Handbook of Visual Analysis 1st Edition Theo Van

Leeuwen pdf download

https://ebookgate.com/product/handbook-of-visual-analysis-1st-
edition-theo-van-leeuwen/

Get Instant Ebook Downloads – Browse at https://ebookgate.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Introducing Social Semiotics Theo Van Leeuwen

https://ebookgate.com/product/introducing-social-semiotics-theo-van-
leeuwen/

ebookgate.com

Hipparcos the New Reduction of the Raw Data Astrophysics

and Space Science Library 1st Edition Floor Van Leeuwen

https://ebookgate.com/product/hipparcos-the-new-reduction-of-the-raw-
data-astrophysics-and-space-science-library-1st-edition-floor-van-
leeuwen/
ebookgate.com

Visual Ergonomics Handbook 1st Edition Jeffrey Anshel

https://ebookgate.com/product/visual-ergonomics-handbook-1st-edition-
jeffrey-anshel/

ebookgate.com

The Conference of the Tongues 1st Edition Theo Hermans

https://ebookgate.com/product/the-conference-of-the-tongues-1st-
edition-theo-hermans/

ebookgate.com
Multilevel Analysis of Individuals and Cultures 1st
Edition Fons J.R. Van De Vijver

https://ebookgate.com/product/multilevel-analysis-of-individuals-and-
cultures-1st-edition-fons-j-r-van-de-vijver/

ebookgate.com

handbook of detergents analysis 1st Edition Heinrich

Waldhoff

https://ebookgate.com/product/handbook-of-detergents-analysis-1st-
edition-heinrich-waldhoff/

ebookgate.com

Handbook of Forensic Drug Analysis 1st Edition Fred Smith

https://ebookgate.com/product/handbook-of-forensic-drug-analysis-1st-
edition-fred-smith/

ebookgate.com

Handbook of biomedical image analysis 1st Edition David

Wilson

https://ebookgate.com/product/handbook-of-biomedical-image-
analysis-1st-edition-david-wilson/

ebookgate.com

Handbook of Stochastic Analysis Applications 1st Edition

V. Lakshmikantham

https://ebookgate.com/product/handbook-of-stochastic-analysis-
applications-1st-edition-v-lakshmikantham/

ebookgate.com
VA00post 14/9/00 10:24 am Page i

HANDBOOK OF VISUAL ANALYSIS

VA00post 14/9/00 10:24 am Page ii
VA00post 14/9/00 10:24 am Page iii

1
2
3
4
5
6
7
8
9
0
11
12 H A N D B O O K O F V I S U A L A N A LY S I S
13
14
15
16
17 EDITED BY
18
19 THEO VAN LEEUWEN AND CAREY JEWITT
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44 SAGE Publications
45 London ● Thousand Oaks ● New Delhi
Replacement title-verso page

ISBN 0-7619-6476-2 (hbk)

ISBN 0-7619-6477-0 (pbk)
© Theo van Leeuwen and Carey Jewitt 2001 Compilation and Introduction
© Philip Bell 2001 Chapter 2 © Malcolm Collier 2001 Chapter 3
© Martin Lister and Liz Wells 2001 Chapter 4
© Theo van Leeuwen 2001 Chapter 5
© Gertraud Diem-Wille 2001 Chapter 6
© Carey Jewitt and Rumiko Oyama 2001 Chapter 7
© Charles Goodwin 2001 Chapter 8 © Rick Iedema 2001 Chapter 9
First published 2001
Reprinted 2002, 2003, 2004

Apart from any fair dealing for the purposes of research or private study,
or criticism or review, as permitted under the Copyright, Designs and
Patents Act, 1988, this publication may be reproduced, stored or
transmitted in any form, or by any means, only with the prior permission in
writing of the publishers, or in the case of reprographic reproduction, in
accordance with the terms of licences issued by the Copyright Licensing
Agency. Inquiries concerning reproduction outside those terms should be
sent to the publishers.

SAGE Publications Ltd

1 Oliver’s Yard
55 City Road
London EC1Y 1SP

SAGE Publications Inc

2455 Teller Road
Thousand Oaks
California 91320

SAGE Publications India Pvt. Ltd

B–42 Panchsheel Enclave
PO Box 4109
New Delhi 110 017

British Library Cataloguing in Publication data

A catalogue record for this book is available from the British Library

Typeset by Keystroke, Jacaranda Lodge, Wolverhampton

Printed and bound in Great Britain by
Cromwell Press Limited, Trowbridge, Wiltshire
VA00post 14/9/00 10:24 am Page v

Contents

List of figures vii

List of tables ix
Notes on contributors xi
Acknowledgements viii

1 Introduction 1
Theo van Leeuwen and Carey Jewitt

2 Content analysis of visual images 10

Philip Bell

3 Approaches to analysis in visual anthropology 35

Malcolm Collier

4 Seeing beyond belief: Cultural Studies as an approach to 61

analysing the visual
Martin Lister and Liz Wells

5 Semiotics and iconography 92

Theo van Leeuwen

6 A therapeutic perspective: the use of drawings in child 119

psychoanalysis and social science
Gertraud Diem-Wille

7 Visual meaning: a social semiotic approach 134

Carey Jewitt and Rumiko Oyama

8 Practices of seeing visual analysis: an ethnomethodological approach 157

Charles Goodwin

9 Analysing film and television: a social semiotic account of 183

Hospital: an Unhealthy Business
Rick Iedema

Index 207
VA00post 14/9/00 10:24 am Page vi
VA00post 14/9/00 10:24 am Page vii

Figures

2.1 The first twenty covers of a quarter of a century of Cleo magazine 11

(1972–4).
2.2 The last twenty covers of a quarter of a century of Cleo magazine 12
(1996–7).
3.1 Antonio Romo in his wheat field, Rio Chiquito/Talpa, New 37
Mexico (USA). Photograph by John Collier, Jr., 1939.
3.2(a) Excerpt from larger panorama of Taos Valley, New Mexico 41
(USA), made in 1964. Photograph by Malcolm Collier.
3.2(b) Excerpt from larger panorama of Taos Valley, New Mexico 42
(USA), made in 1996. Photograph by Malcolm Collier.
3.3 Unloading firewood in Romo year, Talpa, New Mexico (USA). 47
Photograph by John Collier, Jr., c. 1939.
3.4 Conversation in Santistevan’s auto repair shop, Ranchos de 50
Taos, New Mexico (USA). Photograph by Malcolm Collier, 1984.
3.5 Alberto Tafoya home, Rio Chiquito/Talpa, New Mexico (USA).
Photograph by John Collier, Jr., 1930s. 53
3.6 Excerpt from photo map of Stockton Street, San Francisco 55
Chinatown made by class in Asian American Studies at San
Francisco State University. Photographs by Frances Lee,
Hernan Cortez, Suzanne Canieros and Mark Johnson, 1985.
3.7 A photo map spread out for analysis, Asian American Studies 56
class, San Francisco State University. Photograph by Malcolm
Collier, 1986.
4.1 Marlboro cigarette advert, c. 1995. 66
4.2 Robert Doisneau. ‘An oblique glance’, 1948/9. 67
4.3 De Fluminibus. Woodcut, 1483. 71
4.4 Robert Mapplethorpe. ‘Portrait of Clifton’, 1981. 74
4.5 David Hampshire. Photograph of schoolboys, c. 1986. 77
4.6 New Year’s Day, Korem camp, Wollo province, Ethiopia, 1985. 78
4.7 Mike Goldwater. Drought migrant, Zalazele transit centre, Tigray, 1983. 78
4.8 Chris Steele-Perkins. USAid, Sudan, 1985. 82
4.9 Mike Wells. A food convoy passing through Korem Refugee 82
Camp on its way to Makele Camp in Tigray, February 1985.
4.10 Carl Lewis photograph for a Pirelli advertisement. 87
4.11 Sean Bonnell. Chalk Down, from the series Groundings, 1996. 88
5.1 ‘The Third World in Our Street’ (from Bols et al., 1986). 93
5.2 General Emilio Aguinaldo represented as a black dancing girl. 104
Cartoon by Victor Gillam in Harper’s Weekly, 1899.
VA00post 14/9/00 10:24 am Page viii

viii FIGURES

5.3 Advertisement for a Dutch brand of crisps (The Netherlands, 1987). 105
5.4 Carlo Grivelli: ‘Madonna and Child enthroned with donor’. 110
5.5 Jan van Eyck: ‘St Jerome in his study’. 111
5.6 ‘In the good old summer time’, American postcard, 1907. 112
5.7 American advertisement for grapefruit, 1930s. 112
5.8 Advertisement for a Dutch brand of bananas featuring the runner 113
Nellie Cooman, 1987.
5.9 Titian: ‘Allegory of prudence’ (1569). 114
6.1 Rupert’s drawing. 124
6.2 Daphne’s first drawing. 126
6.3 Daphne’s second drawing. 126
6.4 Daphne’s third drawing. 127
6.5 Hermann’s drawing. 130
6.6 Erika’s drawing. 131
7.1 ‘Explore the possibilities’ (City and East London Health promotion). 137
7.2 From ‘4 Boys’ (Family Planning Association). 139
7.3 Sketch of ‘If he won’t use a condom’ (Health Education Authority 142
poster).
7.4 Onion cell text by Ramendeep, aged 11 years. 148
7.5 Onion cell text by Amy, aged 11 years. 149
7.6 British and Japanese ‘exit this way’ signs. 153
8.1 Gaze between speakers and hearers: transcript of Pam and Ann. 158
8.2 Transcript of archaeologists coding and recording the colour of 168
the dirt they are excavating through use of a Munsell chart.
8.3 Transcript of Ann (the senior archaeologist at the site) drawing a 172
map and Sue (her student) making measurements.
8.4 Transcript of use of video in the trial of Rodney King. 176
9.1 Frame one: frame from a scene about 7 minutes into the documentary. 183
9.2 Opening scene of Hospital: An Unhealthy Business. 193
9.3 An operation scene. 198
VA00post 14/9/00 10:24 am Page ix

Tables

2.1 Gender depictions in magazine advertisements: hypothetical 17

values on four variables.
2.2 Duration of content emphasis categories in rank order, pre-election 19
period, per se and as percentage of total Channel 9 broadcast time.
2.3 Modality and gender in magazine advertisements. 21
2.4 Hypothetical data from a pilot trial of 100 classifications with 22
four coders.
2.5 In our hypothetical study three variables can be defined, with 32
several values on each.
2.6 Classification of distance, modality and behaviour of female models 33
on the front covers of Cleo, 1972–4 and 1996–7.
5.1 Denotative signifier and signified in Figure 5.1. 99
5.2 Connotator and connotation in Figure 5.1. 99
5.3 Use of ‘photogenia’ connotators in a chapter on ‘Netherlanders’. 100
9.1 Six levels of tele-film analysis. 189
9.2 The section on ‘stenting’. 196
9.3 Summarizing the analysis. 197
VA00post 14/9/00 10:24 am Page x
VA00post 14/9/00 10:24 am Page xi

Contributors

Philip Bell is the Foundation Chair in Media and Communications, University of

New South Wales, Australia. He has conducted content-analysis based research for
various Australian government bodies and published widely on the representation of
social and political issues in the media, including: the media representation of alcohol
and drug misuse, and road safety. His books include The Media Interview (with Theo
van Leeuwen) and Implicated: the United States of Australia (with Roger Bell).

Malcolm Collier holds an MA in Anthropology from San Francisco State University

(1977), however, his primary training in visual anthropology has been in field settings
in the United States, Latin America and the Arctic. He has worked for many years
in the College of Ethnic Studies at San Francisco State University and has worked
extensively in the fields of visual anthropology, anthropology and education, applied
anthropology, and ethnic studies. He is co-author of the publication, Visual
Anthropology: Photography as a Research Method, and is currently the president of the
Society for Visual Anthropology in the American Anthropological Association.

Gertraud Diem-Wille is Associate Professor at the Institute for Interdisciplinary

Research and Education, Universities of Vienna and Klagenfurt, Austria. Her fields
of research include applied psycho-analysis, organisation development, and civic
education. Her recent publications are, Diem-Wille, G (1996) Femininity and
Professionalism (1997) ‘Observed families revisited – two years on: a follow up study’,
in Reid, S. Infant Observation: the Tavistock model.

Charles Goodwin is Professor of Applied Linguistics at UCLA. His books include

Conversational Organization: Interaction between Speakers and Hearers, and Rethinking
Context: Language as an Interactive Phenomenon (edited, with Alessandro Duranti).
In recent years he has published a range of articles relevant to visual communication,
including ‘Professional Vision’ (American Anthropologist),‘Seeing in Depth’ (Social
Studies of Science), ‘Transparent Vision’ (in Ochs, Schlegoff and Thompson, eds,
Interaction and Grammar), ‘Practices of Color Classification’ (Mind, Culture and
Activity) and ‘Action and Embodiment within Situated Human Interaction’ (Journal
of Pragmatics). His interest and teaching include video analysis of talk-in-interaction,
gesture, gaze and embodiment as interactively organized social practices, language in
the professions, and the ethnography of science.
VA00post 14/9/00 10:24 am Page xii

xii CONTRIBUTORS

Rick Iedema currently works as a Senior Research Associate with the School of Health
Services Management, Centre for Hospital Management and Information Systems
Research, University of New South Wales (NSW), Australia. He has been a researcher
in the area of mental health policy planning (this also provided the basis for his PhD
on organisational semiotics) and has written two major research reports for the NSW
Department of Education, one focusing on the print and broadcasting media and one
on the discourses of administration and bureaucracy. In addition, he has published
articles on a range of linguistic and semiotic topics.

Carey Jewitt is an experienced social researcher with a particular interest in visual

analysis. She is a Senior Research Officer at the Institute of Education, and is currently
working on an ESRC funded project, The Rhetorics of the Science Classroom:
a Multimodal Approach. Her research interests include, multimodal communication,
and visual social semiotics. Her recent publications include a social semiotic analysis
of how discourses of masculinities are visually realised, ‘Images of men’ (1997)
Sociological Research Online, Vol 2, Issue 2.

Martin Lister is Head of the School of Cultural Studies, University of West of England.
He has lectured in a range of subjects, including art history, photo-media and critical
studies, and computers and culture. He has undertaken research, and published widely
in the area of cultural studies, including, Youth, Culture and Photography (with
A. Dewdney), MacMillan (1988), and edited The Photographic Image in Digital Culture
(1995) London and New York: Routledge. In addition he has curated and produced
numerous photo-text touring exhibitions dealing with aspects of contemporary
culture.

Rumiko Oyama taught in the literature and linguistics department of Kobe College,
and the department of Intercultural Communication, University of Kobe, Japan. She
is currently researching the area of cross-cultural semiotics, and is particularly
interested in the ways in which cultural value systems are realised visually.

