Using Documents

in Social Research
Lindsay Prior
Using Documents in Social Research

Copyrighted Material
INTRODUCING QUALITATIVE METHODS provides a series of volumes which introduce qualitative
research to the student and beginning researcher. The approach is interdisciplinary and inter­
national. A distinctive feature of these volumes is the helpful student exercises.
One stream of the series provides texts on the key methodologies used in qual itative
research. The other stream contains books on qualitative research for different disciplines
or occupations. Both streams cover the basic literature in a clear and accessible style, but
also cover the 'cutting edge' issues in the area.

SERIES EDITOR
David Silverman (Goldsmiths College)

E DITORIAL BOARD
"
Michael Bloor (University ot Wales, Cardiff)
Barbara Czarniawska-Joerges (University of Gothenburg)
Norman Denzin (University of I l l inois, Champaign)
Barry G lassner (University of Southern California)
Jaber Gubrium (University of Florida, Gainesville)
Anne Murcott (South Bank University)
Jonathan Potter (Loughborough University)

TITLES IN SERIES
Doing Conversational Analysis Focus Groups in Social Research
Paul ten H ave Michael Bloor, Jane Frankland, Michelle
Thomas, Kate Robson

Using Foucault's Methods Qualitative Research Through Case Studies

Gavin Kendall and Gary Wickham Max Travers

The Quality of Qualitative Research Researching the Visual

Cl ive Seale Michael Emmison and Philip Smith

Qualitative Evaluation Methods of critical Discourse Analysis

Ian Shaw Ruth Wodak and Michael Meyer

Researching Life Stories and Family Qualitative Research in Social Work

Histories Ian Shaw and Nick Gould
Robert L. Miller

Categories in Text and Talk Qualitative Research in Information System

Georgia Lepper Michael D. Myers and David Avison

Qualitative Research in Education Documents in Social Research

Peter Freebody Lindsay Prior

Copyrighted Material
Using Documents in Social Research

Lindsay Prior

SAGE Publications
London. Thousand Oaks. New Delhi

Copyrighted Material
© Lindsay Prior 2003

First published 2003

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

6 Bonhill Street
London EC2A 4PU

SAGE Publications Inc.

2455 Teller Road
T housand Oaks, California 91320

SAGE Publications India Pvt Ltd

32, M-Block Market
Greater Kailash - I
New Delhi 110 048

British Library Cataloguing in Publication data

A catalogue record for this book is available

from the British Library

ISBN 07619 574 6 4

07619 57472 (pbk)

Library of Congress Control Number 2002104873

Typeset by C&M Digitals (P) Ltd., Chennai, India

Printed in India at Gopsons Paper Ltd, Noida

Copyrighted Material
Contents

List of Figures viii

List of Tables vIII

Preface Ix

1 Basic Themes: Use, Production and Content 1

What is a document? 1
Diversity in documentation 5
Documents: production and function 10
Writers and readers - a dynamic relationship 16
Documents and their content 20
Conclusions and key points 26
Research exercises 27
Notes 27

2 Producing Facts 30
From Paris 1748 to Geneva 1998 30
Death: a progress report 32
Enumerating neurotics 43
Conclusions 47
Research exercises 48
Notes 49

3 Documents In Action I. Documents

In Organizational Settings 50
An ethnomethodologist in the archive 50
A Thai village circa 1968 52
Talk and text in the clinic 54
Documents and organizational activity 60
The politics of fog 63
Conclusion: how to do things with documents 67
Research exercises 68
Notes 69

Copyrighted Material
vi CONTENTS

4 Documents In Action I I . Making things visible 70

The tree of knowledge 70
Of fungi and fish 72
Making disease visible 73
Making risk visible 79
Conclusions: making work visible 86
Research questions 88
Notes 88

5 Texts, Authors, Identities 89

Authors. What use are they? 89
A dossier manufactures its subject 92
Autobiography as a technology of self 95
Sybil. A network of text and action 98
Identity as performed through writing - and reading 101
Conclusions 103
Research exercises 104
Notes 105

6 Content, Meaning and Reference 107

Mind, meaning and interpretation 107
Meaning and modernism 109
Referencing death 113
Referencing metaphors 115
Accounting for fatigue 116
Conclusions 122
Research exercises 123
Notes 123

7 Doing Things with Words 125

Traveller's tales 125
The scientific report 127
From Kiriwana to Inishkillane 134
Conclusions 141
Research exercises 143
Notes 144

Copyrighted Material
 CONTENTS vii

8 Documents as Evidence. Researching the

Inert Text 145
Conjectures and refutations 145
Evidence, evidence, evidence 148
The selection of evidence 150
The scope and robustness of data 154
Data extraction: indexing and coding 157
Conclusions: conjectural history 161
Research exercises 163
Notes 164

9 Production, Consumption and Exchange 165

Lessons from southern Sudan 165
Circuits of production and consumption 166
The sociology of the Kula 169
Documents as technology 171
A final plea 172
Research exercises 173
Notes 173

Bibliography 174

Index 189

Copyrighted Material
List of Figu res

1.1 Spatial organization as documentation 9

2.1 Suicide rates for Northern Ireland, 1 968-98 35

3 . 1 Facsimile o f a ward-based nursing assessment record 55

3.2 Extracts from a 'Statement of Special Educational Needs' 58

4.1 The identification o f fungi 72

4.2 Cancer in the family: the fabricated family's pedigree 83
4.3 Visualizing mutations through gels 85
4.4 A normal DNA sequence
4.5 Abnormalities in the DNA sequence

8.1 The growth o f risk as a focus for medicine 1 52

8.2 Funnel plot demonstrating the presence
of 'invisible' knowledge 1 56

9.1 Infor mation contacts among professionals in South Wales 170

List of Tables

2.1 Prevalence o f psychiatric disorders i n private

households by gender 43

6.1 The distribution o f death notices i n Belfast

newspapers in a 1 0% sample of deaths 1 15
6.2 Occurrence of selected words in a 2,31 5-word
patient support group leaflet on chronic fatigue syndrome 1 19

8.1 Occurrence o f key words i n 1 2 interview transcripts 1 59

Copyrighted Material
Preface

The basic idea for this book arose from my dealings with p ostgraduate social
science students during the mid to late 1990s. Such students, when following, say,
a taught p ostgraduate course, are commonly required to complete a dissertation in
addition to other course work.The dissertation provides an opportunity for them
to demonstrate their competence as social researchers, and to display their knowl­
edge of some substantive topic or other. In most cases candidates are required to
describe and openly reflect on aspects of methodology, as well as on specific tech­
niques that they have adopted to execute their research study. That is, they are
required to describe the strengths and weaknesses of the method or methods that
they have used to approach the topic, and to j ustify their methodological stance -
often by making reference to available published research.
When faced with a demand to execute a piece of empirically based research the
preference of the research novice is usually to think in terms of speech and inter­
view, rather than of writing and documentation. Thus, data collection procedures
are very often considered in terms of what is called 'survey' research. Such a survey
is commonly assumed to involve the design of a questionnaire and the administra­
tion of the instrument in an interpersonal interview. The consequent data analysis
is usually undertaken by means of a statistical package. Should students opt for
this route they have a wide range of well-designed and thoughtful textbooks and
manuals to draw upon, and to refer to in the design and execution of their pro­
jects. Good manuals and textbooks are also readily available for those who wish to
undertake qualitative research - of various genres - and, at any one time, there are
numerous first-class texts on the market available to the novice field researcher and
interviewer.
Although many texts suggest strategies for collecting and analysing qualitative
data of various kinds, the emphasis is, as I have just hinted, more often than not on
the spoken word. For those students who wish to centre their work on the study
of documents - or, even, to take account of documents in their research work -
there are very few pronouncements on methodology available. Indeed, the scarcity
of manuals that deal with research into documents is, itself, a rather puzzling phe­
nomenon. Perhaps, it has something to do with the fact that qualitative work, espe­
cially w ithin the anthropological tradition, was developed in the course of
examining life in non-literate societies - societies in which documents seemingly
played a minor role. Perhaps, however, it is also to do with the fact that, as the mod­
ern French philosopher Jacques Derrida has persistently argued, in the metaphysics
of the wester n world, speech has always been privileged over writing. What is writ­
ten is therefore always to be recognized as secondary, marginal and subsidiary. Yet,
despite such antipathy it is clearly the case that writing plays a maj or role in the

