-3, 44, 49, 5 1 , 55-6
consistency puzzle 1 246
and the laguage of flm 67-8,
critic, role of 42, 44, 54-5, 96, 98
academc 91-3, 94
theory of
1 , 52-5 - see also
Cruching Tiger Hidden Dragon 1 4
Currie, G. 1 7
against photographic
transpaency 27n, 27-8, 30
on depiction 3 1-3, 39
on motion in flm 34-6, 37, 39
on narative unreliability 82-3
on te language of flm 65-71
Day-Lewis, D.
in Tere Will be Blood 49
de Saussure, F 58, 62-3
de Sousa 1 04n, 1 24n
depiction 3 1-3
digital fl 1 8
as author 41 -2, 43, 44,
53-, 55
DVDs (digital video discs)
and authorship 56
Eaton, M. l l 1 n
Eco, U. 67
editing 7, 1 1 , 1 4 - see also 'montage'
and the laguage of flm 60-2
in Soviet flm 2
Eisenstei, S.
Battleship Potemkin 2
October: Ten Days that Shook the
World 61-2
on the language of flm 60, 61 -2
and cognition 1 04
nature of 1 03-
ad identifcation 1 1 7-1 8,
1 20-2
Erlich, V. 84n
eror theory 29, 29n, 3 1
evaluation 95-1 01 , 1 02
Feagin, S. 1 22n
feeling 1 03
femnist crtique
of identifcation 1 1 4-1 5
nature of 79, 79n
fction flm 72
Bazin's difculty with 22-3
tee senses of the term 1 -2
flm teor vii-viii
classical vii, 1
cognitive viii, 1 1 6
psycho-semiotic vii, 17, 57, 1 14
seventies - see 'psycho-semotic'
Ford, John 50, 53
Te Searhers 51
form of a flm
and content 1 23-4
and feeling 1 23-7
Foucault, M. 42, 45
faming of the shot
ad motion 14
and nar ation 75
expressive function of 1 1
Friday, J.
against photographic
taspaency 28-9, 3 1
theory of perception 28, 3 1
Gaut, B.
against Scrton 7n
on empathy ad sympaty 1 21
on flm authorship 44, 49-50,
5 1-2
on flm naration 76n
on identifcation 1 1 9-20, l 22n,
1 23n
on interpretation 92-3
on the paadox of horor I 1 1 n,
1 1 1-12
Geh, F 49
German expressionist flm 22
The Cabinet of Dr Caligari 2
Gibson, R.
' Te Priest' 7-8
Godad, J. -L. 64
Goodman, N. 57
Guzzetti, A. 62
Harman, G. 64, 65
Hawks, H. 53
Hitchcock, A. 41
and empathy 1 1 3-23
and evaluation 1 1 3, 1 23
and imagination 1 1 8-20,
1 21 -2, 1 28
and understanding 1 1 9, 1 23
psychoanalytic account of
1 1 4-16, 1 1 6n, 1 28
of colour 34, 34n
of motion in the flm image 34-8
Mtller-Lyer 35-6
imagination, role of
in identifcation - see 'identifcation
and imagination'
in narative comprehension 90-1
implied author 40, 44, 46-7, 52, 55,
56, 75-7
and authorship 47, 5 1-2
ad meaning 88
and nar ation 8 1 , 83
Kael, P.
against auteur teor 545
Kania, A.
on nar ation 76n, 78n
on motion in flm 367
Kawin, B. 75n
Kieran, M. 1 19n
King, W.
on te a status of photography
Klein, W
'Entrance to Beach, Ostia, Italy,
1965' 7
Koyaanisqatsi 1 23
Laca, J.
psychoanalytic theor 1 141 5
Lamaque, P.
on te paadox of fction 1 08n
language of flm
queston of 58
Lst Year at Marienbad 79
Lwrence ofArbia 67
Levinson, J.
on flm nar ation 77n
Levi-Strauss, C. 50
Lewis, D. 79n
linguistics 58
Saussuran 63
litera paadigm, infuence on
flm authorship 40, 44, 45
flm nar ation 72, 73
the language of flm 57
Livingston, P.
on authorship 4, 45n, 46n, 47
on flm authorship 47-50, 52-3
low-angle shots 1 0, 93
1 45
Main, E.
against photogaphic
tanspaency 27n, 29
Matravers, D.
on the paradox of fction 1 06n
McDonald, K. 57n
meaing in flm 87, 94
ad the language of flm 58,
60-2, 66-71
constuction of 88-93
medium-specifc analysis 3-, 1 2
Metz, C.