Theo van Leeuwen is Professor of Communication Theory at the London College of

Printing. He worked as a film and television producer before studying linguistics and
becoming an academic. He has published widely in the areas of social semiotics and
critical discourse analysis and, with Gunther Kress, is the author of Reading Images –
The Grammar of Visual Design (1996), London: Routledge. His current work is on the
semiotics of speech, music and sound.

Elizabeth Wells is a senior lecturer in film, video and photographic studies and Course
Leader, MA Independent film and Video, at the School of Media, The London
Institute: London College of Printing. She is editor of Photography: a Critical
Introduction, Routledge (1997), and curated Viewfindings – women photographers:
‘landscape’ and environment, which toured Britain in 1994/95. She has contributed
numerous essays and review articles to books and magazines within the field of visual
and cultural studies.
VA00post 14/9/00 10:24 am Page xiii

Acknowledgements

The authors and publishers wish to thank the following for permission to use copyright
material.

Network Photographers Ltd., for figure 4.2 Un Regard Oblique, Robert Doisneau and
figure 4.8 Food Convoy, Mike Wells
Art and Commerce Anthology Ltd. for figure 4.4 Clifton, 1981 Copyright © 1981 The
Estate of Robert Mapplethorpe
Magnum Photos Ltd. for figure 4.7 Daily Life in Chad, Chris Steele-Perkins
The Advertising Archives for figure 4.10 Carl Lewis for Pirelli Tyres
Sian Bonnell for figure 4.11 Chalk Down
National Gallery Company Limited for figure 5.4 Virgin and Child with Saints Francis
and Sebastian, Carlo Crivelli and figure 5.9 An Allegory of Prudence,Titian
East London and the City Health Authority for figure 7.1 Explore The Possibilities
Family Planning Association and Comic Company for figure 7.2 for images from
4Boys
Every effort has been made to trace all the copyright holders, but if any have been
overlooked, or if any additional information can be given, the publishers will be
pleased to make the necessary amendments at the first opportunity.
VA00post 14/9/00 10:24 am Page xiv
VA01post 14/9/00 10:24 am Page 1

1 Introduction

THEO VAN LEEUWEN AND CAREY JEWITT

INTRODUCTION

This book evolved from our different relationships to the visual and from our
experiences with visual analysis. Carey had recently evaluated a young men’s sexual
health clinic which had included analysis of the images in sexual health leaflets
and posters available for young men (Jewitt, 1997, 1999). The images reinforced
stereotyped forms of masculinity which, had they been put into words, would have
been unacceptable to most sexual health workers. Interestingly, these resources had
been rejected by many of the young men who used the clinic; the images were key
in this rejection, but most people found it difficult to articulate why they disliked
them. Finding a way to ‘articulate why’, to understand ‘what might otherwise remain
at the level of vague suspicion and intuitive response’, had been difficult (Iedema,
this volume: 201). One of the difficulties was accessing useful information about visual
analysis. Theo had often been involved in a search for this kind of information by
postgraduate students and colleagues in linguistic discourse analysis and pragmatics,
and had rarely been able to fully satisfy them. At our initial meeting to discuss some
of the difficulties involved in visual analysis we came up with the idea for this book.
We wanted above all to produce a book which would be a useful resource for
researchers investigating the visual representation of significant social issues, and
which provided exemplification of a range of methods and perspectives of visual
analysis in sufficient detail to make it possible for readers to actually use the
approaches explained in the book.
By way of introduction we will first give a brief description of the chapters
and then discuss some issues that arise from reading them. One of our aims in doing
the latter is to show how elements from different approaches might be combined
according to the requirements of specific research projects.

DIFFERENT PERSPECTIVES AND METHODS OF VISUAL

ANALYSIS

Content analysis (Chapter 2) has long been associated with investigations of the
way social issues are represented in the mass media, and it has the distinct advantage
of being understood and accepted by most people, including journalists. Philip Bell
VA01post 14/9/00 10:24 am Page 2

2 HANDBOOK OF VISUAL ANALYSIS

shows how it can handle large quantities of data, typically in relation to confirming
comparative hypotheses (‘expectations’), such as that women are more often
represented as engaged in domestic activities than men, or that professionals in the
movies or television series of a given period are more often played by white rather
than black actors. But, as Bell points out in the chapter, it can equally well be applied
to formal issues (for example, whether yellow is more often used on magazine
covers than green), or, indeed, to any issue that allows the formulation of clearly
definable categories and comparative hypotheses. The chapter provides enough detail
to allow readers to construct their own content analyses, including the necessary
statistics.
Visual anthropology (Chapter 3) is concerned with the use of visual records
for the description of the present and past ways of life of specific communities.
In the case of past ways of life, visuals are often (very successfully) used to elicit
memories from informants. In this chapter, Malcolm Collier draws on a wide range
of examples, including studies of Anglo immigration in New Mexico agricultural
communities, student behaviour in Cantonese bilingual classes in San Francisco,
and community schools in Alaska and the Navajo Nation in Arizona. It is very
much oriented towards the practice of visual research, discussing what kinds
of photographic and video records are most useful for the purposes of anthro-
pological research, and what kinds of contextual information to keep. The chapter
ends with a challenging call for the development of a visual language for intellectual
discourse.
Cultural Studies (Chapter 4) has recently developed a specific sub-field of
visual cultural studies, which, as Martin Lister and Liz Wells describe, is premised
on the unprecedented importance of imaging and visual technologies in con-
temporary society, and concerned with all kinds of visual information, its meanings,
pleasures and consumption, including the study of all visual technologies, from oil
painting to the Internet. The chapter sees Cultural Studies as an interdisciplinary
field, and describes it, not as a specific methodology, but as an agenda of questions
and issues for addressing specific images. These questions and issues are then matched
to the conceptual frameworks and methodologies of a range of different disciplines.
This approach makes the chapter particularly useful as a model for integrating
the various approaches discussed in this volume as a whole. The chapter draws
primarily on examples of mass media and art photographic images, from global
cigarette advertising campaigns to documentary photographs of famines in Africa –
the latter, as Lister and Wells point out, having played a key role in constructing a
Eurocentric view of Africa and its peoples as ‘economically and technologically
weak, dependent victims of natural disaster’ (p. 78).
The chapter on semiotics and iconography (Chapter 5) written by Theo van
Leeuwen discusses the visual semiotics of Roland Barthes and the iconographical
method of visual analysis developed by art historians such as Edgar Wind,
Erwin Panofsky and Meyer Schapiro. Both methods are premised on the idea of
layered meaning, of images consisting first of all of a layer of representational or
denotative meaning (the layer of who and what are depicted here) on which is then
superimposed a layer of connotative or symbolic meaning (the layer of what does it
VA01post 14/9/00 10:24 am Page 3

INTRODUCTION 3

all mean). Both methods provide specific pointers for distinguishing and analysing
these layers, and specific criteria for arguing whether or not a layer of symbolic
or second-order meaning is present. The main difference between the two is that
iconography uses both textual and contextual criteria for arguing symbolic meaning,
while Paris school semiotics restricts itself mostly to textual criteria, to pointers within
the image itself. Although iconography has mainly been applied to art works from
the past, the chapter attempts to demonstrate that it can also be applied to contemp-
orary images, using Jan Nederveen Pieterse’s (1992) history of the European and
North American depiction of Africans and Afro-Americans as a main source of
examples.
The chapter on psychoanalytical image analysis (Chapter 6), written by
Gertraud Diem-Wille, a Viennese psychoanalyst specializing in the treatment of
children, argues that children, within the special context of the psychoanalytic session,
produce drawings that are based on the same primary processes of representation
as dreams. Their meanings can therefore be brought out through psychoanalysis, just
like those of dreams. Diem-Wille then goes on to apply this method to what could
be called a socio-psychoanalytical study of what drives highly successful career men
and women. Here the projective drawing technique is seen to lower interviewees’
defences, enabling them to visually express what they are inhibited from verbally
expressing. Although Diem-Wille is reluctant to generalize, the cases she discusses
reveal patterns which have wider validity in understanding the influence of parental
relationships on career choices, and paths, and the role of work in the formation of
identity.
Social semiotic visual analysis (Chapter 7) provides a detailed and explicit
method for analysing the meanings established by the syntactic relations between the
people, places and things depicted in images. These meanings are described as not only
representational, but also interactional (images do things to or for the viewer),
concerned with the modality or perceived truth value of images, and compositional
(for example, positioning images and written text in certain ways). In this chapter
Carey Jewitt and Rumiko Oyama characterize social semiotics as concerned with the
study of images in their social context, and as a critical form of visual discourse analysis
which does not necessarily stop at description but may also seek to influence the
semiotic practices it describes. The chapter shows the method at work in three
different research projects: the study of sexual health materials already mentioned, an
ethnographic/semiotic study of the primary school science classroom and a study of
some of the cross-cultural differences in visual syntax which distinguish British and
Japanese exit signs and magazine advertisements.
In the case of conversation analysis and ethnomethodology (Chapter 8),
visual analysis is not so much a matter of analysing images (with or without
consideration of their context) as of analysing the dynamic unfolding of specific
social practices in which non-verbal communication (pointing, gaze work, and so
on) and images (including signs, maps and diagrams) play a role. The chapter
begins by showing how, without taking account of visual communication, conver-
sation analysis might not only miss out on information but lead to inaccurate
conclusions. Charles Goodwin discusses the role of a range of different kinds of
VA01post 14/9/00 10:24 am Page 4

4 HANDBOOK OF VISUAL ANALYSIS

image in the work of scientists and other experts. Both the production of images
(for example, map-making by archaeologists) and their interpretation (for example,
the interpretation of the Rodney King videotape by police experts during the infamous
trials) take place in situated interactions that use a variety of communication modes
– speech, non-verbal communication (for example, pointing and gaze work) and
images (for example, the archaeologists’ map or the Rodney King video). The same
images, moreover, may be used differently by different participants (for example, a
schedule of plane arrivals and departures is used differently by baggage handlers and
gate agents). The chapter provides specific pointers for analysing existing images
(and for producing useful video recordings for research purposes), and argues
powerfully for studying visual communication in its socially specific and multi-
modal contexts.
The chapter on film and television analysis (Chapter 9) discusses a television
documentary which depicts the conflicts between the clinical and administrative
sections of a large hospital, and demonstrates how this film is systematically con-
structed to favour one point of view over others, and how a study of this kind of bias
might be conducted. Combining elements from film theory and social semiotic
genre analysis, Rick Iedema describes the different levels at which film and television
texts can be studied (the frame, the shot, the scene, the sequence, the stage and the
genre), and then looks at key variables and methods relevant to each of these levels.
At the time of writing the chapter, the author was working as a research consultant
for a large hospital, and hence the chapter is an example of applied visual analysis,
of a form of critical analytical practice taking place inside an institution, with the
aim to influence and change the practices of that institution.

THE IMAGE AS RECORD AND AS CONSTRUCT

Some of the chapters in this volume describe the analysis of images specially produced
for research purposes (Chapters 3 and, in part, 8). Such images are produced to
serve as records of reality, as documentary evidence of the people, places, things,
actions and events they depict. Their analysis is a matter of extracting just that kind
of information from them. The same applies to the therapeutic use of drawings
described in Chapter 6. The analysis of these drawings must bring out, for instance,
what the family relations in a given subject’s childhood were like and how they were
experienced by the subject. Art historians, too (Chapter 5), often analyse images as
sources of factual information, even though, in this case, the images may not have
been specially produced for this purpose. Researchers who use images in this way
are of course aware of the limitations of images (including photographic and video
images) as sources of factual information. Good research images, as Collier describes
them (Chapter 3), should not be overly constructed, or complex, and therefore
harder to read than images produced for the media or as art images; also they should
not be isolated from the series of images to which they belong (for as he points out,
single images and images without extensive contextual annotation are problematic
for research purposes). However, despite their limitations, images are, in this context,
VA01post 14/9/00 10:24 am Page 5

INTRODUCTION 5

regarded as a reliable source of factual evidence, and Collier in fact prefers them
over the ‘deceptive world of words’ (p. 59).
In other cases images are analysed, not as evidence of the who, where and what
of reality, but as evidence of how their maker or makers have (re-)constructed reality,
as evidence of bias, ideologically coloured interpretation, and so on. This is common
in Cultural Studies and semiotic analyses (for example, Chapters 4, 5, 7 and 9), and
in ethnomethodological research when the process of (re-)constructing reality itself
is documented, such as in studies on the way scientists change the apparently unruly
and messy world of photographs into the more orderly world of diagrams by
‘filtering’, ‘uniforming’, ‘upgrading’ and ‘defining’ photographs (Lynch, quoted in
Chapter 8: 163). From these perspectives the image is more unreliable and slippery as
a source of factual information. In this volume Lister and Wells, like Sekula whom
they quote, mistrust the ‘evidence’ of photojournalism and documentary photography:
‘. . . when photographs are uncritically presented as historical documents, they are
transformed into aesthetic objects. Accordingly, the pretence to historical under-
standing remains although that understanding has been replaced by aesthetic
experience’ (Sekula in Chapter 4: 89).
The point for us is not to construct theoretical arguments in favour of reality
or construction, or to arbitrate in the debates on this issue from our editorial position.
Rather we think the point is to urge you to keep the distinction in mind when reading
the chapters in this volume, and to consider that the choice of an appropriate method
of analysis is dependent on the nature of the project in which it is to be used, on the
visual material that is being investigated, and on the goals of the research project.
Indeed, sometimes several methods may be necessary. One of us is currently engaged
in a research project dealing with children’s toys. The project investigates both the
meanings offered to the child by the toy industry and the mass media (through the
texts and pictures on toy packaging, in toy catalogues, in toy advertisements, and so
on) and the way in which these (and other) meanings are taken up in parent–child
interactions and in the child’s actual playing with the toys. Clearly the former requires
a mode of analysis in which the various packaging texts, catalogues and advertisements
are treated as constructs, and the latter the analysis of videotapes specially produced
for the research as ethnographic evidence of parent–child interaction and children’s
play with toys.
The issue of ‘record’ versus ‘construct’ exists because many images have an
element of both and so require a mode of analysis which is sensitive to both. Clearly
advertising images are in the first place constructs and their analysis must reveal
the nature of these constructs. Equally clearly construction has to be minimal (or,
where images are specially produced for research purposes, minimized) if images
are to be used as records of people, places, things, actions or events. Again, it is no
accident that studies aiming at changing practices of representation often choose a
detailed and explicit method of analysing construction and its effects, so as to avoid
the idea that it is all in the eye of the beholder, which in this case would be counter-
productive (cf. Chapters 2, 7 and 9); alternatively they could of course document
the very processes of that construction, as happens in ethnomethodological research
(Chapter 8).
VA01post 14/9/00 10:24 am Page 6

6 HANDBOOK OF VISUAL ANALYSIS

THE UNITS OF ANALYSIS: SINGLE IMAGES VERSUS

COLLECTIONS OF IMAGES

Some of the approaches to visual analysis described in this volume are based on the
analysis of collections of images, others on the analysis of single images. Chapter 2
shows that content analysis requires at least two different sets of data (for example,
images from two different periods or publications) for the purpose of comparison,
and that each of these sets needs to contain a sufficiently large number of similar
images (for example, all advertisements containing images of women from a given
period of time) in order to be both representative and statistically significant. Visual
anthropology (Chapter 3) also uses collections of images, but for different purposes.
Collier describes two kinds of collections: first, collections of many different images
of the same subject (for example, wide views, details and a range of angles of the
same street) which are put together to allow patterns to become visible; and second,
collections of images made to help identify what is depicted in a given image (for
example, a 60-year-old image of an elderly man engaged in wheat harvesting was
compared to other photos of the same man in different settings, other photos of the
area and other photos of wheat harvesting, in order to establish exactly who the man
was, where he was photographed and what he was doing). As discussed in Chapter
5, art historians also use this method when they want to establish the exact who,
where and what of art works of the past. As noted by Collier and Goodwin in
Chapters 3 and 8, films and videos are always collections of images.
In other cases single images are discussed. But, as indicated in Chapters 2,
5 and 7, any method of visual analysis which provides a wide enough range of
clearly defined specific image features and connects them convincingly enough with
particular meanings and/or communicative effects can be used either for the analysis
of single images or quantitatively.