Copyrighted Material
x P R E FA C E

social life of modern societies. It is all the more surprising therefore that social
scientific texts outlining systematic and rigorous methods for dealing with the
written word are, more or less, absent. In fact, it seems fair to say that the world of
writing as a subject of study has been surrendered to the realm of the literary the­
orist. T his is not, of course, the place to speculate on the reasons for such bias. My
aim is merely to alter - at least in small measure - the apparent imbalance.
As anyone who has used documents in social scientific research will know, their
study demands of researchers that they adopt a variety of strategies in both the
planning and the execution of the research. Indeed, almost any study of documen­
tation will serve to contradict the notion that there is a hard and fast line to be drawn
between qualitative and quantitative research. Consequently, rigorous research work
on documents demands a passing knowledge, at least, of a wide variety of social
scientific strategies. Interestingly, many students tend to shy away from research into
documents on the grounds that whilst they might have a good topic for research,
they will not have a suitable 'method' to refer to in the appropriate section of their
dissertation. I hope that this book will go some considerable way to disabusing stu­
dents of that notion. For, as I shall show, a document, and especially a document i n
use, can b e considered a s a site o r field o f research in itself. T he investigation o f that
field requires the adoption of both appropriate research techniques and a suitable
methodological stance, and I hope to outline what those are in the chapters that
follow.
With both postgraduate and undergraduate students in mind, I have attempted
to do a number of things. First, to alert them to the wide range of possibilities
which exist in relation to conducting social scientific research involving docu­
ments. Secondly, to outline the various kinds of strategies and debates that need to
be considered whenever it is intended to integrate the study of documents into the
research programme. Thirdly and finally, to outline a distinctive (non-humanist)
position in terms of which documents may be approached and analysed, and·
thereby provide a basis for that elusive 'methodological stance' that is said to be
wanting in the study of documents. Throughout the book my focus will be on the
analysis of documents in use. Because of my own particular background in the
sociology of health and illness, many (though not all) of the examples that I have
elected to examine and analyse are drawn from that specific area of interest.
I would sincerely hope, however, that readers from across the social sciences will
be able to see - without too much effort - the applicability of my analyses to docu­
ments relating to any and all fields of interest. In every one of the above respects,
of course, a book of this kind can do little more than provide an itinerary of
possibilities. Only the researcher can cover the routes in detail.
Before moving on to matters of substance, I would like to record some debts
that I have incurred in the writing of this book. As I have already suggested, post­
graduate students have been to the forefront in offering ideas and insights. Two
such students in particular deserve mention. The first is Jon Banks, whose work on
chronic fatigue syndrome set me to thinking about the importance of documents
in clinical work. The second is Jon Brassey, for allowing me to call upon his net­
work data - as used in Chapter 9 - and for forcing me to think about some of the

Copyrighted Material
 P R E FA C E xi

concepts that are commonly used in what is called actor - network theory.
Colleagues that I have worked with in the medical school have also acted (usually
unwittingly) as sources of inspiration. Of these I must record my debt to Jonathan
Gray in genetics, Shoumi Deb in neuropsychiatry and Adrian Edwards and Roisin
Pill in primary care. I am especially grateful to Dr Edwards for permission to repro­
duce Figure 8.2. My thanks are also due to Fiona Wood for gathering much of the
data referred to in the section on risk and genetics in Chapter 4, as well as for read­
ing the manuscript. Paul Morris provided the impetus for some of the work
reported on in Chapter 8. In addition, a good deal of the research that I allude to
throughout the book as 'mine' was only made possible by grant aid from a number
of sources. I would list among the latter The Joseph Rowntree Foundation, for some
of the work referred to in Chapters 3 and 5; The National Assembly for Wales, Office
of Research and Development (WORD), for some of the work referred to in Chapters
4, 5, 8 and 9 (Grant references: R98/1/023; ROO/1/050; SG98/134); and The
Economic and Social Research Council (grant reference L218252046), for underpin­
ning work reported on in Chapter 4. I should add that in all cases, the finance was
for purposes other than writing a book, but the writing was nevertheless depen­
dent on the doing. My thanks are due also to the publishers and to the series edi­
tor, David Silverman, for waiting patiently for the long-overdue manuscript. Finally,
I should note that permission to reproduce Figure 1. 1 was kindly given by the
Nuffield Foundation.

Lindsay Prior
Kilbride, Co. Antrim

Copyrighted Material
Copyrighted Material
1
Basic Themes: Use, Prod uction and Content

What is a document? 1
Diversity in documentation 5
Documents: production and function 10
Writers and readers -: a dynamic relationship 16
Documents and their content 20
Conclusions and key points 26
Notes 27

What Is a document?

In 1917 Marcel Duchamp (1887-1968) adorned a common urinal bowl with

the signature of 'R.Mutt', placed it on a pedestal, and presented it as a work of
art to an exhibition arranged by the New York Society of Independent Artists.
The selection committee rejected it as being 'by no definition a work of art'
(de Duve, 1993) and consequently Duchamp resigned from the Society. The
'Fountain', as the artist had entitled the urinal, was just one of many 'ready­
mades' that he presented to the world in the form of objets d'art. The latter
included bank notes - signed by Duchamp himself - and a copy of the Mona
Lisa defaced with a goatee beard and a moustache (a caricature much imitated
in later years). The reasoning behind such apparently ridiculous acts and pre­
sentations was to challenge the very notion of an independent and singular
work of art. For where, Duchamp asked, was the boundary between art and
non-art? Indeed, was it possible for anyone to produce works that were not
works of art?
Although such questions may seem rather distant from the ones that concern
us, they, in fact, go right to the heart of our problem. For it is no easier to
specify what a document is than it is to specify, in abstraction, what is and what
is not a work of art. Nevertheless, by dwelling on Duchamp's questions we can
gain insight into some of the essentially social processes that are involved in acts
of definition. Consider the following.
In 1995, the Royal Academy (London) presented an exhibition of works under
the title 'Africa: The Art of a Continent'. Among other things, the display

Copyrighted Material
2 USING D O C U M ENTS I N S O C I A L R E S EARCH

included a collection of gold weights from West African societies. Such gold
weights took the form of animals and humans in various poses, and as anyone
may see from the exhibition catalogue (Phillips, 1995), they are objects of con­
siderable beauty. Yet, when the objects were originally made they were fashi­
oned with rather immediate practical purposes in mind - the weighing of gold.
This is not to exclude the possibility that they were also designed with a view
to being aesthetically pleasing, but they were certainly not made by 'artists', nor
were they made by people who would have considered themselves in any way
as 'Africans'. Nevertheless, they were presented to us, in the late twentieth
century, as examples of art made in somewhere called Africa during earlier
centuries. So what is it that has turned gold weights into African art?
Any answer to such a question must surely lie in the web of activities that
surround the objects rather than the things in themselves. That is to say, with
the actions of museum curators, critics, and cataloguers who regard the objects
as fit for display in an 'art gallery', and the viewers and visitors who are willing
to pay to see such tools as very fine vehicles for the expression of human aes­
thetic sensibilities. This, not to mention the existence of the art gallery itself,
which offers a platform or 'frame' for the exhibition of such works. The objects
as such cannot contain the answer for they are here defined as weights and
there defined as art. Nor would appeal to the original intentions of the creators
of such objects settle the matter, for their involvement with things was (neces­
sarily) ephemeral. Indeed, whichever way you look at it, the 'artness' of the art
ain't in the things.
In attempting to define the nature of a document one is, of course, presented
with very similar problems to those posed by the attempt to define art. Thus,
paintings, tapesteries, monuments, diaries, shopping lists, stage plays, adverts, rail
tickets, film, photographs, videos, engineering drawings, the content of human
tissue archives and World Wide Web (WWW ) pages can all stand as documents
in one frame or another.Yet, as with the gold weights, their status as documents
depends not so much on features intrinsic to their existence, nor on the inten­
tions of their makers, but on factors and processes that lay beyond their bound­
aries. Indeed, we shall note throughout this book that if we are to get to grips
with the nature of documents then we have to move away from a considera­
tion of them as stable, static and pre-defined artefacts. Instead we must consider
them in terms of fields, frames and networks of action. In fact, the status of
things as 'documents' depends precisely on the ways in which such objects are
integrated into fields of action, and documents can only be defined in terms of
such fields.
Fields or networks of action, of course, engage and involve creators (agents,
writers, publishers, publicists and so on), users (readers, or receivers) and set­
tings. All three realms are implicated in the emergence of documentation. As
for the producers and users, they invariably operate on the documents in terms
of specific projects and systems of relevance (Schutz, 1962) - say, the study of
fine art, or the study of archaeology, or the study of African history. Indeed,

Copyrighted Material
 BASIC THEMES 3

borrowing Schutzian terms, we might say that the social world is made up of
the 'multiple realities' of its creators. And the objects that interest us are
inevitably 'situated' in terms of such systems of reality. That is partly (but only
partly) why one and the same physical artefact (gold weights) can appear in
different guises (as African art here and as functional implements there).
For Schutz and other humanistic social scientists (such as, say, G.H. Mead,
1934) the most obvious point to enter into the study of fields of action is, of
course, through the world of human agents. In fact, for most of anthropology
and sociology 'the field' is commonly defined so as to focus specifically on the
array of activities that human actors engage in - making gold weights, art,
promises, families or whatever.Yet we should remain alert to the fact that there
is far more in heaven and earth than human agency. Indeed, human agents only
ever appear as one component of a field, for it is quite clear that human beings
necessarily live and act and work in a field of things as well as of people. And
there is forever a dynamic to 'the field' in such a way that things, such as docu­
men�s and the information that they contain, can influence and structure
human agents every bit as effectively as the agents influence the things. In that
respect, there is always a ghost of the sorcerer's apprentice present in the exis­
tence of documents and other artefacts.! In this book, that dynamic will be
central, and in many of the examples provided herein we shall be looking at
and analysing the different ways in which documents function in action.
The emphasis that social scientists commonly place on human actors mani­
fests itself most clearly in the attention that they give to what such actors say
and think and believe and opine. And should we wish to study human actors
in a rigorous social scientific manner there are many manuals available to
instruct us as to how we should proceed with our research. Most of these texts
focus on ways to capture and analyse speech and thought and behaviour.
However, few social science research manuals concentrate on the written word
and, more specifically, on documents that contain words. Indeed, when docu­
ments are put forward for consideration they are usually approached in terms
of their content rather than their status as 'things'. That is, the focus is usually
on the language contained in the document as a medium of thought and
action. Yet it is quite clear that each and every document stands in a dual rela­
tion to fields of action. First, it enters the field as a receptacle (of instructions,
commands, wishes, reports, etc.). Secondly, it enters the field as an agent in its
own right. And as an agent a document is open to manipulation by others: as
an ally, as a resource for further action, as an enemy to be destroyed, or sup­
pressed. CWe should not forget that people burn and ban texts as well as read
them.) It is the examination of this dual role that forms the intellectual back­
bone of the current volume.
As I have just stated, in so far as documents have been dealt with as a
resource for the social scientific researcher they have hitherto been considered
almost exclusively as containers of content. Now, as we shall see, document
content is important. We should not, however, let the presence of content