on identifcation 1 1 5, 1 1 5n
on the language of flm 58, 61 ,
62-5, 66, 67
Monaco, J. 57n
montage - see also 'editing'
ad the language of flm 602, 63
Soviet or 'intellectual' 2, l I n, 22,
style 19, 33
mood 1 03, 1 25-7
Moreall, J. 1 1 1 n
as a foral category of flm 1 3-14
reality of in the fl m image 33-8
Mulvey, L.
on identifcation 1 14-1 5
nar ation, flm
three main questions 73
nar ative
chaacterized 72-3
comprehension 83-5, 87-91
flm's tendency towad 62, 63
unrelability 80-3
nar ator
chaacterized 74
knds of 745
necessity of 74-9
Nehamas, A.
on literay authorship 45, 467,
50, 5 1-2
Neill, A.
on empathy and
identifcation 1 1 9n, 1 20n, 1 23n
on the paadox of fction 1 06n
New World 2
and realism 20, 24-6
as representation 5-6, 30
paadox of fction 105-9, 128
characterized 1 06
paadox of horor 1 1 01 3, 1 28
chaacterized 1 1 0
paadox of tragedy 1 1 0
as constructive 88-9, 92
dict realist theory of 28, 3 1
representationalist causal theor
of 31
Perkns, V.
on evaluation 99-100, 1 02
phobias 108
and reaism 1 7-1 8
as representations 48, 23, 30
Pierce, C.
theory of signs 64
point-of-view shot 70
post-stucturalism 42
Prinz, J. 1 00n
projection, mechanism of 34, 34n,
psychoanalysis 91 - see also
'Lacan, J.'
Pdovkin, V.
on the language of flm 59, 60-1
Radford, C.
on the paadox of fction 1 06n
Rashomon 93
Ray, N. 41
realism of flm 1 6-39
photographic 1 7-1 8
1 46
recording, mechanical
of real life or of teatre 4
signifcance for the at stats
of flm 2-3, 20-1
refex responses 1 03
Renoir, J. 33, 41
Rules of the Game 16, 1 8-20,
Roberts, R. 1 04n, 1 24n
Robeson, P.
as Othello 49
Robinson, J. 104n, 1 24n
Rope 99-1 00
Rossellini, R. 1 9
Rules of the Game - see 'Renoir, J.'
Russian formaists 84, 90
Sarris, A. 1 6n
on auteur theory 41 , 53-
Schier, F 32n
Scrton, R.
against photography and flm as
representationa arts 4, 1 5
Searle, 1. 79n
semiotics 41 , 4, 59
ad flm 59, 62-5
serally maufactured flms
authorship of 43, 48
Sesonske, A.
on flm as an independent a
1 2-1 5
Shakespeare, W
Othello, chaacter of 49
shotreverse-shot editing 70
silent flm 2, 1 1
simulation theor 1 22
Sixth Sense 8 1
Smith, G.
on the consistency puzzle 1 25-7
Smith, M. 1 22n
signifcace of introduction to
flm 1 6-17, 33
as a foral categor of fl 1 3-14
depiction of 33
Spottiswoode, R. 57n
Stage Fright 81
structuralism 41 , 42, 44
creation of 1 1 , 99-100
sympathy, compaed with
empathy 1 1 3, 1 20-1
television 37-8
The Ludest Whisper 99-100
The Man Who Wasn 't There 1 23
The Seven Samurai 2
theate, compaed with flm
nar ative 74, 74n
ontology 49-50
reaism 1 7
staging 6
as a foral category of flm 1 3-14
depiction of 33
Toko Stor 1 26
transpaency, photographic 25-8
skepticism about 28-3 1
Trfaut, F 40-1 , 53
Walsh, R. 53
Walton, K. 97n
on fction flm 23
on photographic transpaency 17,
23-8, 29, 39
on the paadox of fcton 1 06n,
1 09n
Waburon, N.
against photographic
transparency 29
Welles, 0. 1 8-19, 41 - see also
' Citizen Kane'
Wilder, W 1 9
Wilson, G.
on flm naration 75n, 78n
1 47
Wlson, G. (Cont 'd
on interretation 88, 93-5, 1 02
on nar ative unreliability 81 -2
Wolleng P.
on flm authorship 4, 50-1
on the language of flm 645
Wollheim, R.
on imaginaton 1 1 8-1 9
You Only Live Once 8 1-3
1 48

Sign up to vote on this title
UsefulNot useful