TEXT, CONTEXT, SOCIAL PRACTICE

There is another way in which the various approaches described in this volume do not
use the same units of analysis. Visual analysis may be based only on what is visible
within the image or collection of images (in the text), as is the case, by and large, with
content analysis (Chapter 2) and also with various types of semiotic analysis (Chapters
5, 7 and 9). It may draw on contextual information, whether gleaned from interviews,
as in the case of visual anthropology (Chapter 3) and the therapeutic interview
(Chapter 6), or from archival research and background reading, as in the case
of iconography (Chapter 5). Or, as in the case of Cultural Studies (Chapter 4), it
may use a range of different kinds of information. The approach of Chapter 8 is
different again. Here the unit of analysis is not the text, or the text together with
external, contextual information, but the enacted social practices in which images
are used.
The same applies to the question of word and image. Images may be analysed
without any recourse to the verbal or written information which may accompany
VA01post 14/9/00 10:24 am Page 7

INTRODUCTION 7

them (for example, the catalogue of an exhibition or the introduction to a book of

artistic photographs). The images may have been designed to be self-sufficient, or
capable of being inserted into many different contexts, as certainly is the case with
much contemporary fine art and ‘classic’ photojournalism (such as Cartier-Bresson
in Chapter 4). At the very least there is as much of a case here for analysing the broader,
supra-contextual meaning potential of images, as for analysing their meanings in a
particular context, for example a particular exhibition. But visual analysis may also
include the accompanying text, or even see word and image as one indivisible unit of
analysis, as in the social semiotic analysis of layout (Chapter 7). In such an approach
the catalogues, text panels, and so on of an exhibition would be included in the
analysis, as a secondary source of information, serving, for instance, to anchor who
or what is depicted or what is symbolized. But the analysis would still not include the
way this meaning potential is actually taken up in the interactions of specific social
actors visiting the exhibition, as would be the case in an ethnomethodological
approach. Again, the choice of method depends on the nature of the material analysed
and on the goals of the analysis. There are good arguments for analysing images in
relative independence of their context (for re-contextualizing them, in other words),
and for analysing them together with the physical context and or social interaction
in which they are embedded (as, for instance, in the analysis of reading advertisements
on hoardings or in magazines in Chapter 4).

TEXT, PRODUCER, VIEWER

These different perspectives on text, context and social practices have implications
for the way in which the producers and viewers of images may be included or
implied in the analysis. A mode of analysis which restricts itself to the evidence of
the text may not, as Bell puts it, ‘by itself demonstrate how viewers understand and
value what they see or hear’ (Chapter 2: 26), or what producers, deliberately or
otherwise, intend to communicate. On the other hand, text analysis can show what
representations include and exclude, what they prioritize and make salient, and
what differences they construct between different people, places and things.
The degree to which producers and viewers should be included and how this
is to be achieved again depends on the kinds of images analysed, and on the aims of
the research. Research aimed at discovering how a general audience understands
advertising images should perhaps not include information about the production
of advertisements which such an audience would not be familiar with. On the other
hand, if a researcher wanted to change the audience’s perceptions, for instance
through media education, a ‘look behind the scenes’ might become relevant. Again,
research aiming at critiquing, say, racism or sexism in certain representational
practices clearly has an interest in linking these practices to specific social institutions,
but in the analysis of art images (Chapters 4 and 5) producers may be depicted,
not in terms of the formal or informal institutions within which their work is (or
was) situated, but as individuals working on the basis of traditions, influences and
inspirations – and here the audience will be less often taken into account. Important
VA01post 14/9/00 10:24 am Page 8

8 HANDBOOK OF VISUAL ANALYSIS

and innovative research often thrives in the ‘blind spots’ of specific research
approaches and methods, applying the methods of one approach to the kind of
material studied in another – for instance, considering the individual in an area where
most research has concentrated on institutions, or institutions in an area where most
research has concentrated on individuals.
When social practices are taken as the units of analysis (as in Chapter 8), the
difference between producers and consumers is much diminished. Both the archae-
ologist showing an apprentice how to construct an archaeological map and the police
expert showing a jury how to interpret the Rodney King video use the resources of
their expert knowledge to construct meaning. This method could also be used to show
how the script of a television series is produced, or how university students learn new
interpretations of a television series in Media Studies courses – all aspects of visual
communication which could never be revealed, for instance, through content analysis.
At the same time, this approach would not bring out what aspects of social life
television series more generally include and exclude, what kind of interpretations are
generally favoured in the mass media, and so on – questions which content analysis
and semiotics are well placed to answer.

CONCLUSION

Clearly some methods of analysis are more methodical than others. Some lay down
very precise criteria for analysis, so that the impression may arise that visual analysis
can be done ‘by rote’, and described as a kind of recipe, a procedure to be followed
step by step, without the need for any form of initiative, let alone inspiration. Content
analysis, with its more or less mechanical statistical processing of data, and social
semiotic analysis, with its proliferation of features and precise criteria for analysing
them, tend most clearly in this direction. Anyone who has actually tried these methods
knows that there is a great deal more room for initiative and, indeed, inspiration than
is sometimes acknowledged in the way these methods are described. These methods
remain an art of interpretation, but one that follows certain rules of accountability.
Other forms of analysis provide less precise rules for conducting the analysis.
Cultural Studies and ethnomethodology, for instance, certainly depart from precise
theoretical positions, research questions and principles of research, but they do not
provide a large number of analytical categories and nor do they explicitly construct
research work in terms of a ‘step-by-step procedure’. The approach described by
Collier in Chapter 3 provides an intermediary position. Collier sees visual analysis
as a complex process which alternates between stages that require an intuitive grasp
of the whole and stages that require the hard work of structured analysis, of careful
and methodical checking and double-checking. For Collier, it is both necessary to
‘observe the data as whole’, to look at, listen to ‘its overtones and subtleties’, to ‘trust
your feelings and impressions’ and to ‘go through the evidence with specific questions
– measure distance, count, compare. Produce detailed descriptions’ (Chapter 3: 39).
He sees visual analysis as both art and science: ‘It is both necessary and legitimate to
allow ourselves to respond artistically or intuitively to visual images . . . However,
VA01post 14/9/00 10:24 am Page 9

INTRODUCTION 9

while creative processes are essential to discovery, artistic processes may produce
only fictitious statements if not combined with systematic and detailed analysis’
(p. 59). In our opinion, this can be usefully applied to all visual analysis.

REFERENCES

Jewitt, C. (1997) ‘Images of men’, Sociological Research Online, 2 (2).

Jewitt, C. (1999) ‘A social semiotic analysis of male heterosexuality in sexual health
resources: the case of images’, International Journal of Social Research Methodology,
1 (4): 263–79.
VA02post 14/9/00 10:25 am Page 10

2 Content Analysis of Visual Images

PHILIP BELL

INTRODUCTION

Cleo is a magazine aimed at female readers from eighteen to thirty-five years

of age. First published in Australia in 1972, it celebrated its twenty-fifth birthday
by reproducing all three hundred of its front covers in its November 1997 spectacular
with Kylie Minogue as its model on the cover. The first and last twenty covers from
the quarter of a century of Cleo are reproduced as Figures 2.1 and 2.2 respectively.
What differences can be seen if one compares the two groups of covers
published 23–5 years apart? How has Cleo’s ‘image’ (and the images which make up
its ‘image’) changed? To answer this, either in conversation at the hairdresser’s or as
an academic thesis, requires identification of observable dimensions of the images in
question (for example, are the models who are depicted older or differently dressed?),
as well as a judgement about how frequently various visual features appear in the
periods that one chooses to compare. In short, the answer requires a (visual) content
analysis.
This chapter outlines the assumptions, practices, limitations and advantages
of explicit, quantifiable analysis of visual content as a research method. It then returns
to the Cleo covers to exemplify the process.

CONTENT ANALYSIS AND WHEN IT CAN BE USED

Generalizations about what is shown on television, in the press or in advertisements

require observable, more or less ‘objective’ evidence. Consider the ubiquitous claims
and counter-claims about, say, the depiction of women in magazine advertisements
or the prevalence and effects of televisual or cinematic violence. When such issues are
debated, the frequency and the meaning of identifiable classes of discrete visual
content are at stake (as opposed to more complex forms like narrative structures or
the complexities of different genres).
Whether it is explicitly labelled content analysis or not, making generalizations
about the relative frequencies of visual representations of particular classes of people,
actions, roles, situations or events involves implicit or explicit classification and
quantification of media-circulated content.
VA02post 14/9/00 10:25 am Page 11

CONTENT ANALYSIS OF VISUAL IMAGES 11

Figure 2.1 The first twenty covers of a quarter of a century of Cleo magazine (1972–74).
VA02post 14/9/00 10:25 am Page 12

12 HANDBOOK OF VISUAL ANALYSIS

Figure 2.2 The last twenty covers of a quarter of a century of Cleo magazine (1996–7).
VA02post 14/9/00 10:25 am Page 13

CONTENT ANALYSIS OF VISUAL IMAGES 13

Not surprisingly, therefore, content analysis has provided one of the most
widely cited kinds of evidence in Media Studies for many decades. First, in relation
to newspapers and radio (for you can analyse verbal content as well as visual)
and, later, directed at television and occasionally at the cinema. Perhaps the method
has been widely used because it seems like the ‘commonsense’ way to research what
the media show, or because it appears to require little theoretical analysis. In short, it
is the most basic way of finding out something about the media’s meaning and allows
for apparently general statements to be made about aspects of representation which
non-specialists, journalists and experts alike can understand. However, as will be clear
in the following discussion, content analysis is quite a technical procedure. It is also
of limited value in many research contexts, and might best be thought of as a necessary
but not sufficient methodology for answering questions about what the media depicts
or represents. Content analysis alone is seldom able to support statements about the
significance, effects or interpreted meaning of a domain of representation. For
example, using content analysis to show that prime-time television depicts a high
level of physical inter-personal aggression does not, by itself, show that viewers are
affected in any particular way (either by imitating what they see or by inhibiting their
own aggressive behaviours). Claims about the effects of what is shown raise questions
which need to be addressed by further, different kinds of research which are not
considered in this chapter (see Chapters 4 and 9).
Let us now become more precise and technical in defining and describing
the range of procedures for visual analysis that content analysis refers to. First, a
general definition: content analysis is an empirical (observational) and objective
procedure for quantifying recorded ‘audio-visual’ (including verbal) representation
using reliable, explicitly defined categories (‘values’ on independent ‘variables’).
To illustrate each component of the definition, consider the example of
making generalizations/representations about women and men in advertisements in
‘women’s’ compared with ‘men’s’ magazines. To begin to observe how women and men
are depicted requires an explicit hypothesis (or hypotheses) without which the com-
plex field is too diverse, ill-defined and therefore unable to be systematically analysed.
So content analysis begins with some precise hypothesis (expectation) or
question about well-defined variables. In our example, these variables could include
types of magazine, size of published advertisement, (defined) pose of represented
models (say, standing, seated, walking, running) and depicted context (e.g. office,
home, outdoors). One explicit hypothesis might be that women will be depicted in
fewer outdoor situations than men, in both kinds of magazines. Only when one or
more hypothesis is formulated will the relevant variables become apparent.
Note that the kinds of hypotheses which content analysis usually evaluates are
comparative. That is, the researcher is usually interested in whether, say, women are
depicted more or less frequently than men in relation to some variable or quality, not
in the absolute frequency of certain depictions taken in isolation. Obviously,
if 20 per cent of women in television commercials are shown in ‘outdoor’ settings,
these data tell the researcher very little unless the comparable frequency for men (in
otherwise similar contexts) is known. Content analysis is used to test explicitly com-
parative hypotheses by means of quantification of categories of manifest content.
VA02post 14/9/00 10:25 am Page 14

14 HANDBOOK OF VISUAL ANALYSIS

Let us continue discussing our (admittedly quite simplified) examples by

drawing attention to other aspects of content analysis. First, the choice of the
scope, number or scale of the visual content to be studied must be explicitly described
before detailed observations begin. This involves deciding what ‘corpus’ or what
‘sample size’ or what ‘field’ or ‘domain’ of representation is to be studied: how many
magazines, how many advertisements, how many editions of one or more newspapers,
and so on (see Chapter 3). For example, because it is comparative, a hypothesis might
specify two or more periods (say, five years before World War II and five years after)
during which to compare some aspects of gender representation. Because World War
II brought women increasingly into the paid workforce this hypothesis would be of
sociological interest. In all such cases, the question arises of how many data are
enough? Can we generalize from our samples of images to the population from which
it has been selected? This question ultimately requires statistical decisions and
inferences, but it is clearly important to ensure that the corpus (the sample(s)) is
representative of the variables specified in the relevant hypotheses. Obviously, if a
researcher only samples from one women’s magazine she or he could not generalize
the findings to all women’s magazines, nor even draw conclusions about the particular
magazine beyond the period sampled. (Women’s magazines have changed radically
in the last several years.) We will return to this issue after outlining the technical
procedures to follow when conducting a well-designed content analysis.
To summarize this introductory discussion: visual content analysis is a
systematic, observational method used for testing hypotheses about the ways in which
the media represent people, events, situations, and so on. It allows quantification of
samples of observable content classified into distinct categories. It does not analyse
individual images or individual ‘visual texts’ (compared with psychoanalytical analysis,
discussed in Chapter 6, and semiotic methods, discussed in Chapters 4, 7 and 9).
Instead, it allows description of fields of visual representation by describing the
constituents of one or more defined areas of representation, periods or types of images.
Typical research questions which may be addressed using content analysis
include:

1 Questions of priority/salience of media content: how visible (how frequently,

how large, in what order in a programme) different kinds of images, stories,
events are represented? ‘Agenda Setting’ studies of news broadcasts would be
an example of this kind of question.
2 Questions of ‘bias’: comparative questions about the duration, frequency, pri-
ority or salience of representations of, say, political personalities, issues, poli-
cies, or of ‘positive’ versus ‘negative’ features of representation.
3 Historical changes in modes of representation of, for example, gender, occu-
pational, class, or ethically codified images in particular types of publications
or television genres.