Copyrighted Material
4 U S I N G D O C U M E NTS I N S O C I A L R E S EARCH

bedazzle us to the exclusion of other qualities. For, above all, a document is a

product. It is a work - often an expression of a technology. And, in the ordi­
nary way of things, products are produced - they are produced by humankind
in socially organized circumstances. Consequently, one set of questions that
may quite justifiably be asked by the social scientific researcher concern the
processes and circumstances in terms of which document 'X' has been manu­
factured. It is a theme that recurs throughout the book, and one that is specifi­
cally addressed in Chapter 2.
Naturally, documents are not just manufactured, they are consumed. Further,
as with all tools, they are manipulated in organized settings for many different
ends, and they also function in different ways - irrespective of human mani­
pulations. In short, documents have effects. So a further route of analysis for the
researcher is to ask questions about how documents function in specific cir­
cumstances. Questions of functioning will be dealt with mainly in Chapters 3-5.
Naturally, the way in which a document functions is often affected by its
content, but content is not always determinant. (Indeed, at the risk of jumping
the starting gun, I should point out that the content of a document is never
fixed and static, not least because documents have always to be read, and read­
ing implies that the content of a document will be situated rather than fixed.)
In any event, the analysis of content (dealt with in Chapters 6-8), production
and use form three of the corner points around which this book is built.
Our focus, then, will be on the study of documents in their social setting -
more specifically on how documents are manufactured and how they function
rather than simply on what they contain.2 Of the three dimensions, however,
the most fundamental is undoubtedly that which relates to matters of function
or use. For, by asking how things function we can move away from a strategy
that views documents solely as resources to be scoured for evidence and data,
and into the high plains of social scientific research.3 In most social scientific
work, of course, documents are placed at the margins of consideration. They
are viewed as mere props for the real (human) action that takes place in and
through talk and behaviour. Once one has read Derrida (1976), of course, one
sees that such a position is entirely consistent with a long and inbred tradition
within western philosophical thought. That tradition, according to Derrida, has
persistently valued speech over writing and relegated the latter to a marginal
and subsidiary role.4 Writing thereby appears as an alienating force in the
world. Yet, we are aware that the modern world is made through writing and
documentation, a point that was emphasized, above all, in Max Weber's per­
ceptive analysis of 'bureaucracy'. 5 It is somewhat telling that the lesson has
seemingly been forgotten. Indeed, given the role and significance that written
documentation plays in most human societies it is strange to note just how
little attention has been paid to it by social researchers. In this book, of course,
we can hopefully move documents 'up front', and seek to mark them out as a
legitimate field of social scientific enquiry. in their own right. How that is to
be done will be evident in the remainder of the text. For now we shall merely

Copyrighted Material
 BASIC THEMES 5

concentrate on a few preliminaries: first, to consider how documents as a category

of 'things' can encompass a wide variety of media; secondly, to assess the
significance of text in documentation.6 Both of these tasks can be achieved by
consideration of a few early examples.

Diversity in documentation

We normally think of documents in somewhat uni-dimensional forms - as

fixed and static texts. And it is certainly the case that in this book, text will
figure prominently. Yet, documents are not coterminous with text. Indeed, it
is clear that contemporary documents often express their contents - ideas,
arguments, narratives or whatever - in multi-modal forms. That is to say
documents frequently contain pictures, diagrams, emblems and the like, as
well as words. W hat is more, an electronic document can also add sound to
. the multiple dimensions in terms of which it may ordinarily be composed.
In some ways, therefore, it is quite artificial to restrict analysis of documents
to text (see, for example, Bauer and Gaskell, 2000). Nevertheless, for ease of
analysis, it often makes sense to focus on documents in which written words
serve as the mastercode (Kress, et al., 1997), and that is what we shall do. One
implication of this book, then, is that the analysis of the scribbled word can
serve as a paradigm for the analysis of documents in their entirety. In order
to emphasize the significance of non-textual features of documents, and to
illustrate something of their potential complexity, however, it will prove useful
to consider just a few examples of things that do not necessarily involve text,
but which may, nevertheless, be used as documents. We will begin with
a consideration of sculpture and paintings. I shall offer here an example of
how such objects can function as documents in fields of activity, and suggest
some ways in which they may be approached as a field for social scientific
enqUIry.
At the risk of seeming overly gloomy at such an early stage of the book,
I would like to consider something about the representation of human attitudes
towards death in the western world during the last couple of centuries - in
particular with respect to sculpture. I am not so much thinking of the grand
and rather eloquent sculptures that may be found in palaces, abbeys and cathe­
drals - though the tombs of the grand and powerful have much to tell. Instead
I am thinking of everyday, routine, funerary sculpture as might be found in
almost any of the older cemeteries in the English-speaking world. Such sculp­
ture may, of course, vary from work that is nothing more than rough shaped
stone, to rather elaborate representations of angels, cherubs and the like.
Outside of the aforementioned cathedrals and abbeys, however, it is unlikely
that one could find any form of , modern' funerary sculpture dating before the
mid point of the seventeenth century. It is an absence that can be used to reveal
a great deal about western attitudes to death.

Copyrighted Material
6 USING DOCUMENTS IN S O C I A L R E S EARCH

One might be tempted to think that the absence of personal gravestones

dating from before the latter half of the seventeenth century is something to
do with the ravages of time. Perhaps all of the older graveyards have been
destroyed, lost or overgrown. It is not so. Very simply (and as far as the western
world is concerned) it is only in the seventeenth century that resting places of
ordinary men, women and children were individualized and marked. That is to
say, before this period the lot of the common person at death was to be dis­
patched to the common funeral pit, whereas after the seventeenth century and
in the world of Protestant Christianity at least, it came to be regarded as more
appropriate for 'decent' people to rest in their own private (family) plot. The
cemetery nearest to my own home, for example, has its earliest stones and slates
dated to the 1 720s, even though the area has been inhabited for some 4,000
years or more. (I should add that it is an area that fell beyond the borders of the
western Roman Empire - an empire in which personal funerary monuments
were extensively used.) In that respect, we have to suspect that the very pres­
ence of a cemetery with personalized graves ' documents' a long-term change in
social sensibility and social behaviour in relation to death. I leave aside, of
course, the fact that large numbers of poor and institutionalized people, as well
as stillborn infants, were buried in common plots right into the twentieth century,
not to mention the fact that by the first half of that century cremation, as well
as burial, came to be regarded as a legitimate form of disposal.
Further, were we to look in more detail at our traditional cemetery we
would undoubtedly note other changes. Thus, the sentiments and images on
the gravestones would, for example, appear to have altered from one century
to another. Skulls, crossbones and hourglasses would probably be present on the
earlier seventeenth- and eighteenth-century stones, whilst urns, truncated
pillars and angels on a human scale would appear on some of the later,
nineteenth-century, graves. Equally, we might note that the seventeenth- and
eighteenth-century inscriptions gave a quite heavy emphasis to the body, as in
the frequent use of the inscription 'here lies the body of XYZ'. This would
stand in contrast to nineteenth- and twentieth-century sentiments that are
more likely to ignore, and therefore to elide, the body/soul distinction that
earlier peoples took to be central. Again, as we move through the twentieth­
century cemetery we would probably see a far greater uniformity of expres­
sion on gravestones - often noting standardized heights, spacing and sizes of
the stones, fewer elaborate figures, and a marked secularization of expressed
feeling and sentiment. In such ways we might consider the shape and size of
cemeteries, their internal arrangements, the sentiments expressed on the
stones, and the very stones themselves as documenting changes in both western
attitudes towards death and features of common social relations, the growing
individualization of the dead being, perhaps, the most powerful example of the
latter. In fact, by examining the spatial organization of the cemetery we can see
how the document acts back on its creators. It achieves this in part by emphasiz­
ing the significance of individuals and intimate (as opposed to, say, organizational)

Copyrighted Material
 BASIC THEMES 7

relationships, and by emphasizing the significance of a durable personal marker

in the face of a transient life (Bauman, 1992). Indeed, it provides one means by
which we modern humans attempt to colonize the future.
The use of a cemetery as a document to changing human sensibilities was
most powerfully exercised by the French social historian of death, Phillipe
Aries, in his broad review of the Western tradition - The Hour oj Our Death
(1981). More limited and restricted overviews of the cemetery as a social doc­
ument have been produced by others, such as for example Prior (1989) and
Stannard (1977). Archaeologists, of course, use ancient cemeteries as docu­
ments in a routine manner and their work, more than any, gives credence to
the observation that the arrangement of things - spaces, stones, boundaries -
can serve as documents every bit as much as the inscription of text.
In a similar manner we can also use other (non-textual) forms of documen­
tation to plot changing sentiments and behaviours of all kinds, including those
at the moment of death. I shall cite only one example. In the Musee des Beaux
Arts in Bayonne (south-west France), there is a painting by J.-E. Fragonnard
(1732-1806) of the final moments of the Duc de Berry. The Duke is in his
large and spacious bed - as one might expect - attended by a priest, but
surrounded with a room full of people behaving rather as if the Duke's death
were something of a spectacle. I cannot pretend to convey every detail of the
artist's eye to you as a reader. However, it is clear that whatever was going on
in the Duke's chamber, the assembled collection of worthies suggests to us the
fact that, during the late eighteenth century, death was not to be considered a
lonely and isolated business. Instead, it appears as a social and religious event of
considerable importance - perhaps, even, a communal event. And this theme
of a public death is represented many times over in many different paintings of
both lesser and greater figures during the seventeenth, eighteenth and nineteenth
centuries. (In his Images oj Man and Death (1985), Aries provides a selection of
similar images relating to European and American people of all social stations.)
Only in the twentieth century does death truly appear as a private and isolated
affair, and only then are dying people secreted away in hospital wards and
hospices to die in seclusion. This move from the public to the private sphere -
a process sometimes referred to as the sequestration of death (Seale, 1998;
Giddens, 1991) - is, of course, only one of many changes that is represented in
European and American painting. It is also only one of thousands of ways in
which paintings and visual images of all kinds might be called upon as docu­
mentary evidence of fundamental alterations in our sensibilities to that infinite
array of human activity that we call 'social' . In this respect we might say that
paintings, gravestones, cemeteries, wills and endless other items can often be
used to reveal something of the different ways in terms of which death and
dying have been 'performed'.7
Naturally, it is possible to study representations of death and dying by using
written records alone. That is, by the use of documents in the narrow sense of
the term. Thus Richard Cobb (1978), for example, trawled through the dossiers

Copyrighted Material
8 USING DOC U M ENTS I N S O C I A L R E S EARCH

of those who died violent deaths in Thermidorian (revolutionary) Paris by

examining in considerable detail the records of the Basse-Geole de la Seine.
The results of his work throw light on the minute detail of the everyday lives
of ordinary Parisians (such as features of dress and occupation), as well as on
the detail of the nature of suicide and homicide in Paris between 1795 and
1801. In fact, as Cobb's interest in the wardrobes of his suicides suggests, we
should not become beguiled by text. People think with things as well as words,
and, very often, the arrangement of things is as significant as the arrangement
of words. Human activity finds expression through many media and the shape
of things can be regarded as every bit as important as the flow of text. Just to
emphasize that point, I shall end this section by looking at how architectural
plans (which report on the arrangement of spaces) can be appropriated as
documents in a social scientific frame.
By providing visible statements about the ways in which constructed spaces
are to be divided and sub-divided, architectural plans inevitably express ele­
ments of the discourses in terms of which they are conceived. A society that
lived in circular domestic spaces, for example, is more than likely to have
different ideas about the 'place' of things and people than a society that lives in
square or rectangular spaces (Bourdieu, 1977). And this notion that a study of
the manner in which we organize space can reveal basic rules about how daily
living is to be managed and arranged can be carried over to the study of more
detailed areas of social life. The arrangements of desks in a classroom, the
arrangement of classrooms in a school, the positioning of a school, or prison,
or psychiatric hospital, or a factory can all be used to reveal evidence about
how a society orders social relationships. In some previous work I described
how the changing nature of hospital plans can be used to reveal changes in
such things as our attitudes and behaviour to mental illness, infectious diseases
and infancy (Prior, 1988; 1992). It would be inappropriate for me to review
the details of the arguments here. Instead I shall offer you just one plan of a
hospital ward (Figure 1.1). It is a plan for a children's ward area 1963. As I have
stated elsewhere (Prior, 1992), the very development of hospital wards for
children serves to mark out, in large part, the image of childhood as a separate
stage of development in western society. The plan in Figure 1.1, however, dis­
plays interesting developments over and above those that are apparent in hospi­
tal plans that emerged between 1880 and the 1950s. In particular, we may note
how rooms have been designed· to incorporate mothers as well as children - in
homage to the significance of the maternal bond that had been much discussed
in post-war ' child psychology' . The open area for play and lessons is also remark­
able. In terms of play, it is notable for the way in which such activity has been
moved in an open space - to facilitate good observation by the nursing staff. In
terms of lessons it is notable as early recognition of the sick child's 'need' for a
learning environment. The generally open design of the ward is also intentional,
and reflects increasing professional concern with surveillance of children in
hospital. (The apotheosis of this concern appears in the late twentieth-century

Copyrighted Material
 French windows

0 0
to garden

MM

oo 00 0 00
DIDo
Space for play, lessons

0 0
and meals
c::::::J c::::::J c::::::J c::::::J
0

0
Divan for
Mother

c::::::J c::::::J c::::::J c::::::J

� 7-Bed Bay
5-Bed Bay
OJ
l>
U>
�
Do Do
0
t(5.
� c::::::J c::::::J ()

CD �
0.. 101 Clothes Cupboard 101 I
m
� Sink Sink
�
m
CD Nurses' Station
U>
I I
� o 0 Linen Store
o
Clean-Utility
Room I I Treatment
Sister's
Office Dirty-Utility
Room
Room
Kitchen B
o

FIGURE 1 . 1 Spatial organization as documentation (Nuffield Foundation, 1963)

(l/
10 USING DOCUM ENTS IN SOCIAL R E S EARCH

use of two-way mirrors through which parents could be secretly observed inter­
acting with their offspring.) Beyond that the general design of the ward reflects
a whole series of ideas about the nature of childhood health and illness and the
role of social contact in sickness.
Architectural plans and paintings, then, offer us clear examples of documents
in which text plays only a minor role, whilst the cemetery offers us a clear
example of a document that represents and reflects through a mixed array of
physical and cultural dimensions. Their status as documents (rather than, say, as
art) is, of course, dependent on the frame in terms of which they are 'situated' ,
and henceforth appropriated. In our case, that frame has been a social scientllc
one. As we shall note shortly, social scientific documents can themselves be
appropriated, in turn, by other actors and agents and metamorphosed into legal,
political or religious documents, or they may be used for the establishment of
personal identity or to create new facts and new things. Before we consider such
issues, however, it would be as well for us to look at how documents come into
the world in the first place. That is, to examine how they are created.

Documents: production and function

In the Gospel of Saint John (I, i) we encounter a truly majestic claim, namely
that, 'In the beginning was the word.' The German author Goethe was later to
play with this phrase in the first part of his Faust. Goethe's Faust is one of those
figures that has little compunction in forming a contract with the Devil in
order to achieve his worldly ends, and so it is not perhaps surprising that he
inverts the Gospel claim so as to assert that, 'In the beginning was the deed.'
(See von Goethe, 1986.)
The two sentences point to a rather beguiling opposition between word and
deed that, fortunately, we need only approach in a social scientific manner. In
that light it seems safe to assert that documents are created in the context of
socially organized projects in such a way that word and deed belong together.
One plausible line of social research that remains open to all investigators,
therefore, is to follow a document through its social trajectory - to examine
how it is manufactured or produced in specific contexts of thought and deed.
Naturally, the production of documents is a complex business, and, in practice,
it is often bound up with processes of consumption - recall that the afore­
mentioned gold weights needed museum visitors and curators as well as crafts­
men to turn them into art. For the sake of clarity, however, we will maintain,
in this chapter, an analytical division between the production and the con­
sumption of documents, always keeping in mind that it is the active, dynamic
assimilation of 'things' into fields of action that is the key to understanding the
process of fabrication.
Possibly, the first point that needs to be made is that the birth and life of
documents rests on the foundations of a collective rather than individual

Copyrighted Material
 BASIC THEMES :U

action. In western societies, of course, we are usually keen to think in terms of

individual authors as the makers of documents. And individual authors are,
indeed, important figures, but as I will illustrate with the following examples,
the author as subject is only one side of a many-sided phenomenon. Indeed,
following Foucault ( 1 979), we are far better off thinking in terms of an 'author­
function', rather than an author as subject. That is to say, to consider the con­
cept of an author as a tool that can be recruited so as to bypass a consideration
of complex social referents. (Indeed in modern scientific and medical journals
the concept of authorship is being forsaken in favour of the concept of'con­
tributor' - see, for example, Smith, 1 997a; 1 997b). To illustrate the broader
issues, however, let us consider some famous texts in this light. I will begin with
a look at Homer's The Odyssey.
The Odyssey gets an early mention in Aristotle's Poetics (fourth century BeE) .
Aristotle summed up the epic poem of about 12,000 lines in three or four
sentences.