WHAT TO ANALYSE: ‘ITEMS’ AND ‘TEXTS’

What is the ‘content’ of content analysis? As the examples mentioned already

indicate, the material analysed by means of content analysis may be visual, verbal,
VA02post 14/9/00 10:25 am Page 15

CONTENT ANALYSIS OF VISUAL IMAGES 15

graphic, oral – indeed, any kind of meaningful visual/verbal information. To analyse

is to break a thing down into its constituent elements. So the visual/verbal units of
meaning which are the objects of content analysis are those defined by the medium
in which they are produced as isolable, self-contained or separate, like paragraphs,
framed images, pages or news photographs. I will call these ‘texts’ whether or not
they are verbal. A display advertisement is a visual text. A news item (on television)
is also a text, because it has a clear frame or boundary within which the various
elements of sound and image ‘cohere’, ‘make sense’ or are cohesive. Texts, then, are
defined within the context of a particular research question and within the theoretical
categories of the medium (gallery painting, television) and genres (such as portraits,
news or ‘soaps’) on which the research focuses.
The texts that are to be analysed are the meaningful and cohesive units framed
within the medium and genre(s) at issue. Visual content analysis usually isolates
framed images (in publications) or sequences of representation (scenes or shots in
television or film). But, unlike semiotic analysis, content analysis classifies all the
texts on specified dimensions (what will be called ‘variables’ in this chapter) to
describe the field or totality. It is not concerned with ‘reading’ or interpreting each
text individually. By contrast, semiotic analysis is qualitative and usually focuses on
each text or text-genre in much the way that a critic focuses on a particular gallery
painting or on the aesthetics or cultural connotations of a particular film or class of
films (see Chapters 4, 7 and 9; in addition, Chapter 5 demonstrates an iconographic
approach to individual texts and Chapter 6 describes a psychoanalytical approach to
individual texts).

HOW TO DO CONTENT ANALYSIS

Like all research methodologies, content analysis is an effective procedure only if

precise hypotheses and clearly defined concepts underpin its use. Categories of
(visual) content must be explicitly and unambiguously defined and employed
consistently (‘reliably’) to yield meaningful evidence relevant to an hypothesis. To
observe and quantify categories of content it is first necessary to define relevant
variables of representation and/or salience. Then, on each variable, values can be
distinguished to yield the categories of content which are to be observed and
quantified. The concepts of variables and values are critical to an understanding of
the procedure, so I will define these more completely below.

Variables
Visual or verbal representations differ from each other in many ways – on many
dimensions or qualities. A content variable is any such dimension (size, colour
range, position on a page or in a news bulletin); or any range of options of a similar
type which could be substituted for each other – for example, a list of represented
participants (male/female; adult/child) or a number of alternative ‘settings’ such as
kitchen, bathroom, street, automobile, shop, and so on. Variables like size, represented
VA02post 14/9/00 10:25 am Page 16

16 HANDBOOK OF VISUAL ANALYSIS

participants, settings, priority, duration and depicted role consist of the set of options
which are of the same class or type as defined for the purposes of the research
project. Notice that in content analysis a variable refers to aspects of how something
is represented, not to ‘reality’. So if someone is shown as elderly in a television ‘soap’
(by means of make-up or clothing) it is this which the researcher observes, judges
and classifies, not the ‘real’ age of the actor. This point is sometimes referred to as
judging the ‘manifest content’ of an image or text. It is the content as represented
that is analysed, not some independently or ‘externally’ defined, and certainly not
any aspect of ‘reality’ not actually depicted. So all the variables defined are those on
which particular representations differ from one another.

Values

A variable consists of what we will call values. These are elements which are of
the same logical kind. That is, elements can be substituted for each other because
they belong to the same class: these constitute the values on a particular variable.
For example, all the occupational roles in which people are, or could be depicted
in television commercials would constitute a variable. But such a ‘role’ variable
would be distinct from, say, the variable of ‘depicted physical setting’, from which
it is conceptually independent. Of course, some roles usually occur in particular set-
tings, but that is an empirical contingency and does not imply that the two variables
are the same (Chapter 9 provides a more detailed discussion of this issue). To ensure
that one does not confuse two variables, a useful test is as follows: could a defined
value on one variable be substituted for another on that variable or not? If not, one
has confused two distinct variables.
To summarize, then, a content analysis begins with the definition of relevant
variables and of the values on each. Each variable is logically or conceptually in-
dependent of every other distinguished in a particular research project.
The values defined on each variable should also be mutually exclusive and
exhaustive (another technical criterion which will be discussed below).
Schematically, the specification and definition of variables and the values of
each can be represented as in Table 2.1. Here, four variables have been distinguished.
Within each, a different number of values have been specified. In research on gender
representation, to continue one of our hypothetical examples, the table might
appear as in Table 2.1.
Note again the logic of such a system: each variable is independent of all others.
All values on a given variable are then mutually exclusive: any observed element of
representation can only be classified into one value on each variable, and values that
are defined should be exhaustive – the values should cover all the possible cate-
gorizations on the respective variable. This may require a ‘miscellaneous’ or ‘other’
value to be allowed on some variables as it is not always possible for the researcher to
anticipate the complete range of values (for instance, occupational roles, to continue
our example) that may be found during the course of research. So ‘politician’ may be
a value or role which is only infrequently found. It may be included in a ‘miscellaneous’
VA02post 14/9/00 10:25 am Page 17

CONTENT ANALYSIS OF VISUAL IMAGES 17

value if its limited visibility is not considered important for the hypotheses being
evaluated. However, the fact that a role is only infrequently depicted may be of great
interest to the researcher, in which case such a value should be explicitly included in
the range of values which make up the defined variable.

Table 2.1 Gender depictions in magazine advertisements: hypothetical values on four

variables.

Variables
Gender Role Setting Size

Male House duties Domestic Full page

Nurse

Values Female Business Public Half full page

executive

Flight attendant More than half

page

Teacher Double page

2 5* 2 4

* These values are purely indicative; in fact there are likely to be many more values on this
variable.

Definitions of variables and values

The principal difficulty confronted by researchers wishing to conduct a comprehen-

sive content analysis is to define unambiguously both the variables and values so
that observers (usually called ‘coders’) can classify represented content consistently
(reliably).
Instead of speaking of ‘categories’, I have distinguished variables and values.
Each variable, and within it each value, should be defined in terms of one principal
feature of representation. This is important even when a variable is physical and seems
to be relatively unambiguous. The ‘duration’ (in seconds) of television advertise-
ments is clearly an unambiguous measure, but the size of display advertisements may
not be. Magazines vary in size. So the absolute size (area in square centimetres or
inches) of an item such as an advertisement may not be an appropriate variable to
allow comparison across magazines. If it is not, the researcher needs to define the
variable (size) and its values to best demonstrate comparative salience or ‘noticability’.
In the example in Table 2.1 this was done by classifying the size of the published
advertisement relative to a standard page of the respective publication (as two page,
one page, half page, etc.).
Other ‘subjective’ variables (those which require coders to make subjective
judgement) demand very clear definitions and criteria. Otherwise judges (coders)
VA02post 14/9/00 10:25 am Page 18

18 HANDBOOK OF VISUAL ANALYSIS

may apply inconsistent criteria, or one coder may change her or his criteria through
the course of coding a large number of instances.
Apart from physical features of images or other visual items (such as size,
duration and monochrome/colour) the overall topic or subject of news items or the
product category of advertisements will generally be of interest to researchers. In such
cases, the question of the appropriate scale of analysis becomes important. Does one
need to define, say, camera angles or shot lengths in relation to represented participants
(such as political figures) in analysing television news? Or is a global judgement about
the focus of news items all that is required? In the latter case, the variable will be
named something like ‘news story topic’, and values will be defined to permit coders
to classify items to show the ‘agenda’ of issues of one or more news programmes. For
instance, Bell et al. (1982) compared the agendas of different networks’ news coverage
during an Australian Federal (National) Election by defining values such as the
following on the ‘news topic’ variable:

(a) Celebrities & personalities: Major emphasis given to the personality(ies) per se,
rather than their expertise or skills. Item focuses on the person as newsworthy
owing to glamour/wealth/bravery/prestige, or as embodiment of some other
desirable or undesirable characteristic(s). Excludes Vice Regal, Monarchy (see
Vice Regal & Monarchy). For example:
– Kirk Douglas arrives for Australian Film Awards.
– Former beauty queen graduates from police academy.

(b) Crime: Major emphasis given to the manifest criminal event or its conse-
quences: robbery, murder, swindles etc. or to technical or biographical stories
of crime/criminals. For example:
– Inquiry into penetration of the Painters and Dockers Union by organized
crime.
– Jewel thieves arrested.

(c) Disasters: Major emphasis given to climatically induced phenomena causing

human suffering (flood, heatwaves), geological events of similar kind (earth-
quakes, avalanches). Excludes accidents (see Accidents & Chance Events). For
example:
– Newcastle district bushfire.
– New South Wales Government aid to drought areas.

As these definitions show, coding something as complex as audio-visual news

stories demands general, yet precise definitions. The kinds of data that are yielded
by analysing a large corpus of material such as television news coverage during
an election campaign are illustrated in Table 2.2 taken from the same research
monograph.
VA02post 14/9/00 10:25 am Page 19

CONTENT ANALYSIS OF VISUAL IMAGES 19

Table 2.2 Duration of content emphasis categories in rank order, pre-election period,
per se and as percentage of total Channel 9 broadcast time.

Content mins:secs %

1 Sport 112:16 24.6

2 Federal Election 81:38 17.9
——— 10%
3 Accidents and chance events 35:31 7.8
4 Crime 27:42 6.1
5 Celebrities and personalities 24:24 5.3
6 Disasters 23:00 5.0
7 War 21:21 4.7
8 Arts, entertainment and leisure 13:00 2.8
9 Demonstrations and protests 11:45 2.6
10 Health and safety 10:12 2.2
11 Investment and shares 9:21 2.0
——— 2%
12 Economy 7:51 1.7
13= Industrial relations and disputes 7:17 1.6
13= Prisons 7:17 1.6
15 Defence 6:50 1.5
16 Energy 6:48 1.5
17 Elections – other 6:01 1.3
18 International relations 5:51 1.3
19 Vice regal and monarchy 5:25 1.2
——— 1%
20 Police 3:58 0.9
21 Legal 3:32 0.8
22 Security 3:23 0.7
23 Transport 3:02 0.7
24 Science, medicine and technology 2:54 0.6
25 Environment 2:36 0.6
26 Minerals and resources 2:28 0.5
27 Communications 1:55 0.4
28 Administration 1:50 0.4
29 Rural 1:42 0.4
30 Trade unions 1:40 0.4
31 Weather 1:37 0.4
32 Social welfare 1:30 0.3
33 Employment 0:25 0.1
34 Education 0:22 0.1
——— 0%
35= Aboriginal affairs 0:00 0.00
35= Animals 0:00 0.00
35= Housing 0:00 0.00
35= Migrant and ethnic affairs 0:00 0.00
35= Religion 0:00 0.00
35= Women’s issues 0:00 0.00

Total time 7:36:24

VA02post 14/9/00 10:25 am Page 20

20 HANDBOOK OF VISUAL ANALYSIS

Note that some content values are not represented in Channel 9’s coverage.
These may be very significant as absences (as issues not on the broadcaster’s ‘agenda’
despite their sociopolitical importance, for instance, the low priority of Aboriginal
issues).
Visual content analysis may focus on more ‘micro’-level variables than does
the above. Indeed, there is no limit to how precise and finely grained an analysis
may be other than the ability of the researcher to define clearly, and coders to apply
reliably, the specified criteria. Whether a model smiles or not, whether they look
at the camera, whether they are clothed in certain ways, their skin or hair colour,
can all be subjected to content analysis and then to quantification. But whether a
model is ‘attractive’, ‘young’ or ‘American’ (as depicted) is unlikely to be clearly and
unambiguously definable. This is because each of these variables is a composite of
more specific variables. Like ‘size’ (above) each would require analysis into specific
dimensions (variables) of definable values which are one-dimensional rather than
‘composite’ (and therefore potentially vague).

Quantitative results: comparisons and cross-tabulations

Content analysis classifies extensive fields of representation in quantitative terms. As
was emphasized earlier in this chapter, the kinds of hypotheses which such quanti-
fication helps to test are those which compare one field of representation with another.
The examples referred to above included comparisons between ‘men’s’ and ‘women’s’
magazine advertisements in terms of visual representations of gender roles. The data
set out in Table 2.2 are not comparative. However, that table was originally published
with relevant comparison tables showing how other television channels prioritized
their respective agendas during the same period (according to the same variables and
values, of course). Such tabulation of data by more than one variable (here, television
station by topics) is typical of the comparative cross-tabulations which result from
content analysis.
Another example illustrates this more succinctly. In Table 2.3, advertisements
showing one or more female or male figures, or parts of the body (for example, an
arm wearing a watch), or groups of males or females, have been analysed according
to the ‘visual modality’ of the respective advertisement published in popular
magazines. ‘Modality’ here refers to the ‘truth value’ or credibility of (linguistically
realized) statements about the world (Kress and van Leeuwen, 1996) (see Chapter 7).
These authors point out that visual images also ‘represent people, places and things
as though they are real . . . or as though they are imaginings, fantasies, caricatures, etc.’
(Kress and van Leeuwen, 1996: 161). The data in Table 2.3 are based on defined values
of modality according to visual features which have then been cross-tabulated by the
gender of represented participants (persons).
These data support the following conclusions (among others):
1 Males are more likely than females to be shown in ‘factual’ style advertise-
ments.
2 Most magazine advertisements (approximately two-thirds) adopt a standard
modality.
VA02post 14/9/00 10:25 am Page 21

CONTENT ANALYSIS OF VISUAL IMAGES 21

3 When more than one model is shown, gender makes no difference to

advertising modality.
4 Women are depicted in approximately 70 per cent of advertisements in the
magazines in question; men in approximately 30 per cent of comparable
advertisements.

Table 2.3 Modality and gender in magazine advertisements.

Modality
Person Standard Factual Fantasy Total

Male n 139 54 40 233

% 59.7 23.2 17.2 100.0
Female n 310 54 92 456
% 68.0 11.8 20.2 100.0
Body part n 27 2 11 40
% 67.5 5.0 27.5 100.0
Female group n 36 10 6 52
% 69.2 19.2 11.5 100.0
Male group n 29 9 8 46
% 63.0 19.6 17.4 100.0
Total n 541 129 157 827
% 65.4 15.6 19.0 100.0

Such conclusions are relevant to a visual content analysis based on theoretical

concepts from semiotics. As has been emphasized earlier in this chapter, any
theoretically specific variable and the values of which it is comprised, can be quantified
provided that it is explicitly defined and can therefore be reliably coded.