A certain man has been abroad for many years; he is alone and the god Poseidon keeps a
hostile eye on him. At home the situation is that suitors for his wife's hand are draining
his resources and plotting to kill his son. Then, after suffering storm and shipwreck, he
comes home, makes himself known, attacks the suitors: he survives and they are destroyed.
(Aristotle, 1 995: 1 7)

The Odyssey, then, is a narrative. The poem itself is estimated to be about 2,700
years old. Whatever its age, it is a masterful piece of literature. As a printed text,
however, it can be traced back only to 1 488. Before that date it must have
existed merely as a handwritten document - presumably contained, at one
stage, on papyrus rolls. What is interesting from our (social scientific) viewpoint
is that the poem as it presently exists is still attributed to a single poet called
Homer - who, it is claimed, wrote it. This is somewhat odd, because, for
various reasons, it is clear that the Odyssey epic was created in an oral rather
than a literary tradition. For example, not one person in the story is described
as being able to read or write. Moreover, in the peasant society of the day,
poetry was something more likely to be 'performed' than written (Knox,
1 996). Yet, whether it was written or spoken, it certainly seems plausible to
argue that the poem was both created and embellished over the 27 centuries
by a large number of people - speakers/performers, listeners, scribes, editors,
publishers, translators. For example, whoever the originator of the poem may
have been, he, she or they produced the narrative and the plot of the Odyssey
with an audience in mind - a plot that only resolves itself in the final moments
of the text. Indeed, the narrative would have been consciously designed to
resonate with audience expectations and experience, and as such one might say
that the audience would have had a role in authorship. Thus, we can assume
that over 2,000 years performers of the epic would have found bits of the
narrative that 'worked' better than others, and that too would have impacted
on the poem. In modern times, translators of the epic will have imposed their

Copyrighted Material
12 USING DOCU M E NTS IN SOCIAL R E S EARCH

rhetorical style on the poem, and so forth. In other words, it is not unreasonable
to suggest that it is a collective rather than an individual genius that is involved
in the poem's rather lengthy production. (Indeed, the eighteenth-century
Neapolitan philosopher Giambattista Vico considered the works of Homer to
be a product of the Greek people rather than of any individual within the
culture.) Naturally, such collective acts have, at different times, been guided by
various kinds of rules - concerning the structure of the text, the nature of read­
ing, the art of translation and so on (see Knox, 1 990) . For example, there were
clearly rules that determined the style and flow of the ancient Greek compo­
sition, one of which involved the use of the hexameter (six metrical units) .
There were also rules governing the various epithets that could be attached to
people and things (the dawn of Odysseus is always 'rosy fingered', for example,
and clouds always 'scud' across the sky) . Above and beyond that, the poem
expresses rules - about such things as honour, fidelity, justice, and the relation­
ships between gods and mortals - and in that sense reveals to us significant
aspects of everyday life in the ancient world (Finley, 1 977) .
What is of interest here is that the impact of these clear traces of collective
action are routinely effaced and then subsumed under the author-function -
so that the 'author' functions as the creator. Naturally, it is not for us to decide
on the historical issues concerning The Odyssey one way or another. Our inter­
est in the example is almost entirely in the fact that western literary tradition
rests on the illusion of an identifiable author as a unifying force for textual
materials. And we seem to hold to the belief even when we are aware that the
existence of a text depends on broad social and organizational processes of pro­
duction, rather than simple acts of personal genius and inspiration. In this light,
it is useful to consider (though only briefly) one of the key texts of twentieth­
century linguistics - Saussure's Course on General Linguistics. The latter was
published after Saussure's death in 1 9 1 5 (see Saussure, 1 983), and was compiled
in large part from the lecture notes of his students. Although the text is referred
to as Saussure's and quoted and cited as such, there is a sense in which the
'author-subject' had relatively little to do with the final production. (Indeed,
the full reference for the book, provided in our bibliography, is of some interest
in this light.) And this use of the author-subject (that is, a named person) to
endow a sense of unity and order to a document is common to most (though
not all) forms of written and published discourse.s In fact it is of interest to
note which kinds of documents are normally expected to be associated with
identifiable author-subjects (books, plays, poems, scientific papers) , and which
do not (acts of legislation, committee reports, theatre tickets, tax forms, mod­
ern maps, train tickets and so on) . In a parallel manner it is of interest to note
which kinds of documents require signatures and which do not. (It is also
instructive to ask what a signature supposedly adds to a document - a rather
interesting question taken up by Derrida ( 1 977) and some of his critics.)
What our ancient Greek Odyssey tells us, then, is that documents are essen­
tially social products. They are constructed in accordance with rules, they

Copyrighted Material
 BASIC THEMES 13

express a structure, they are nestled within a specific discourse,9 and their presence
in the world depends on collective, organized, action. The same lessons are
evident in the study of almost any text, but let us take yet another notable
document, merely to emphasize the issues.
The Bible is a text that reports on the word of God, and without questioning
the truth of that claim in any way, we can still recognize it as a text that has
been mediated in different ways through human intervention and production.
For example, in the contemporary world there are many different English
translations of The Bible. There is a so-called authorized or King James Bible
( 1 6 1 1 ) , an American Revised Standard Version (1 952) , a Modern English
Version (1 966), a New International Version ( 1 972) , and many others. Each
text differs, one from the other, subtly but noticeably in the words and
emphases that it contains, as well as in the order of the various books and
verses. Differences of order and content are perhaps not surprising since all
versions are dependent on translations of Hebrew and Greek texts - texts and
fragments of text that were later translated into Latin and used as the Vulgate
or common version of the Bible. English language versions, of course, emerge
out of a specific Protestant tradition. And among the earliest and most beautiful
of English language Bibles was that produced by Tyndale during the 1 530s. It
was his text that formed the basis for the later King James ( 1 6 1 1 ) Version. And
it is from Tyndale's work that we have borrowed many of the best-known
English phrases - 'the salt of the earth', 'the powers that be', 'eat, drink and be
merry', and 'a law unto themselves' are all Tyndale's phrases (Daniell, 1 989). The
acceptability and convenience of such phrases is of course a direct product of
the translator's art. They are also phrases to which few, if any, would take excep­
tion, but taken in its entirety, Tyndale's Bible was objected to - so much so that
he was strangled and burned to death for having produced it.
Clearly, were we to become interested in the fine detail of biblical language,
and concerned with nuances of meaning and matters of emphases, we would
need to take the history of the relevant production and translation processes
into account. In particular, we would need to examine the various twists and
turns by means of which a biblical text and the manner of its division into
chapters and verse- came into our hands. In so doing we would come to rec­
ognize the fact that documents not only are produced in accordance with rule­
governed procedures, but always exist as resources in schemes of action. They
both express and represent a set of discursive practices. As such they can be
recruited as allies in various forms of social, political and cultural struggle.
Indeed, we know that people frequently mobilize and use the detail within
documents for social, political and, as in our example, religious purposes and
projects. In fact, Tyndale was murdered presumably because his translation
expressed and represented a discourse of Protestant individualism at a time
when the established authorities felt it necessary to assert the power of the one
true church and its monopoly on biblical exegesis. Mter his death, opponents
of the 'one true church' recruited, mobilized and assembled both Tyndale and

Copyrighted Material
:14 USING DOC U M E NTS IN SOCIAL RESEARCH

his Bible into a project relating to what we vaguely call the Protestant
Reformation. Note that the (Tyndale) Bible, in this sense, is cast into the early
modern world as an actor in its own right and was therefore also liable to
destruction and suppression at the behest of the authorities.
This Frankenstein-like quality of documents,IO that is the capacity of
humanly created artefacts to serve as active agents and counter-agents in fields
of social action, is not to be underestimated. Indeed, even writing - as script -
can serve as an agent. Thus, in many cultures, we know that specific forms of
script have been monopolized by a socio-political elite so as to underpin
particular bases of power and exploitation, and to exclude outsiders such as
peasants and foreigners from making an impact on elite cultures. Goody ( 1 968)
provides many examples of such practices, and indicates how various social
strata have (in the past) recruited script as a powerful political ally, the practices
of the Chinese literati being the most telling of these.
This last point serves to emphasize how a study of the use of documents can
be as telling as a study of content. In the next example, we can see how mat­
ters of content have often become central solely because of the use to which
the document can be put. Though just to keep our eyes on the ball, I should
emphasize that the value of the example lies in the issue of production rather
than of use.
Questions concerning the authenticity of a document often arise in the
research arena. Such questions are, of course, essential - and they often shed
direct light on important issues concerning the reliability of text as evidence.
Indeed, the few studies that have been devoted to the use of documents in
social scientific research (see, for example, Scott, 1 990) have often concentrated
on issues of authenticity to the exclusion of issues of use and function. It is per­
haps yet another reflection of the general emphasis that is given to document
content rather than use. Yet, establishing authenticity is far from being the only
task of the social scientific researcher. Equally as important are questions about
how documents are produced and recruited - and the recruitment of docu­
ments, as with the recruitment of soldiers, is not always dependent on a fitness
test. In this vein a useful example to call upon involves Anne Frank's Diary.
During the 1 980s, following a series of ill-founded and crude objections
relating to the authenticity of the diary, the latter was subjected to extensive
forensic examination by the State Forensic Science Laboratory in the
Netherlands (Barnouw and Van Der Stroom, 1 989) . The use of a forensic plat­
form enabled investigators to concentrate, in the main, on such items as the
glue, the ink, the paper, and especially the handwriting contained within the
documents. All of these confirmed the authenticity of the time and place in
terms of which the diary was produced (Amsterdam, 1 942-4) , together with
the approximate age of the writer (mid teens) . A few other investigations
focused on the routes by means of which the diary came into public view, and
in tracing such routes we can see how Anne's diary did in fact undergo a number
of important transformations. For example, her father, Otto, originally typed