RELIABILITY

‘Reliability’ refers to the degree of consistency shown by one or more coders in

classifying content according to defined values on specific variables. Reliability can
be demonstrated by assessing the correlation between judgements of the same
sample of relevant items made by different coders (‘inter-coder reliability’) or by
one coder on different occasions (‘intra-coder reliability’).

Measuring reliability

As content analysis claims to be objective and therefore capable of being replicated,

it must be reliable or consistent if its results are to be of value. Reliability is a simple
VA02post 14/9/00 10:25 am Page 22

22 HANDBOOK OF VISUAL ANALYSIS

but important concept. It can be thought of as a quantitative index of the consistency

or lack of measurement error in a particular content analysis. Obviously, if a tool
used for measuring any quality (say the lengths of objects) does not yield the same
value each time it is applied to a given object, then it is not a reliable or consistent
tool. Although content analysis is not literally a tool in the physical sense, it is a
method of classification and of quantification, so its definitions must be precise
enough to be used reliably.
To achieve high levels of reliability, the researcher must:

(a) Define the variables and values clearly and precisely and ensure that all
coders understand these definitions in the same way.
(b) Train the coders in applying the defined criteria for each variable and value.
(c) Measure the inter-coder consistency with which two or more coders apply
the criteria (definitions) using a set of examples similar to, but not part of,
the research corpus.

If only one coder is to be employed a pilot (trial) study should be conducted

to measure intra-coder reliability. Have the coder classify 50–100 examples on all
relevant variables. A week later, repeat the procedure (without, of course, referring
to the previous results). Correlate the two sets of classifications.
Several methods have been devised for quantifying reliability. The two which
require the least mathematical sophistication are per cent agreement and ‘pi’.

Per cent agreement

Simply calculate how frequently two coders agree in their judgements, ensuring that
they are both tested on the same, large number of items (that is, on a representative
range of variables and values). For the mathematical, an index can be calculated as
in the example of Table 2.4, which tabulates (hypothetical) data from a pilot trial of
100 classifications with four coders.

Table 2.4 Hypothetical data from a pilot trial of 100 classifications with four coders.

% Agreement Coders

80 1, 2
60 1, 3
90 1, 4

Calculating reliability – average the three percentages:

Reliability = (80 + 60 + 90)%/300 = 76.7% (0.767)

If these were actual results, the researcher might choose not to employ the
most ‘aberrant’ coder in the next stages of the research project, so that the average
reliability (between the other three coders) was higher (0.85). Even here, however,
the level of reliability is below what is usually recommended, namely 0.9 or 90
per cent.
VA02post 14/9/00 10:25 am Page 23

CONTENT ANALYSIS OF VISUAL IMAGES 23

With four coders, such an index is rather artificial. It assumes (arbitrarily)

that one coder is the norm and averages the others’ respective agreements with this
norm.
There are two rather technical points to bear in mind. First, a high frequency
of items classified as ‘miscellaneous’ or ‘other’ will spuriously inflate the apparent
level of reliability. Less than ten per cent of items should fall into this category on
any variable. Second, the fewer values there are on a given variable, the more likely
there is to be agreement between coders which is based on chance rather than on
similar judgements according to the definitions. So binary or tripartite classifications
would need to be close to 100 per cent reliable.

‘Pi’: a more sensitive measure of reliability

Because the index (above) ignores the fact that two coders may agree in their
judgements purely by chance (and that this is, of course, a function of how many
values have been specified on particular variables), a better index has become more
widely used. For example, if a coder is asked to judge whether images show either
a male or female, 50 per cent agreement could occur by chance (without even
looking at the images). With five values on a variable, 20 per cent agreement would
be expected. And so on. But such agreement is not really indicative of reliability.
So Scott (1955) has proposed one of several more subtle formulae for assessing
reliability by taking account of chance agreements based on the number of values
distinguished in a given content analysis: ‘pi’.

Pi = (per cent observed agreement) – (per cent expected agreement)/(1 per cent
expected agreement)

However, this assumes that the researcher can state the expected percentages for all
values on all variables in advance of the coders’ judgements being made. This is
because the percentage of expected agreement is the sum of all the squared per-
centages of all categories. So if there are six values on a variable, and if the expected
and obtained percentages are as below, the pi coefficient or index would be calculated
thus:

Variable: occupational role Expected frequency E.f. squared

Business 20% (0.20) 0.0400

Hospitality 10% (0.10) 0.0100
Sport 30% (0.30) 0.0900
Academic 20% (0.20) 0.0400
Leisure (non-sport) 15% (0.15) 0.0225
Other 5% (0.05) 0.0025

Expected frequency = sum of the squares of the e.f. values = 0.2050.

If two coders classifying a sample of 100 images depicting occupational roles

in advertisements, for instance, agreed on 95 per cent of cases, then the reliability
index would be: (0.950–0.205)/(1.00–0.205) = 0.94 (the ‘pi’ value).
VA02post 14/9/00 10:25 am Page 24

24 HANDBOOK OF VISUAL ANALYSIS

However, if the obtained agreement between the coders in the pilot or trial
were found to be low, or if there were fewer or less evenly distributed ‘expected
frequencies’, the value of the index would be reduced. As a rule of thumb, a ‘pi’
value of at least 0.80 should be obtained. If this is not achieved in a pilot trial, the
researcher should re-train coders and/or redefine values and re-test the degree of
reliability.

LIMITATIONS AND EXTENSIONS

The main limitations of quantitative content analysis concern the relatively un-
theorized concepts of messages, texts or ‘manifest content’ that it claims to analyse
objectively and then to quantify (see Chapter 4).
The categories of visual ‘content’ which are most frequently quantified in
media research arise from commonsense social categories, such as ‘roles’ depicted,
‘settings’ shown, gender and age of represented participants in images. Such
variables are not defined within any particular theoretical context which analyses
visual semiotic dimensions of texts. That is, the framing, visual ‘angles’, scale of
photographic ‘shot’, and so on, that are part of the discourse of visual analysis are
seldom incorporated into (visual) content analysis (Chapter 7 discusses these in terms
of ‘semiotic resources’). Nor are categories from, say, Marxist or neo-Marxist theory
(to take another contentious theoretical paradigm) seen as appropriate to quantifi-
cation. Indeed, T.-W. Adorno (the famous cultural critic) has quipped that ‘culture’
is, by definition, not quantifiable. Other critics of content analysis point out that the
inferences that are made from quantification to qualitative interpretation (especially
those involving notions of ‘bias’) are fraught with difficulties. The cultural complexity
of visually coded texts means that either only the most simplistic, socially conventional
cate-gories can be studied, or content analysis imports tendentious or highly inter-
preted abstractions into its ostensibly ‘objective’ definitions of variables and/or values.
Winston (1990) argues that research purporting to demonstrate televisual bias against
businessmen fails because, in part, it fails to understand the (semiotic) codes of
television – the way television means to its audience. The result is that coding
categories used in research into this question were frequently moralistic and decon-
textualized. For instance, whether a programme depicting a businessman is a comedy
or a drama would be relevant to how an audience ‘reads’ that character’s behaviour
as ‘negative’ or ‘positive’. However, Crooks, Conmen and Clowns, the study criticized
by Winston, ignored the ways in which different genres are understood by their
respective audiences. (The genres and semiotic codes of tele-films are discussed in
Chapter 9.)
In a widely influential paper, Stuart Hall (1980) makes a similar point: ‘violent
incidents’ in cinematic genres like the Western are meaningful only to audiences who
know the genres’ respective codes – the goodie/baddie opposition, the conventions
relating to resolving conflict, and so on. In short, content analysis cannot be used as
though it reflects unproblematically or a-theoretically the social or ideological world
outside the particular context of the medium studied. Second, content analysis cannot
VA02post 14/9/00 10:25 am Page 25

CONTENT ANALYSIS OF VISUAL IMAGES 25

be easily compared with some assumed ‘reality’ by which to make claims of ‘bias’ or
‘negative’, let alone ‘true’ or ‘false’, representation. Winston (1990) discusses these
‘inference’ problems in convincing detail. Third, generalizing from content analysis
results can be difficult, and claims made for the consequences of the quantitative
picture of media content may go beyond what is validly licensed by the data. For
example, it is sometimes assumed that users (viewers, audiences) understand or are
affected by texts, genres or by media content generally in ways that reflect the kinds
of analytical categories used in content analysis itself. An example is that of ‘bias’ in
television reporting of political issues. It is often assumed that more coverage of one
political party than its competitors means that viewers will be more likely to favour
it. But this is, obviously, an unwarranted inference. At the very least, such an inference
would require additional research in order to be validated.
Visual representations raise further theoretical problems of analysis. Many
highly coded, conventional genres of imagery (the footballer saving a goal, the ‘cover
girl’ from Cleo, the politician engulfed by microphones, the academic expert in front
of rows of books) have become media clichés. To quantify such examples is to imply
that the greater their frequency, the greater their significance. Yet, the easy legibility
of clichés makes them no more than short-hand, stereotypical elements for most
viewers who may not understand them in the way that the codes devised by a
researcher imply (see ‘Looking: form and meaning’ in Chapter 4 on page 70).

VALIDITY: GOING BEYOND THE DATA

To conduct a content analysis is to try to describe salient aspects of how a group

of texts (in our case, images or visual texts) represents some kinds of people, processes,
events, and/or interrelationships between or amongst these. However, the explicit
definition and quantification that content analysis involves are no guarantee, in
themselves, that one can make valid inferences from the data yielded by such an
empirical procedure. This is because each content analysis implicitly (or, sometimes,
explicitly) breaks up the field of representation that it analyses into theoretically
defined variables. In this way, it is like any other kind of visual or textual analysis.
Semiotics posits as semantically significant variables such as ‘modality’ or ‘represented
participants’, or conceptual versus narrative image elements (see Chapters 4 and
7). All systematic methods of visual analysis (even psychoanalytic or iconographic
– see Chapters 5 and 6) postulate some features of images (and not others) as sem-
antically significant, within the images’ usual contexts of exhibition and reception.
Content analysis does this also. What is at stake in conducting this, or any form
of visual analysis, is the degree to which the resulting statements about the field
analysed can be said to describe features that are, in fact, semantically significant
for viewers/spectators/‘readers’ of the images. In other words, does the analysis
yield statements that are meaningful to those who habitually ‘read’ or ‘use’ the
images?
This brings us back to the point at which this chapter began. Whenever we
generalize about differences between sets of images, we must implicitly define as
VA02post 14/9/00 10:25 am Page 26

26 HANDBOOK OF VISUAL ANALYSIS

well as observe and quantify the visual material in question. We might, for example,
claim that the front pages of broadsheet newspapers show more, although smaller,
images, of more active participants, than do tabloid papers published concurrently.
Such a claim is a hypothesis which content analysis could test. To accept the claim
as true is to accept that something like a systematic content analysis could be
conducted and that it would demonstrate the differences claimed.
The criticism that is most frequently levelled against content analysis is that
the variables/values defined (the ‘categories’) are somehow only spuriously objective.
It is claimed that they are as subjective as any semantic variables despite being
‘measured’ or at least counted. However, such a criticism can be turned around, to
point to the fact that not only content analysis but all visual or verbal semiotics, formal
and informal, are only as valid as the explicitness and reliability of their respective
theoretical concepts.
To make inferences from the findings of a content analysis or any kind of
theoretical analysis of a group of images means that one goes ‘beyond the data’ making
a prediction about the salience, the social or ideological importance, the visual
significance of one’s findings. So it is best to think of these findings as ‘conditional’ –
as bound by their theoretical and methodological contexts, but as ‘true’ until and
unless they are contradicted by further evidence.
New theories will propose that different variables of images are semantically
significant and posit new definitions by which new generalizations can be tested. The
principal virtue of content analysis is that it is explicit, systematic and open to such
theoretically motivated, but empirically grounded critique. Thoughtful, provisional
content analysis need be no more ‘positivistic’ or ‘pseudo-scientific’ than competing
methodologies, but it is important to propose the findings of such analyses as
conditional to their context. Some questions are not usefully analysed by such methods,
and many claims have been made for content analysis which are impossible to justify.
But, as one approach among others to describing how and what images mean, content
analysis can be a useful methodology. The question of how one can generalize or ‘go
beyond the data analysed’ (as I have put it) is the problem of ‘validity’, to which I
now turn.
Content analysis, by itself, does not demonstrate how viewers understand or
value what they see or hear. Still, content analysis shows what is given priority
or salience and what is not. It can show which images are connected with which, who
is given publicity and how, as well as which agendas are ‘run’ by particular media. Or,
to put these claims more cautiously, content analysis can demonstrate such patterns
of media representation provided that one accepts the validity of the categories (values
and variables) defined in the research.
‘Validity’ refers to the (apparently circular or tautological) concept of how
well a system of analysis actually measures what it purports to measure. Valid
inferences from a particular content analysis, given this definition, will reflect the
degree of reliability in the coding procedures, the precision and clarity of definitions
adopted and (a less obvious factor) the adequacy of the theoretical concepts on which
the coding criteria are based. Validity refers to the confidence one can have in the
results showing that the stated theoretical concepts offer a discriminating description
VA02post 14/9/00 10:25 am Page 27

CONTENT ANALYSIS OF VISUAL IMAGES 27

of the field being analysed. ‘Images of violence’ (to return to a controversial example)
will be defined in different ways by different theoretical and pragmatic interests. For
instance, a media student might see the televising of accidental causes of physical
injury or death as a depiction of ‘violent’ episodes. However, an alternative theory of
televised violence might exclude accidents, arguing that ‘violence’ should only refer
to what, in their dramatic contexts, are visibly intentional actions which are shown
to cause physical and/or psychological harm. The relative validity of a particular
content analysis, given theoretical differences in the terms in which an hypothesis
can be formulated, refers to the degree to which inferences are justified from the
findings to the theoretical statements. Whereas reliability refers to ‘internal’ consistency
in one’s method, validity refers to the external or inferential value of one’s research,
given its theoretical context. This cannot be quantitatively assessed, but it is important
to ask of one’s own (and, indeed, of others’) research whether it actually demonstrates
what it purports to demonstrate, whether the variables and values do allow one to
‘measure’ the concepts incorporated in the hypotheses and their respective theories.
As we have seen, content analysis provides a quantified dimensional descrip-
tion of fields of representation. The methodology can be used to provide a background
‘map’ of a domain of visual representation. Having conducted a content analysis, the
researcher can then interpret the images or the imagery in qualitative ways (using
semiotic or some other more individual, text-oriented theory such as those described
in the other chapters of this volume). Typical or salient examples can be further
analysed to fill out the qualitative description of ‘what the data mean’. So, having
shown how frequently and in what contexts, say, images of passive females occur, a
researcher might discuss the psychoanalytic or ideological significance of the images
in terms of metaphors, photographic style, historical or social context (see Chapters
4 and 7).