Copyrighted Material
 BASIC THEMES 15

up her handwritten documents and, in the process, exercised a degree of

editorial discretion. Indeed, he extracted only what he considered to be the 'essen­
tial' detail from her diary - omitting painful observations - and he also altered
spellings and punctuation. These difficulties were accentuated by the fact that
Anne herself wrote two versions of her ' diary' and, in addition, a further text
entitled 'Tales from the Secret Annex'. The earliest English language (1 952)
edition is a composite of all three sources. Indeed, it is possible to see that there
are many Anne Frank diaries - and it was reported during 1 998 that some new
pages of the diary had been discovered. (Note that a 'diary' - as with a 'poem'
or a 'record' - is always a 'situated' product and not something that can be
defined according to its intrinsic features.)
Details of the kind relating to the history of Anne's diary are essential
and important because, within the rhetoric of social scientific and academic
discourse in general, the establishment of a document's authenticity is, as we
pointed out above, central to its acceptability as evidence. That is why it is vital
that the conditions under which a given document has been produced are
made known and made publicly available. In this particular case these have
been made evidently transparent. Yet, we have also to ask why anyone would
question the authenticity of a teenage girl's ' diary'. Challenges to authenticity,
of course, are normally associated with highly valued or politically charged
documents. Indeed, as was pointed out above, it is the position that a document
holds in a network or web of activity that contains the key to its use and recep­
tion. Thus few people would be interested in investigating the precise history
of George and Weedon Grossmith's The Diary of a Nobody ( 1 999, first published
1 892) - an intensely amusing, fictional account of a City of London clerk's
daily life, very much in the ' got up, had breakfast, went to work', genre. This is
essentially because it is a diary without social and political consequence
(though its supposed author, Mr Pooter, would be devastated to read that
judgement) . Anne Frank's Diary, on the other hand, serves as both a witness to
and an advert for the wide variability that exists in the application of human
moral codes. As an advert for the darker side of such morality, it can and does
play a powerful and influential role - certainly influential enough for right­
wing parties to seek to suppress it or to denigrate it.
The Bible, The Odyssey and Anne Frank's Diary each exhibit elements of
collective action. The contemporary forms of all three documents emanate
from many different sources and the analysis of the routes and actions by means
of which the documents come into the world requires close investigation. As
we shall see in Chapter 2, the study of the ways in which documents are pro­
duced can constitute a topic for social scientific study in itself. As well as illus­
trating themes such as those mentioned above, such study can also highlight
essential features of social action and interaction in general. Above all, when we
begin to look at the manner in which, say, official documents (reports, statisti­
cal tables and so on) are manufactured, we will also come to recognize the rule­
driven procedure that lies behind the production process. It is through the

Copyrighted Material
16 USING DOCUME NTS I N SOCIAL RESEARCH

routine application and interpretation of such rules that facts about society and
the world in general come to be known and made. Normally, of course, we
prefer to gloss over issues relating to the genealogy and manufacture of docu­
ments and simply use them as resources for further study and research. Yet, the
above examples serve to illustrate just how a study of the ways in which docu­
ments are produced (and how they are used or consumed) in socially organized
circumstances is every bit as important as a study of content. Naturally, in the
empirical world all three features of documentation are interlinked. Indeed,
before we move on to consider matters of content we need to consider issues
of use and function a little further, and that is the aim of the following section.

Writers and readers - a dynamic relationship

The consumption of documents is as important as their production, though in

terms of the dynamic that is created within fields of action the producer/
consumer distinction is of heuristic use only. So before we turn to consider the
consumption of documents proper, it may be as well to emphasize how those
who use and consume documents are not merely passive actors in the com­
munication process, but also active in the production process itself.
The critical role of the consumer as an active agent is, perhaps, most clearly
seen in the contemporary world where computerized documents are common.
For as we know from our own experience, a reader of a web page, or any
screen-based document, can these days easily cut, paste, edit and re-edit text to
suit the user's purpose. Yet, even though cutting and pasting has become a
visible, traceable process in the electronic world, we should be aware that every
reader in every age has cut and edited documents to suit a personal agenda. It
is just that the process was previously invisible and non-traceable. Skipping
chapters and pages is, after all, a well-established technique for getting from
page 1 to page n . Indeed, and as de Certeau (1 984: xxi) has argued, 'the act of
reading has . . . all the characteristics of a silent production' . More importantly,
perhaps, the social scientific researcher needs to be aware that a reader can enter
into the creative/productive process long before he or she opens a document.
Indeed, the reader can influence a writer at the exact moment when the text
is first being inscribed on paper. Consider the following examples.
In a not so distant world, the written letter was a frequently used form of
communication. Indeed, it was an essential means of communication both
within countries and between countries. So, in the history of the USA, for
example, the letter has come to assume a very special role, especially those
letters that European emigrants to the USA wrote to their families in the 'old
country'. In fact, and as we shall see, they have subsequently formed an impor­
tant resource for social scientific study. Naturally, and in the everyday run of
things, such letters contain details of routine and mundane events, and in that
sense provide unremarkable authors' accounts of the writers' activities, wishes,

Copyrighted Material
 BASIC THEMES 17

desires and plans.Yet, what is important from our current viewpoint is that such
letters were usually written with the interest of readers in mind. In other
words, authors had images of what the readers would want to hear about, and
what might hold their attention. So, the absent reader sitting in Europe was, in
such ways, present in the USA when the letters were first scribbled down, and
was able to fashion the letter according to his or her needs and wishes.
The exact same point can be made with respect to other kinds of document.
Take, for instance, a diary or autobiography. Now every personal document of
that kind is written with some reader or audience in mind. Some diarists, of
course, write as though an audience is of no significance. That is, as if the doc­
ument can be detached from the social fields in which it is created and lodged.
The Diary oj Samuel Pepys (see the 1 970 edition) offers one example of the
style, for Pepys was always keen to give the impression that his accounts were
produced by the day and off the cuff without thought to a reader. Yet, we have
good reason to suspect that he consciously fashioned and shaped his diary for
future public consumption - and thus for reader/audience approval.We can, of
course, only guess at how Pepys saw his receivers. In the case of John Stuart
Mill ( 1 989) , on the other hand, he tells us on the first page the kind of reader
that he has written his Autobiography for, and it is easy to see how his image of
the receiver shapes the ensuing document. These routes whereby the reader
interpolates a presence into the writer's world are, in fact, numerous, and occur
across the entire range of documentary material. Thus, in verbatim accounts of
interviews and interrogations we see how the questioner can determine the
shape of an author-subject's narrative by posing key questions - questions that
stimulate the unfurling of the narrative in specific directions. Similarly, in the
posed photograph we see a product that has been negotiated between both the
creator and the sitter such that it would be difficult to determine who, exactly,
the author-subject of the photograph might be. Such acts of interpolation
of 'readers' into the world of the author are, however, often seen at their
best when considering creative works of a musical or artistic nature. So let us
consider yet a further example.
In the world in which the composer Dmitri Shostakovich (1 906-75) oper­
ated, the function of art was seen in terms of the glorification of the Soviet
Union, the Proletariat, the 'Party' and 'socialism'. Music that glorified individ­
uals and individualism was largely frowned upon. That is, unless the individual
was someone of the status of a Comrade Stalin - a man who stood as the
living embodiment of the Proletariat, the Party and the Soviet Union. When
Shostakovich composed his music, therefore, he was obliged to take such con­
straining factors into account (Fay, 2000). Often, he dealt with the tensions that
arose between his creative capacities and the demands of the Party by writing
his music in different intellectual keys. That is to say, he would write a piece of
music that supposedly bolstered the principles of communist ideology whilst,
at the same time, interpolating himself into the composition. One example of
this is evident in his String Quartet Number 8 (Opus 1 1 0). It is a thoroughly

Copyrighted Material
18 USING DOCUMENTS IN SOCIAL RESEARCH

depressing piece of music. It was composed in 1 960 just after Shostakovich had
visited Dresden. At that stage, Dresden still exhibited the wreckage of war and
the results of the carpet bombing to which the city had been subjected by the
Royal Air Force. During the same year Shostakovich had agreed, under pres­
sure, to become a member of the Communist Party of the Soviet Union. He
dedicated Opus 1 1 0 to the 'victims of fascism and war' . This inscription was
taken by the Communist Party's apparatchiks as approval for Soviet ideology
and the Dresden connection was obviously a welcome one. Interestingly, at the
heart of the quartet is a four-note motif. It is a motif derived from Shostakovich's
own name (DSCH, or what we know as D, E flat, C, B) . It is a trick that
Shostakovich had used in some of his other works - to put himself, rather than
the Party, at the core of his own music. And there are also other devices and
references that DSCH used in Opus 1 1 0 that were somewhat ambiguous. For
example, in the last movement there is a repetition of some very aggressive
chords that sound very much like three dark beats, evidently representing
falling bombs. Much later, however, Shostakovich was to admit that he derived
the chords from the sounds made when the Soviet secret police pounded
apartment doors in the midst of night.
So what we have in Opus 1 1 0 is a piece of music made with the Party audi­
ence uppermost in mind, whilst the author, and his ideas, appear only by means
of subterfuge. In that sense we might say that the Party (and the Soviet secret
police) were as influential as Shostakovich was in creating the string quartet.
That, perhaps, is to go too far, but we can legitimately claim that audiences
(including those comprising secret police) forever interpolate their presence
into texts - if for no other reason than texts (documents) always have readers.
And this despite the fact that such acts of construction are inherently unstable
and tend to alter in line with changing hierarchies of relevance and socio­
political context.
Naturally, the most important reader of a musical (or of a theatrical) text is
the performer him- or herself. And it is in the gap that exists between the writ­
ten notation on the page and performance of music on the stage that allows us
to sit through the 'same' concert (or stage play) many times without any sense
of weary repetition. Performance reinvents the notation (text) . This is as true
of a Mozart opera or a Beethoven symphony as it is of, say, a rendition of Frank
Sinatra's 'My Way' . And given such considerations, we may be tempted to ques­
tion the very primacy of authors over readers/performers. This is just what the
literary critic Roland Barthes ( 1 9 1 5-80) did - though mainly with respect to
issues of meaning and interpretation. Thus in an article entitled 'Death of the
author' ( 1 977) , Barthes argued that a text's unity lies not in its origin, 'but in
its destination', that is to say, with the reader. So only the reader can provide a
sense of unity to a text and it is, consequently, on readers rather than writers
that we ought to concentrate. (A similar argument was advanced in S/Z
(Barthes, 1 990). We need only quarrel with Barthes to the extent that he con­
siders 'reading' as an inner, subjective and personal act, whereas in this book we

Copyrighted Material
 BASI C THEMES 19

shall consider reading as performance. The task of the social researcher is therefore
to study how readers use and consume text and notation in everyday life
(routine performance) . To this end, let us consider a final example.
During the closing years of the nineteenth century, when ethnographic
research was in the early years of development, a number of English anthro­
pologists studied family trees in the Torres Straits society of the southern
Pacific. A.C. Haddon and WH.R. Rivers had travelled to the Torres Straits in
1 888, and they asserted an interest in Torres Straits society as an entirety. They
consequently examined family life, economic and political activity, artistic life,
religion and so on, as well as taking an interest in the physiology and psycho­
logy of the inhabitants (Haddon, 1 904). Among other things that they 'recorded'
were the family links of the people being studied. Such ways of recording
human relationships were quite foreign to the Torres Straits inhabitants
and, in retrospect, it seems as if the locals failed to be entirely open with
the anthropologists - or perhaps the anthropologists failed to ask the right
kinds of questions and thereby ended up with a limited narrative. The mis­
understanding arose in relation to the issue of adoption, which is relatively
common in the society in question. Thus, in Murray Island, for example,
adults can adopt children without any apparent formal declaration. People can
also have more than one name, and can often change names, seemingly at will.
Haddon was unable to grasp the significance of these points and despite the
fact that he lived on Murray Island for some time, he clearly misunderstood
the nature of adoption and inheritance patterns. Such misinterpretation is in
itself rather interesting not least because it calls into question the extent to
which outsiders can gain knowledge of insider procedures, even when they
adopt ethnographic methods. More appropriate to our concerns, however, is
the fact that although Haddon's genealogies were developed in an anthro­
pological project and from an academic platform of Cambridge University,
Torres Straits inhabitants currently use them for quite different purposes -
that is, to establish rights to property. In fact, Haddon's genealogies are used to
press land claims.
What the Torres Straits example illustrates, then, is that the nature of the
consumption process can alter the entire nature of the document. Thus the .
Cambridge University Reports were intended to be consumed as objective
scientific (anthropological) accounts of family and kinship in an exotic (non­
western) society. The detail of the family trees, for example, is meant to illus­
trate patterns of consanguinity and affinity - as expressed through a discourse
of scientific anthropology. Later researchers might have used such reports as
data for their theories of kinship. Yet the contemporary inhabitants are using
the very same documents in an entirely different context and presenting and
lodging them in different kinds of platforms. That is to say, they are lodging
them as evidence in courts of law concerned with the definition of property
rights. In that field of action, users (or consumers) have effectively turned the
anthropological documents into legal documents.

Copyrighted Material
20 USING DOC U MENTS IN SOCIAL RESEARCH

One further point: whatever the shortcomings of the anthropological

investigations it is at least clear that drawing up a genealogy is not simply a matter
of inscribing names on paper. It is something that has implications for social
action - action that cannot always be foreseen by the authors. In an entirely
different context, Berg (1 996) has referred to this kind of implication as 'action
at a distance' - the process whereby making inscriptions in �me setting neces­
sarily leads to something being done in another setting. Ir't this respect one
might view the act of drawing up a will , or completing a marriage certificate,
as standing in exactly the same kind of relationship to legal action as the com­
pletion of a Torres Straits genealogy by an anthropologist. More importantly,
perhaps, we can begin to see, once again, how the documents themselves
become 'actors' in the social process. That is to say, documents can enter into
systems of action in their own right, and are not just passive items operated
upon by human agents. Their very existence, in that sense, can influence the
actions of human beings. It is in this respect that one is drawn, yet again, to
consider the fable of the sorcerer's apprentice, which certainly resonates with
the Torres Straits genealogies. Other examples abound. Indeed, there are cir­
cumstances in which a document may be said to be more pertinent to action
than a person. This is evidently so in the case of documents of identification,
such as passports and birth certificates, where it is possession of the relevant
document that can trigger a series of actions - such as admittance into a
country or an organisation (whilst the corporeality of the person can count for
little in the absence of the documentary substance) . It is a point that echoes
throughout Nikolai Gogol's Dead Souls (see Gogol, 1 996) , a story constructed
around the theme of a man who bought the names of dead serfs (dead souls)
from their landowners so that the landowners would not be liable to taxation -
a taxation that was levelled on the 'books' rather than the bodies. And that is
only one set of circumstances in which one becomes aware of the role of the
document as actor. In that light it is, as Atkinson and Coffey ( 1 997) have
argued, rather puzzling as to why researchers continue to produce accounts of
complex, literate, social worlds ' as if they were entirely without writing'.

Documents and their content

So far I have highlighted aspects of the context in which documents are

produced and consumed. Both the framework of consumption and that of pro­
duction form an important area for study in the social sciences. Such a focus
on context, however, leaves us at risk of ignoring content. As I have stated
before, most of the available methods texts that deal with research into docu­
ments have focused on content almost .exclusively. In this section I intend to
outline some of the broader issues that have to be confronted by the social
scientific researcher when : ontent is integrated into a research programme.

Copyrighted Material
 BASIC THEMES 21

Content analysis can take many forms. In its simplest, empirical, sense it can
involve little more than enumerating the frequency with which certain words,
items or categories appear in a text. Some of the earliest methodological state­
ments on content analysis appear in Goode and Hatt ( 1 952) , wherein they
concentrated mainly on the analysis of political speeches and the like. During
later decades the methodology of empirical content analysis developed in more
quantitatively complex ways. A good overview of the method is provided in
Weber ( 1 990).
Naturally, enumeration of words and themes has its place, but only within a
well-considered theoretical frame. For whilst simple counting strategies can
reveal much about the focus of a document and what its dominant concerns
appear to be, they only add up to anything insightful once the function of the
document has been identified. In other words, the enumeration process must
always rest on an informed analysis concerning the nature of the 'facts' and
'categories' to be counted. Thus a document might function as the carrier of a
message, an object to be translated, an impediment to understanding, or, yet, as
a prop to interaction. We will look at a few examples in Chapters 6 and 7.
Perhaps it could suffice for now to state that if we wish to move beyond the
surface content of a document and into its functioning, then deeper and more
sophisticated strategies of analysis may be required. But let us consider the
possibilities for a relatively straightforward content analysis first.