AN EXAMPLE: TESTING SEMIOTIC HYPOTHESES

Images carry connotations and invite individual reminiscence. They may convey
a sense of duration or of nostalgia through codes of colour, framing and through
their public context. An image can engage the viewer in a fetishistic and compulsive
urge to look and look again, encouraging the sense that the viewer ‘owns’ the image
or that it is part of his or her ‘identity’ (see ‘Looking: recognition and identity’, in
Chapter 4, page 83). None of these experiential possibilities can be defined and
quantified very reliably. Yet dimensions of ‘interactive’ meaning (involving how
the viewer is invited to relate to an image) and of ‘textual’ meaning (including how
images are formally composed or balanced) can be defined and their frequency
counted.
In Chapter 7 of this book, the systemic functional (semiotic) approach to
image analysis is presented. By definition, it is the ‘manifest content’ of images that
content analysis dissects and counts, so any unambiguously definable aspect of a
group of images can be quantified. Insofar as semiotics also involves an empirical
(observational) methodology, it is possible to use the analytical concepts derived from
VA02post 14/9/00 10:25 am Page 28

28 HANDBOOK OF VISUAL ANALYSIS

this type of theory as the basis for quantification and, hence, for comparative
generalization. So to conclude this chapter, I will illustrate how it is possible to
quantify some important dimensions of what Kress and van Leeuwen (1996) call the
inter-personal semiosis of images focusing on depictions of women on women’s
magazine covers. I will use their semiotic model to define three variables, each with
several values. These are meant only to illustrate the way in which a content analysis
of explicitly semiotic variables could be conducted.
The specified variables and values will be used to code the two sets of twenty
cover-page images from Cleo magazine reproduced as Figures 2.1 and 2.2. These will
be the basis of a partial or small-scale content analysis to illustrate that:

1 Comparative hypotheses can be formulated and tested using reliable categories

relating to the semiotics of visual images.
2 Content analysis can be conducted by giving precise definition to theoretical
concepts (in this case from semiotics).
3 Objective criteria need to be specified for categorizing visual material. The
criteria must be defined unambiguously.

This small-scale analysis of media content will be set out in a conventional

‘social science’ format, as one might find in psychology or sociology, for example:
Hypotheses will be formulated; a Procedure (or Method) will be described; Results
and discussion sections of a typical empirical report will be presented. Under these
headings, other conventional sub-sections will be named (such as variables, values,
reliability). So this ‘worked’ example of a theoretically motivated empirical analysis
will serve as a summarizing overview of the method of visual analysis (content
analysis) that has been the subject of this chapter.
Figures 2.1 and 2.2 show respectively twenty front covers of Australian Cleo
magazine from 1972–4 and twenty from 1996–7. These will be compared by means
of a content analysis.

Hypotheses
Following Kress and van Leeuwen (1996) it is hypothesized that the inter-personal
meaning of Cleo front pages has changed between 1972 and 1997 in the following
ways:

1 The later images present models as more socially distant than in 1972–4
covers.
2 The modality of the images in 1996–7 is lower than in the earlier period.
3 The models’ gaze at or gaze away from the camera is not different between the
two sets of images, but the pose of the models (their head, body dispositions)
are less ‘powerful’ in the more recent examples.
4 The 1996–7 models are more frequently blonde, less frequently brunette, than
in the earlier period.
5 The later covers depict models who look younger than do those from the ear-
lier period.
6 Models occlude the magazine logo more frequently in the later period.
VA02post 14/9/00 10:25 am Page 29

CONTENT ANALYSIS OF VISUAL IMAGES 29

Method: definitions and criteria

Let us begin by defining each variable and the respective values on each before using
these as criteria to code our small samples of images. For the sake of exposition, I
will simplify relevant theoretical semiological concepts (which are the basis for
the definition of variables and values in the terminology of this chapter).

Variable 1: social distance

Kress and van Leeuwen relate the represented social distance of participants in
an image (in our examples, the depicted models) to Hall’s (1966) discussion of
‘proxemics’ (the psychology of people’s use of space):

In everyday interaction, social relations determine the distance (literally and figura-
tively) we keep from one another. Edward Hall (e.g. 1966: 110–20) has shown that
we carry with us a set of invisible boundaries beyond which we allow only certain
kinds of people to come. The location of these invisible boundaries is determined by
configurations of sensory potentialities – by whether or not a certain distance allows
us to smell or touch the other person, for instance, and by how much of the other
person we can see with our peripheral (60 degree) vision. . . .
With these differences correspond different fields of vision. At intimate distance,
says Hall (1964), we see the face or head only. At close personal distance we take in the
head and the shoulders. At far personal distance we see the other person from the waist
up. At close social distance we see the whole figure. At far social distance we see the
whole figure. At far social distance we see the whole figure ‘with space around it’. And
at public distance we can see the torso of at least four or five people. It is clear that
these fields of vision correspond closely to the traditional definitions of size of frame
in film and television, in other words, that the visual system of size of frame derives
from the ‘proxemics’, as Hall calls it, of everyday face-to-face interaction. (Kress and
van Leeuwen, 1996: 129–31, italics added)

On this basis a variable of visual representation, social distance, can be divided

into six values:

(a) Intimate.
(b) Close personal.
(c) Far personal.
(d) Close social.
(e) Far social.
(f) Public.

These are defined in terms of how much of the (human) participant’s body is
represented in the frame of the respective image, as in Kress and van Leeuwen
above (see Chapter 9 for a discussion of social distance in relation to the analysis of
tele-films).

Variable 2: visual modality

Again, following Kress and van Leeuwen, visual modality results from
VA02post 14/9/00 10:25 am Page 30

30 HANDBOOK OF VISUAL ANALYSIS

The degree to which certain means of pictorial expression (colour, representational

detail, depth, tonal shades, etc.) are used. Each of these dimensions can be seen as a
scale, running from the absence of any rendition of detail to maximal representation
of detail, or from the absence of any rendition of depth to maximally deep perspective.
And on each of these scales there is a point that represents the way the given pictorial
dimension is used in what could be called standard naturalism. To the degree that the
use of a dimension is reduced, it becomes, at least in one respect, more abstract, ‘less
than real’. To the degree that is amplified, it becomes ‘more than real’, and we asso-
ciated this with a ‘sensory coding orientation’, an emphasis on sensory pleasure (or
displeasure, as in the case of ‘more than real’ horror images), and an attempt to come
as close as possible to a representation that involves all the senses. (Kress and van
Leeuwen, 1996: 256)

However, the authors argue that modality is context dependent. That is,
modality is coded according to particular ‘orientations’ appropriate to different
conventional domains of representation. The domains they distinguish include
scientific/technological, abstract, naturalistic and sensory. A ‘sensory’ coding orient-
ation is appropriate to images which provide sensuous and sensory pleasure to the
viewer. Highly saturated colour conveys high modality in paintings such as nude
or ‘still life’ genres and in most modern display advertising (see Chapter 7).
So a variable, modality, can be defined as the represented ‘realism’ of an image,
given the sensory coding orientation, based on degrees of colour saturation. Three
values could be distinguished on this variable: ‘high’, ‘medium’ and ‘low’. These
are relative terms and criteria may be difficult to specify. But, as has been pointed
out in the introduction to this chapter, all values must be defined clearly if reli–
able classifications are to be made, so, in principle, the case of modality poses no
unique methodological difficulties. We might, then, set the criteria for the three values
thus:

(a) High sensory modality: image uses highly saturated colours naturalistically.
(b) Medium sensory modality: image uses, for example, less saturated, ‘washed
out’ or ‘ethereal’ use of pastels.
(c) Low sensory modality: image is monochrome (black–white) only.

Variable 3: behaviour
Kress and van Leeuwen also discuss the way interaction between the viewer and the
people shown in images is affected by the gaze of the represented participants.
Simplifying this to allow our illustrative example to be outlined, the principal
differences between what they call ‘image acts’ relate to the ways in which images
make ‘offers’ or ‘demands’, as it were, to their respective viewers (see Chapters 7
and 9):

There is . . . a fundamental difference between pictures from which represented

participants look directly at the viewer’s eyes, and pictures in which this is not the
case. When represented participants look at the viewer, vectors, formed by partici-
pants’ eyelines, connect the participants with the viewer. Contact is established, even
if it is only on an imaginary level. In addition there may be a further vector, formed
by a gesture in the same direction. . . .
 Another Random Document on
Scribd Without Any Related Topics
is excellent. ‘It is impossible,’ he says, ‘to lay down any definition of
God which will even satisfy man’s conception of God.’ What, then, is
the good, he asks, of holding up this ‘magnified non-natural man’ of
yours, and asking me to fall down and worship it? Common-sense
revolts against such an idea and common-sense, dear Doctor, is, will
you not agree, for once right?”
“You surprise me, Sir Horace,” said Maddock. “Are you too going
to spend your time and trouble in demolishing the survivals of verbal
inspiration?”
“Certainly not! I am only trying to see wherein common-sense is a
safe guide as a biblical critic. We are agreed, then,—you, that is, the
Judge and I—that we must unite in opposing many of ‘the
statements which,’ as the Judge says, ‘the orthodox are pleased to
call evidence.’ Because, for instance (to continue with the Judge’s
own words), ‘the fallible man Paul says in a letter to Timothy that
the Scriptures were inspired, it does not make them so.’ We are
agreed here?”
“We are agreed here,” said Maddock, with deliberation.
“Or again, to take another instance, when Matthew and Luke, for
whatever purpose, strive in their genealogical tables ‘to give Jesus’ (I
always use the Judge’s words) ‘a divine origin, conceived of a virgin
by the Holy Ghost, and yet to connect him with David by making
Joseph the natural father of Jesus.’—are we not here faced by two
ideas which ‘no one short of an ecclesiastical metaphysician,’ or, as
you say, a ‘very bad critic,’ would or could ‘reconcile?’—We are still
agreed, of course.”
“We are still agreed—to a certain extent.”
“Nay, let us go further, then, and chime in with the Judge to the
effect that ‘on far stronger evidence (if evidence it can be called)
than that which supports’—let us say, almost all—‘of the events or
miracles’ of the Scriptures, ‘the Roman Catholic Church propound to
the world their miracles,’ which ‘the Protestant section of Christianity
reject as incredulous.’”
 “Proceed,” said Maddock.
“Nay, let us go further still, and notice how we no longer look on
the Genesis account of the Creation as more than allegory, of the
Flood as being strictly accurate; of the tower of Babel as, again,
more than allegory, and so on in many other similar cases. And how
in the same way we do not look upon the statements of Christ, and
after him of the author of the ‘Revelations,’ of the close approach of
the Apocalypse, as literal but only figurative. ‘The statement of
Jesus,’ as the Judge puts it, ‘as to his coming again before the then
generation have passed away does not mean that he will so come:
‘generation’ being merely used figuratively, but when he does come
he is still to come in the clouds of heaven, and with great glory,
sounds of trumpets, rushings of winds, and mourning of tribes; for’
(Gildea paused)—‘all this has not yet been falsified by the event.’
This is, I think, undoubtedly the conclusion at which common sense
arrives, but common sense is of course wrong.”
“Common-sense is wrong,” said Maddock.
“Common-sense too, as exemplified in this its typical blockhead
who cannot ever rise beyond the spiritual Rule of thumb and Three;
common-sense observes of the development of divine Truth, as
exemplified in the Christian theology of yesterday and to-day, that its
‘golden rule apparently is to adopt those interpretations’ of its
Scriptures ‘which best satisfy the exigency of the particular position
of the time being,’ and thus we have no further guarantee that the
God of to-day will be the God of to-morrow than that the God of
yesterday is certainly not the God of to-day. ‘Heaven forgive me,’
exclaims ‘that great poet and brilliant philosopher,’ Heine, ‘but I often
feel as if the Mosaic God were but a reflected image of Moses
himself.’ And we all remember with what contempt Taine speaks of
this God of Christianity, revised and amended to suit the latest
edition of scientific and historical discovery—rooted up out of the
earth and momentary intercourse with man—driven out of the
clouds and the occasional interposition of his strong right hand—
spied and telescoped from the radiant bowers of the stars, and
finally lodged out of sight, and all but out of mind, in the eternal
infinitudes of Time and Space! After all, then, may not our good
Judge have had, not of course a perception, but a faint glimmering,
of sapience, when he spoke of the position taken up by the orthodox
biblical criticism as critically ‘not only untenable, but absolutely
suicidal?’ The thought is, as we agreed before, simply appalling.
Spirits of Butler, Paley, Neander, Weiss, Westcott, Lightfoot, and
many another mortal or immortal immortal, rise and thunder ‘No!’
When this exponent of the average secular intelligence declares that
contemporary Theology is an impossible compromise between
Reason and Absurdity; that the Protestant is quite inconsistent who
with one face rejects ‘the events or miracles propounded by the
Roman Catholic Church because they involve a violation or
suspension of unvarying natural laws; because such things do not
happen, and because reason refuses to give credence to them,’ and
with another face accepts as truth the sojourn of Jonah in the belly
of some sea-monster (at present conveniently extinct, even to the
bones), or the communications of, what Gordon describes as,

‘that duffer at walls,

the talkative roadster of Balaam:—’

rise, I say, and in Olympian accents demonstrate to him and his

benighted audience, that these were but links ‘in the development of
divine Truth,’ and that ‘one lesson at a time of this difficult kind was
enough, and as history shows more than enough, for human
weakness.’”
“You are a treacherous and malicious young man,” said Maddock,
laughing in spite of himself, “and have no right to quote my words in
such an irreverent and grotesque manner!”
“It is my orthodox ingratitude,” said Gildea, “—And yet,” he added
suddenly, with a complete change of tone and manner, “in less than
fifty years polemics like these will be looked upon as childish, and,
those who spent their life and energy upon them, as we now look on
the mediæval Schoolmen. It is a sad thought.”
Maddock was a little puzzled at these swift chameleon changes in
his friend.
“And now,” said Gildea, looking up with yet another change of
tone and manner, “and now we have done with the negative side of
the good Judge’s criticism and can turn to the affirmative.—But that,”
he added, “must, I am afraid, be after lunch—if you will, Doctor?”
“I will,” said Maddock, “and you shall not then find me so passive,
for your treachery and malice are now quite laid bare to me.”
Gildea smiled.
“But not my loyalty and admiration? Believe me, Doctor, that, if it
were only for this one remark of yours, I could never fail in my
interest and gratitude to you. ‘Our blackfellows,’ you say, ‘had no
punishment for offences against their elementary ideas of purity but
spearing. And it was infinitely better that they should spear for
impurity than lose their first step towards a higher life.’ ... But here
we are,” he said, “This is the house. Fitzgerald and Hawkesbury have
to leave us soon after lunch. Mrs. Medwin and her niece, Miss
Medwin, are coming later to make tea for me, and then we are
going out for a sail in the yacht. Mr. Medwin is thinking of a
legislative career, and so Alcock is to be cultivated. Can you come
with us? You know how pleased it would make us all.”
The Doctor explained that he was due at his hotel at half-past
three to meet Mrs. Maddock, and both he and Gildea expressed their
due regrets at his not being able to make one of the party on the
yacht.

III.
Gildea led the way upstairs and ushered Maddock into the sitting-
room. It was in reality two rooms joined together by a large folding-
door, which was now thrown open and draped with four looped-up
curtains, two of some dark-red material behind two of delicately-
wrought muslin. The two rooms were of the whole depth of the
house, the large bay-windows, open and with a glass-door in the
middle of them open also, at one end looking out over the city, at
the other over the harbour. A grass-slope, and a garden with flower-
beds and rustling trees, spread all round and down to the water’s
edge; while, a little way out, the “Petrel” rode at peaceful
anchorage, her boat behind her. Maddock was for the moment so
taken up with the beauty of the place within and without—the room
with all its harmonies of form and colour, the garden and harbour
scene—that he did not notice that someone was standing, half
hidden by the curtains, in the next room on the hearth-rug. Then
Gildea passed through and greeted this person whom he brought
forward and introduced to Maddock as Mr. Hawkesbury.
Hawkesbury was a small but well-made man with a tendency to
muscular leanness. His face was striking and interesting, and
betrayed a strongly-defined individuality. At one moment he might
have been called handsome, and his manner frank, free, and open:
at another his features took such a contracted intensified look, and
his movements were so nervously acute, that the whole man
seemed to have suffered distortion. It seemed as if he were
suddenly seized by some keen pain, spiritual and physical, and was
being racked by it. When Gildea entered, there was for a moment a
trace of this latter manner in Hawkesbury: his sensitive pride found
something antagonistic in, what seemed to him, the consummate
luxury which surrounded him and even in the consummate culture of
its owner: he was almost asking himself what right this man had to
spend so much money and care in decorating a few rooms for a few
months, this man whose life was so radically selfish? Hawkesbury’s
was, he might have said, the feeling of one who was a socialist and
worker by intense conviction, finding himself opposed to one who
was an aristocrat and hedonist by the mere chance of birth and
fortune. But, when Gildea met and greeted him with the frank sweet
unconscious cordiality of an equal whose acquaintance is pleasant,
the dark look passed from Hawkesbury’s face and he gave himself
up to the simple pleasure of the situation. His unexpected
introduction to Maddock, who represented to him the more or less
sumptuous aristocrat of religion, for a moment, it is true, threatened
to bring back the evil spirit to him; but Maddock, with his fine social
tact, almost divining the state of affairs, was equally frank, sweet,
unconscious and cordial in his manner, and Hawkesbury was at his
ease.
The three men stood talking together, Maddock in the middle, in
the bay-window that looked out over the harbour.
“Why, Sir Horace,” said Maddock, “you will never be able to get
away from this enchanting place again! Are you sure you do not
intend to make it into a home? You did not honour your Melbourne
rooms with such care—such choice of furniture, and....” (He raised
his arm and outspread hand, smiling humorously).
“‘Man delights not me,’” answered Gildea, “‘No, nor woman
neither, though by your smiling you seem to say so.’” The smile
broke out on Hawkesbury’s face too. It was soothing and very
pleasant to find these two talking in his presence of such an intimate
matter as that alluded to here: he was not accustomed, in the
company of, what in Australia and even England goes by the name
of, ladies and gentlemen to this complete absence of social and
individual constraint.
Then Edgar, Gildea’s valet, ushered in someone else, Mr.
Fitzgerald, and there was a movement and introductions between
Maddock, Hawkesbury, and the new-comer, the three being left
alone for a moment while Gildea was giving some directions to
Edgar about domestic arrangements.
Maddock and Fitzgerald fell almost immediately into a
conversation, Hawkesbury playing the part of silent member. The
Doctor was interested in finding out what the impressions of a
cultured Roman Catholic were of Australia and more particularly of
Victoria and New South Wales. He asked a few questions, the
answer to which, he thought, would show him whether Fitzgerald
had observed things with care and sympathy, and was answered
with a gentle readiness that pleased and satisfied him. The two men
felt themselves to a certain extent on common ground, and,
Fitzgerald touching incidentally on the education question, they
began to parallelise each other’s views with cordiality.
“We quite recognise,” said Fitzgerald, “all the difficulties of the
case—the danger of the unfair influence of catholic teaching over
protestant children, or vice versa, just as each happens to be
stronger in the particular place and school. But we would accept this
danger—accept it, even supposing we were the losers by it—rather
than have the present state of things continue. As our Archbishop
said only the other day at Leichardt: ‘Besides the faculties of intellect
and of reason, there are certain passions of the soul,’ and to develop
the former and wholly neglect the latter is to send a boy out into the
world with only one eye. You have prepared him for the temporary
business of life, and unfitted him for the glorious service of eternity:
you have given his ship fine sails, and forgotten to add a rudder! He
may be an acute man of business, but he will be a bad citizen; for, in
taking away from him his sense of religion, you will take away from
him his sense of morality, of honesty, of integrity! We can, at the
present stage, see for Australia no future save that of corruption—a
corrupt political life, a corrupt national life, the unlimited worship of
Mammon!”
“I agree with you to a large extent,” said Maddock, “and we all
know that, practically speaking, the talk about ‘religious education at
home’ is mere verbiage. If the education of a child is secular, his
spiritual lungs, so to speak, end in being able to inhale no other air
and thrive on it.”
“And,” Fitzgerald said, “the education is secular! Every effort is
being made to drive the voluntary schools out of the field. Their
state aid here in New South Wales is withdrawn: in England it is
reduced to a pittance and hedged about with annoyance. And this,
although the educational reports, drawn up by a secular commission,
show that, at any rate the catholic schools educate on the average
both better and more cheaply than the state-schools do! We only
ask for fair play, and now it has come to this pass that we cannot
get it! All over England the protestant voluntary schools are failing
and disappearing. But we, we Catholics, who cannot, as Protestants
do, console ourselves with the reflection that the atmosphere of the
state-schools, if secular, will be tempered by that of our own beliefs
—we will not fail and disappear! We are the poorest of all religious
bodies in England; but I will venture to say, that not a single case
can be found of a catholic school which has surrendered itself up, as
these others did, into the hands of the Secularists. Our educating
priests and laymen have to suffer much privation: I know, shall I say
hundreds, of them who deny themselves all but the bare necessities
of life; but—we stand our ground!... You see,” he added smiling
gently, “we Catholics cannot labour under any delusion here. We
recognize that this is a stupendous crisis in the world’s history. We
will have no compromise and secular tempering of the wind to the
shorn Christian. We will stand to our guns, and, if we must perish,
perish there!”
Maddock was impressed, and so even was Hawkesbury. This
man’s enthusiasm was so quiet, so clear, and yet so radiant. Gildea
returned and joined them.
“We were speaking of the popular education,” said Fitzgerald,
turning to him, “and I would persuade Dr. Maddock that his cause
and ours are here identic.”
“I need no persuading,” said Maddock, “I have for some time been
persuading myself!”
“And yet,” Fitzgerald put in gently, “the alliance between us and
you seems farther off than between us and the Dissenters.”
“And that, I think,” Gildea said, “is because you have more in
common. You are afraid of one another. In the one case, you know
that the frontier of your alliance will be observed, in the other there
is a chance that it may not. At present the most dangerous
opponents of Catholicism in England are, what they call, the High
Churchmen. The Church of England is a compromise between
Catholicism and Protestantism; hence its adaptiveness, hence its
strength! It more nearly, in my opinion, approaches ideal Christianity
than any other sect in existence. It unites the Faith, the Poetry, of
Catholicism, with the Freedom, the Prose, of Protestantism.”
“We thank you,” said Maddock.
“Logically speaking, however,” added Gildea, “it is an absurdity.”
They all began to laugh.
“Ah,” said Maddock, “I was right when, even while thanking you,
Sir Horace, I thought to myself: Timeo Danaos, et dona ferentes.”
“The Christianity of the Future,” Gildea proceeded gravely, “lies, I
believe, in two transformations—in Catholicism learning that its
kingdom is not of this world, that it no longer requires a Pope, a
Rome, as a Palladium whereby it may fight; in a word, in learning
the lesson of Protestantism, of Freedom: and in Protestantism doing
the converse, and absorbing into itself the catholic Faith, the catholic
Poetry!”
“And what are the Secularists going to do in your Future?” asked
Hawkesbury, “are Messrs. Arnold and Huxley to be put up on a shelf
in your spiritual Museum, in two large spirit bottles, labelled
respectively ‘Culture’ and ‘Science?’”
“Culture,” answered Gildea, “is, after all, but Secular Catholicism,
just as Science is but Secular Protestantism. They too will each learn
their lesson of the other.”
“Humph!” said Maddock, who again had a faint suspicion that
Gildea was mocking, “and so, after all, Sir Horace is an optimist.”
“We do not lay stress,” Fitzgerald said gently, “on the temporal
power of the Holy Father. As Sir Horace implied, this temporal power
was once the one shining light in a chaotic world, and it was well
that it should be set on a hill. But now the light is diffusing itself. It is
our wish that, as the Vatican Œcumenical Council declared:
‘Intelligence, Knowledge, and Wisdom may grow and perfect
themselves—as much with the mass as with individuals, with one
man as with the whole church!’ We are no foes to Freedom. What
we are foes to, is Anarchy! At the Reformation you gave the right of
deciding on the deepest religious questions to every ignorant man
that chose to discuss them, and the seamless robe of Christianity
was rent into a hundred pieces! Look at all these miserable little
protestant sects and sub-sects, Plymouth Brethren, Primitive
Methodists, Ana-baptists, and I know not what noisy, ignorant
fanatics. At the Revolution, you did the same for social questions,
and what is the result? The Dynamiters of Russia, of Germany, of
Ireland, initiated by what you, Dr. Maddock, so well call ‘such
gentleness as was revealed in the diabolical deeds of the
Commune,’—to say nothing of those of the Reign of Terror.”
Maddock half-deprecated, half-approved by a gesture and an
inarticulate sound.
“Yes, but,” said Hawkesbury with the thrilled voice of suppressed
passion, “has not history justified the Reformation? and how can you
say that it will not justify the Revolution? These, as it seems to me,
are the two fiery portals which lead to Religious and Social Liberty.
But you are right to depreciate them: they knew nothing of the
poetry of Culture and Catholicism, or of the prose of Protestantism
and Science. They were volcanic eruptions of the People. Heine says
well, when he talks of ‘the divine brutality’ of Luther, and we do not
shrink from the same phrase for Hugo or Whitman. Sir Horace has
painted us a Future which is indeed heavenly. It is thronged with
sweet-singing angels, and there is not a shadow in its perfect light.
But what has become of the men, and what, O what, has become of
the devils? They have no place in this Future. You do not care for
the People, I say, except as you care for your dog which, if he is
quiet and docile, shall have a kennel and the bones and scraps from
your table; or, if he is surly, shall be chained up; or, if he goes mad,
shall be shot! Ah believe me, gentlemen, the People has a place in
the Future, for the People, and none other, is the Future! ‘All for the
modern,’ cries Whitman, ‘all for the average man of to-day.’ But you
—you only care for the Upper and the Middle-class. Your scheme of
civilization does not reach to the People. The Upper-class is
exhausted: it needs invigorating. ‘Cultivate the Middle-class,’ is the
cry, ‘Give us Higher Education for the Middle-class!’ This is the whole
social teaching of the best representative man you have, Matthew
Arnold. Now we, we Socialists as you call us, love the People, and
(you will pardon me) hate the Middle-class;—the dispossessed, the
sufferers, not the possessors, the usurpers! The People is the
Prodigal Son. What sympathy have we, then, with a man like Arnold
who has devoted himself to the edification of the Elder Brother?
Arnold says once that he has evolved that perfect style of his which
we know so well—that style which encloses a minimum of ideas in a
maximum of catch-words—or, as he likes to call it, ‘plain popular
exposition’—for the especial benefit of the British Philistine, the
divine Middle-class, who otherwise could not be got to read him! He
would have done better, perhaps, if he had not turned to the setting,
but to the rising sun. The People are the masters of the Future, and
the People’s great men will be the great men of the Future.”
There was a pause. Then:
“There is much truth in what Mr. Hawkesbury says,” says Gildea,
“Just at present we think too much of the ultimate Culture of the
Middle-class and too little of that of the People. But the fact is, that
the question of the Middle-class is pressing: they are, as you say,
Hawkesbury, the possessors; they are the Present! And this, I think,
is why men like Arnold, who believe that, in the organization of the
Present, lies the only hope of the success of the Future, are so
anxious about it. It is a case, as he believes, of ‘Culture or
Anarchy’—Culture now or Anarchy then. And Carlyle, a disciple of
whom Mr. Hawkesbury has, in the admirable Preface to his second
book of Poems, declared himself to be; Carlyle too, who laid much
stress on what he calls ‘the radical element’ in himself, yet mocks at
‘Mill and Co.’ as he says, in whom he declares the opposite element
was ‘so miserably lacking.’ Carlyle had no respect for ‘Rousseau
fanaticisms,’ even in a man like Mazzini: he saw that, if the Middle-
class were purblind and slow, the Socialists were only purblind and
quick. Supposing that we grant that the Dynamiters of Russia are
justified in meeting an absolutely dense despotism with violence,
what excuse but impatience can we find for the Dynamiters of
Ireland? The first have no means of free agitation, the second have
every means. Ireland has been wronged: no one denies it; and
never, in the whole course of her history, has England shown such
alacrity as she is doing now to right the wrong; never, not even for
herself. But the Irish Socialists are impatient: their cry is for
everything to-day, this very hour! To grant it them would be the
greatest unkindness possible. Well, they too have taken to dynamite
as a hypochondriac takes to opium. The Russian Nihilists are noble
people, none nobler, but they taught fools and knaves an appalling
lesson when they inaugurated the reign of terror in Petersburg. At
the present moment, as Heine clearly foresaw, the Civilization, not of
Europe, but of the whole world is in danger.”
“You speak well, Sir Horace,” said Maddock, “and express my
opinions better than I could myself, but—Timeo.”
He, Gildea, and Fitzgerald smiled. Hawkesbury was grave. There
was a pause. Then:
“I think,” he said, “that you do the People wrong. These extreme
Socialists, the Nihilists as they are called, are not from the People,
but from the Middle-class. They are, as a rule, men who have
received the best education of the time, and who yet find
themselves unrecognized and unrewarded. Most of them are
journalists. It would astonish you, I think, to see the amount of
really first-rate talent that is being flogged to death in the shafts of
the modern Press. These men cannot work in shops and banks: the
narrow material life has been made impossible to them. The only
opening for the life they would—nay, that they must live, or perish,
is that of Literature. Literature caters for the Middle-class, the ruling
class. These men, then, are the slaves of the great caterers, the
newspaper editors. One of the most thorough Socialists I ever knew,
Holden, in fact, was on the regular staff of the English Spectator, the
organ of the enlightened portion of the Middle-class; and there, as
he said to me, he went as near Socialism as he could for
threepence! (Threepence is the price of the paper.) This same man
wrote, too, political articles for a distinguished radical politician, and
I have seen the proof-sheets of these hacked and mauled by the
patron to suit the palates of the Radicals. It was this man who once
seriously contemplated dropping a bomb in the House of Lords, to
show that herd of hereditary liars, as he put it, that there was such a
thing as justice in the world! He loved the People: he hated the
Middle-class, but the People cared nothing for him. It is, then, I
think, a mistake to lay the paternity of Nihilism to the charge of any
but the over-fed tyrannous Middle-class.”
“What you say,” Maddock said slowly and courteously, “is very
interesting and instructive, Mr. Hawkesbury, and I perceive that the
ground which you, and I think I may say Mr. Fitzgerald,” (Fitzgerald
smiled and bowed), “and myself have in common is large enough to
admit of our working—at any rate not in opposition to one another.
Is not our mutual object the enlightenment of the unintelligent mass
of the People and of the Middle-class? I am, I am sure, grateful to
you, sir, for the manner in which you have brought this home to me.
I always felt that underneath all our differences—I mean, the
differences of our beliefs, religious or social—we had a common
ground, the advancement of a really good and true Civilization, and
now, I think, I know this. He renders us a great service who makes
our feelings self-conscious, who turns them into the articulate
thought of words.”
There was a slight pause.
“And now,” said Gildea, in his half-amused way, “we will, if you
please, go down to lunch. Mr. Alcock particularly asked me not to
wait for him, and we have waited, it seems unconsciously, for over
half-an-hour.”
They went down together into the dining-room, chatting lightly
and pleasantly.

IV.
 The dining-room was the corresponding room on the ground story
to the sitting-room up above. It was quite as well furnished, but in a
different style. A fine rather than an exquisite form of beauty had
been sought after. It was a saying of Gildea’s that a dining-room
ought to give you an impression somewhat similar to that of a
beach-brake in spring: the architecture and furniture should have
clear outlines, the colours should be clear, the lights should be clear.
All massiveness and duskiness was to be avoided. A meal ought to
be a repast, not a feast: we should rise pleasantly satisfied, not dully
satiated. In a sitting-room, on the other hand, the sworn abode of
the sweet and delicate talk and music of women, just as the dining-
room was that of the serene discussions of men, there should be
something of the lush luxuriance in shape and colour of the
midsummer woods, knights and ladies and all the figures of romance
and fairy-tale passing together. But such an arrangement of rooms
as this, he would say with his bright half-mocking smile, was at
present like a damsel of the Middle Ages suddenly awakened in the
dull derisive streets of London or Manchester. This will only come to
pass in that wonderful Future, when we have all learned that Beauty
and Truth are synonyms, and Keats has statues and altars like
Sophokles of old.
Considerable time, wealth and trouble had been spent on this
house. Sydney and Melbourne had been ransacked for beautiful
things worthy of Gildea’s ideas of “the nest,” as he called it to
himself, that he desired; for this was indeed one, and not the least
remarkable, of his freaks. It had been aroused in this fashion. One
afternoon, sauntering across a road in the Domain, he had almost
been run over by someone riding a splendid bay horse. Looking up,
with a fine touch of anger, he had perceived that it was a lady, who
was looking down at him with a look, he suddenly felt, so precisely
his own that, the ludicrous aspect of the thing coming upon him, he
smiled. She too, at once following his change of feeling, smiled, and
then in a moment, with a slight courteous movement of hand and
body, had passed. It had all taken place in a few seconds. Her face
and form made up between them, he thought, the most beautiful
woman he had ever seen, and he had not seen few so-called
whether in Europe or elsewhere. Beauty in women was, according to
Gildea, a thing which was not in reality to be seen in the present
world, implying, as it did, perfection of form and perfection of spirit,
καλον κἀγαθον. The Athens of Perikles had produced female beauty;
in the face and form of the Venus of Milo the highest physical and
spiritual perfection of the time is apparent. Florence too, in such a
woman as Vittoria Colonna, had produced female beauty, and the
Renascence had incarnated it in a Marie Stuart; but, so far, our
Modernity was not ripe for it. Lovely female faces it, as all times, had
in abundance, but these faces knew nothing of spiritual perfection:
they knew nothing of life, they were not beautiful. And the female
faces that did know of life, the faces of women like George Sand,
Charlotte Bronte, George Eliot, were quite wanting in physical
perfection. They imply mental passion, the struggle of pain: they
have not reached to the serene pleasure of spiritual sovereignty. No,
Beauty, καλον κἀγαθον, is to be a produce of the Future when
Modernity has passed through the pangs of its travail and, in the
bright light of health and youthfulness, “grows in wisdom and
stature” to the perfect self.—But this face that he had seen for a
moment, was, he thought, really beautiful.
A few yards from him a man was standing looking back at the
rider passing along under the trees. Gildea came to him, and asked
him courteously if he happened to know who the lady was?
“No,” said the man, “I don’t know who she is, but I often see her.”
And on this incident Gildea had founded a freak which had for
some time amused him. He intended to see this woman again, and,
if he was correct in his supposition (which he used amusedly to
doubt to himself) that she was some phenomenal anticipation of the
Future, to possess her. He set about choosing and furnishing a
house, therefore, which should, as far as possible, be worthy of such
an individual, and much amusement it occasionally afforded him. A
private enquiry-office was meantime seeking her out; and, about a
month ago, Gildea to his surprise had been informed that she was,
beyond doubt, a Miss Medwin, niece of the well-known squatter,
english, eccentric even to the extent of riding about and shooting in
man’s clothes on one of Mr. Medwin’s stations in New South Wales,
and, moreover, strongly suspected of having had, and of still having,
an intrigue with a Mr. Frank Hawkesbury, a writer and man of
uncertain means, in Melbourne. Gildea laughed much on receiving
this unasked-for report, (He had just by accident made the
acquaintance of Hawkesbury), and his interest in his freak somewhat
revived; but his all but conviction that he was incorrect in his view of
Miss Medwin (if it were indeed she), prevented him from having any
great interest in the matter or any great anticipations of success. As
usual, however, he was satisfied to find that he had any interest or
anticipations at all. He learned from Mrs. Medwin that she was in a
short time coming to Sydney for a week or so on her road up to one
of Mr. Medwin’s New South Wales stations to which she had not
been for years, and would be pleased to see him. A few days ago,
then, she and Miss Medwin had arrived, and were waiting for Mr.
Medwin who was detained by business in Melbourne. Hence Gildea’s
invitation to Mrs. Medwin and her niece, to come and make tea for
him and go for a sail in the “Petrel.”
The party arranged itself round the table, Maddock at one end,
Gildea at the other, an empty place on Gildea’s right hand for Alcock,
Hawkesbury on his left with Fitzgerald next to him. Maddock, as
before, could not help observing with admiration the beautiful room
in which they were sitting. Hawkesbury, however, following out a
train of thought suggested by his own last words, sat serious,
looking at the table-cloth.
The lunch began. Gildea and Fitzgerald could both, when they
pleased, excel in that graceful sweetness of manner which is
supposed to be the peculiar gift of women. They pleased now. The
talk flowed lightly and pleasantly, and soon returned to, what
seemed to be to them all, the most interesting topic—the People.
Fitzgerald spoke of the far greater ease and leisure of the People
here than in England, and that led on to a consideration of the
question of Labour here.
 “Carlyle declared long ago,” said Hawkesbury suddenly, “that the
great question of the time was no other than the organization of
Labour. Well, Labour is at last organizing. The consequence is that,
as Mr. Fitzgerald remarked, there is greater ease and leisure among
the People, not only here in Australia where Labour is comparatively
scarce, but even in England where it is plentiful.—The question here,
however,” he added, “shows signs of complication. The employers
are to form—nay, have already formed—a union: ‘The Victorian
Employers’ Union.’ The only wonder is that it is in Victoria and not in
England that this idea has first been adopted. In Trades-Unionism in
England, let me say it at once, there have been many abuses; but,
let me hasten to add, not nearly so many abuses as there were
under the old despotism of Capital. Trades-Unionism, which so few
people seem to understand, originally meant the combination of
many oppressed small units against a great oppressing unit. Now it
means more: it means the determined effort of the People after
happiness.”
“That is very true, I think,” said Gildea, “The People, ever since
the deception practised upon them by the compromise Reform Bill of
’32, have been slowly learning to organize themselves and to rely on
themselves alone. Such a fact soon makes itself apparent. There is
not a single considerable political measure since ’32 which has not a
socialistic tendency.”
Hawkesbury acknowledged Gildea’s remark, and proceeded:
“The People, and by the People I mean of course the masses, is
everywhere realizing that there is something better worth living for
than frantic competition and the scramble for wealth. Trades-
Unionism, then, is the sworn foe of all this. I am not speaking either
for or against Trades-Unionism: I am simply stating what it wants,
what it is! The Trades-hall delegates, in the late conference anent
the Bootmakers’ strike in Melbourne, refused to let a bootmaker
work for more than eight hours a day, although, by so doing, he
might better himself, and by not so doing might keep himself for
ever a mere journeyman. ‘Further argument with men of such a way
of thinking,’ says Mr. Bruce Smith, the chief mover of the ‘Victorian
Employers’ Union,’ ‘further argument seemed useless.’ And it was
indeed as it seemed; for these men were of opinion that if, in the
frantic competition and scramble for wealth, one or two journeymen
did rise and become rich, hundreds and thousands would have to
lead lives which would not stand too favourable a comparison with
those of dogs. ‘Therefore,’ the delegates would say, ‘we will check
this frantic competition and scramble for wealth, and we will even be
so wicked as to sacrifice the one or two possible journeymen who
might rise and become rich, for the sake of the actual hundreds and
thousands whose lives otherwise would not stand too favourable a
comparison with those of dogs.’ Well, and what will be the end of
this new phase of the great battle of Capital versus Labour on which
we seem to be now entering here? Let me not be thought a
terrorist, if I remark, what is indeed patent to all, that, in a country
with a franchise like ours, Labour, if driven into a corner and
confronted by Capital triumphantly brandishing its sword of ‘Frantic-
competition-and-the-scramble-for-wealth—Labour, I say, might make
things excessively uncomfortable for the community in general and
Capital in particular. I am not hinting at mobs and sticks and stones.
I am merely stating a fact that is patent to all. Our good friends the
Landed-proprietors, videlicet the squatters, have experienced in
Victoria and elsewhere—are indeed now experiencing even in
Queensland[13]—the undoubted benefits of a little judicious
legislation. Might not someone suggest to the ‘Victorian Employers’
Union’ and Mr. Bruce Smith, who seem to have such quaint notions
of what Trades-Unionism really wants and is, that the same fate may
possibly be in store for our other good friends, the Capitalists?”
“It is a pity,” said Gildea smiling, “that we have not a Capitalist
here to answer you. But, I think, I know what one of them, Mr.
Alcock, would say. He would say that the great law of Nature is this
very frantic struggle which you deprecate, and that, if you attempt
to put a check on it, you will only end by first arresting and then
destroying all progress. He would oppose the interference of
organized Labour quite as much as of organized public opinion, that
is to say the State. He would of course recognize all the evils of the
frantic struggle, but he would say that it yet contained the great
ascending and progressive power of Nature, it was yet capable of
Evolution; whereas the artificial state of popular leisure and ease
contains the great de-scending and retrogressive power of Nature,
Dissolution.—But here,” he said, “at the very nick of time, he comes
himself.”
Edgar, who had just left them, returned ushering in Alcock, who
came forward with somewhat off-hand apologies to shake hands
with Gildea. He was then introduced to Maddock and shook hands
with him, compromising the matter, as he thought, with the others
by a bow and an expression of his pleasure at making their
acquaintance. He sat down in his place and, having told Edgar what
he chose to eat, was ready for a few moments’ talk before setting
somewhat vigorously to work on the victuals. Gildea explained to
him the conversational context, and what he himself had ventured to
say in the person of the typical scientific capitalist.
“Well,” Alcock said, with a half-pleased half-amused look on his
face, when Gildea had finished, “I will observe that, on the whole,
you didn’t put my sentiments so badly, Sir Horace.—I am opposed to
all state interference,” he declared, turning to Maddock, “It doesn’t
pay in the long run; it enervates people! Look at this New South
Wales here. They can’t put a bridge across a creek now, without
petitioning government for assistance! In England a half-dozen men
or so would have got together and settled the matter themselves.
And they want more state interference in Victoria! Why, it’ll drain out
all their independence, and energy; and, in twenty years, they’ll be
as lazy and lackadaisical as they are here in New South Wales!
Competition’s the law of Nature.” By this time Alcock’s mouth was
full, and he was beginning to enjoy the delicate food and wines, for
he was hungry and thirsty. There was a pause.
“True,” said Fitzgerald, gently breaking it, “but does not Mr. Alcock
too think, that it is just where the law of Nature ends that the law of
Humanity begins? Surely this is the essential position of Christianity,
that it says to the brutality of Nature: ‘Thus far shalt thou go, and no
further.’”
“You can’t,” answered Alcock with his mouth full, too intent on the
victuals to be more explicit, “You can’t interfere—impunity—great
law—nature—struggle—existence—survival—fittest.”
“Here, then,” said Fitzgerald who ate little and drank less, turning
to Hawkesbury, “we are at one, I think, as opposed to the pure
Scientists?”
“I do not believe,” Hawkesbury said, “and I do not think any
Socialist believes, in carrying the initiative of the individual to the
extent that Herbert Spencer would like. But we are not in favour of
state interference. We want to nationalize things, the land, the
unearned increment, the great public enterprises, but we include in
this term the State also. The State at present means the tool of the
Middle-class, worked by Capital and the Land Interest. This
arrangement partakes too much of the nature of a political joint-
stock company to please Socialists.”
“And you think,” asked Gildea, his hand on his wine glass, looking
at Hawkesbury, “you think that when the People wins, as it of course
ultimately will win, the control of things, that it will not work the
State in its own interest, just as the Aristocracy did and as the
Middle-class does?”
“You know,” Hawkesbury said, “I believe in the People! The People
is the only unselfish part of society. Their one desire is for justice
and mercy; and, when they could not get it themselves, they have
always died readily for those who, they believed, wished to give it
them. Herein lies the secret of all great popular devotions—from that
of Christ to that of Napoleon.”
“I,” said Alcock, “do not believe in the People, as you call them,
and their unselfishness has not yet come under my notice. The
People, like everyone else, are led by what they believe to be their
interests, their immediate interests, and our great effort should be,
by giving them a good sound practical education, to get them to see
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com