The content of the immigrant letter

During the period 1 9 1 8-20 Florian Znaniecki and WI. Thomas published a
multi-volume work entitled The Polish Peasant in Europe and America (see
Thomas and Znaniecki, 1 958). It reported on a study of family and commu­
nity life among Polish immigrants to the USA during the very earliest part of
the twentieth century. In order to. gather data on family and community life
Thomas and Znaniecki relied very heavily on the analysis of documents. Such
documents were of various kinds and included parish records, life histories. and
letters. In retrospect, the methodological basis of the study looks rather shaky,
but the general design of The Polish Peasant has much to recommend it. (The
authors did, in fact, devote some time and effort to composing a methodologi­
cal statement on their work, but it strangely lacks detailed connection with the
substance of their study.)
The use of immigrant letters as a source of social scientific data was proba­
bly not original - even in 1 9 1 8 - but it was, nevertheless, inventive. Thomas,
in particular, was concerned with individual attitudes - towards possessions, the
family, social relationships and so forth. The immigrant letter in this respect was
seen to function as a repository of attitudes. For example, the very fact that
such letters were written at all indicated that Polish immigrants were willing
to invest a considerable amount of time and effort in maintaining family links

Copyrighted Material
22 USING DOCU M ENTS IN S O C I A L R ES EARCH

across two continents. On the other hand, the actual content of the letters
suggested to Thomas that in many key respects social solidarity was breaking
down in the Polish community. For example, the letters were said to reveal a
considerable degree of conflict about such matters as marriage partners and
social relationships. As with many researchers Thomas and Znaniecki can be
accused of finding in the data only what they wished to see, and the theme of
social disorganization was already firmly implanted in the sociology of Thomas
well before he had ever looked at any letters. It is not surprising, therefore, that
social disorganization in the American urban Polish community is what
Thomas saw the letters to reveal. As I have already hinted, Thomas and
Znaniecki were not all that clear on basic facts about where the letters were
obtained from, or how many letters were received and analysed, but their work
nevertheless gave a spur to the use of such documents in the study ofAmerican
history. Oscar Handlin's The Uprooted (1 953) , for example, may be said to have
been partly inspired by ideas drawn from The Polish Peasant.
Although The Polish Peasant exhibits a considerable degree of theoretical
complexity for its day, the use of the letters is relatively straightforward. They
are used essentially as a data source for content analysis. That is to say, certain
themes were identified - social disorganization, patterns of family interaction,
individualization, and so forth - and then the researchers attempted to assess
how frequently these themes appeared in the letters. On the face of things, that
is a perfectly reasonable line of approach. In fact, however, it encourages a con­
siderable adulteration of the data sources. For, by imposing a pre-organized
conceptual grid - derived from professional social scientific work - on a data
source, it is more than possible that the social detail of the letters themselves
was lost. For instance, one could use immigrant letters to examine the way in
which, say, 'self' has been conceived among different groups and at different
times (see Barton and Hall, 2000). Letters written with references to 'us' and
'we' for example, signify a different orientation to the self and other from
letters written in terms ofT and 'me'. (The use of paragraphs, punctuation and
other matters of style may also indicate changing aspects of everyday culture.)
How the language and style of the letter writers (and readers) is to be linked
to the conceptual concerns of social scientists is, of course, a problematic issue.
We will pick up on some of the relevant themes in Chapters 6 and 7. For now,
we need only to note that relatively straightforward enumeration strategies
have their place in content analysis and have been used very successfully in the
social sciences. Yet, in the sociological tradition, especially, there have been
other, much more sophisticated approaches to matters of content and at this
introductory stage we need, at the very least, to be aware of them.

The meaning of the Song

In the Old Testament there is a book entitled the Song of Songs. It was origi­
nally written in Hebrew with touches ofAramaic, and borrowings from Greek

Copyrighted Material
 BAS I C TH E M ES 23

and Persian. Its earthly genealogy is unknown, but it was probably written
between the tenth and second century BeE. Depending on the translation, it
can appear as a poem of considerable elegance and beauty. It also has a struc­
ture of considerable complexity - though that does not concern us here - and
it has been variously interpreted. But what is its meaning and where does such
meaning reside?
In rabbinical writing, for example, the Song of Songs has been considered as
an allegory of God's love for Israel. In Christian commentary it is Christ that
is invoked, or God's love for the Church. Underlying these pious interpreta­
tions, however, is a poem that clearly touches (I think that I have the correct
verb here) on issues of sexual love and carnal knowledge. Thus, there is a great
deal of reference to loins and jewels, as well as references to breasts as ripening
fruit and so on. In that sense the Song of Songs may be interpreted as an excep­
tionally sensual poem of erotic secular, rather than theological, interest. It is in
any event a poem that has multiple layers of 'meaning' and therefore varied
possibilities for interpretation (Bloch and Bloch, 1 995) .
Issues of meaning and interpretation have been central to social science since
the late nineteenth century, though sociologists have been more concerned
with the meaning of action rather than the meaning of text. Nevertheless, the
problems about what is to be interpreted and how it is to be done are similar
to considerations of both action and text. Max Weber (1 864-1 920), for exam­
ple, regarded the interpretation of meaningful action as one of the central tasks
of sociology. In fact,Weber is a useful example to us because he developed what
is probably his most well-known 'thesis' on the basis of textual interpretation.
That thesis (often referred to as the Protestant Ethic Thesis) concerned the
affinity between the logic of capital accumulation and productive capacity on
the one hand, and the belief systems of European Protestants on the other. And
in order to establish the credibility of his claims, Weber drew widely on his
reading of texts (theological writings, diaries, etc.) of mainly English Puritans.
He thereafter felt able to outline what he considered to be a typical world­
view of early Protestants (Weber, 1 930) . Clearly, Weber's selection of textual
materials could be scrutinized and criticized as appropriate or inappropriate,
and it has been claimed, for example, that he placed an undue emphasis on the
writings of ascetic English Puritans in his characterization of the 'Protestant
Ethic'. Those problems, however, need be of no concern here, but we can
usefully ask questions about how Weber's interpretation of text related to the
matters of meaning. Whose meaning was Weber in search of, for example? Was
it, perhaps, the authorial intent of people such as Baxter (one of Weber's sources)
and Baxter's contemporaries? If so, then how are we to know that Weber inter­
preted Baxter's intentions accurately? Perhaps Baxter's writings, like the Song of
Songs, are open to various and multiple interpretations. Perhaps Baxter himself
would have placed significantly different interpretations on the same work at
different stages in his own life. Equally, Weber, as a reader of Baxter's Christian
Directory ( 1 673), at the cusp of the nineteenth and twentieth centuries, would

Copyrighted Material
24 USING DOCU M ENTS IN S O C I A L R E S EARCH

have offered a different interpretation of the work as compared with a late

eighteenth-century reader and so on. There can be few doubts that Weber
brought a great deal of himself and his learning to his reading. So perhaps we
would have to take account of Weberian meanings as well as Puritan intentions
when we assess the essays that are collected together under the title of The
Protestant Ethic and the Spirit if Capitalism (essays that were brought together in
book form by Talcott Parsons only in the 1 930s).
The lesson here is that words may be conveyed by text, but meaning and
interpretation are undertaken by human actors and attempting to access sub­
jective intentions and meanings of actors is a difficult (some might say impos­
sible) task. There are clearly a large number of problems to be addressed in our
one example alone. We will elaborate on some of those difficulties in Chapter 6,
in particular. It may as well be stated here, however, that in terms of the frame­
work adopted in this book, a search for 'meaning' is akin to a search for pigs
that fly. Indeed, we shall note that when we look at the content of documents
it is schemes of referencing that need to be analysed rather than systems of
meaning. In other words, our emphasis needs to be on the social activities
through which texts are appropriated rather than psychological properties of
the reader.

The structure of a dictionary

Pick up a dictionary. The dictionary will contain words and definitions of

words, and the words are arranged in some kind of order. In the western world
that order is alphabetic and runs from a to z; it couldn't be simpler or more
natural you may think. I I Yet dictionaries are not always organized in such a
manner. Consider a Chinese dictionary, for example, which clearly cannot be
organized in an alphabetic sequence. Instead, the inscriptions are organized in
terms of common picture parts or 'radicals' . So one way of organizing words is
to group together all those words with the same pictorial root - say the symbol
for the Devil or evil - irrespective of exact meaning. The individual words and
their meanings are, to say the least, central to the content of a document, but
it is also legitimate for us to take into consideration that structure in terms of
which the dictionary is held together.
This notion of a structure as a kind of scaffolding out of which individual
and specific images or objects are constructed is one borrowed from Saussure
( 1 983) . The latter generally distinguished between language and speech,
language containing, if you like, a finite and universal structure out of which
is generated an infinite number of sentences, spoken in an infinite variety of
circumstances. In fact, for Saussure, language was a prop erty of collectives - of
humanity as a whole as it were. Speech was something exhibited by individu­
als and was dependent on language. In later years these basic ideas were
extended to many other fields of human endeavour and action. Roland Barthes
( 1 985), for example, attempted to apply the distinction to the study of women's

Copyrighted Material
 BASIC THEMES 25

fashion and style. Claude Levi-Strauss, rather more intelligibly, applied the
notion to such things as the study of Totemism ( 1 969) .
Structuralist thought had a powerful impact on West European social science
during the 1 960s and 1 970s, and the development of that thought saw endless
twists and turns and intellectual acrobatics. Franyois Dosse (1 997a; 1 997b) has
traced many of the key developments in his majestic and masterful history of
structuralist thought. The detail of the history is far beyond the concerns of this
book. One point of common interest in structuralism, however, is that it is a
method that often eschews the search for the elusive 'meaning' of a document,
and focuses instead on how what is said has been arranged. In this respect Levi­
Strauss's Totemism provides a first-class example. It deserves a few moments
consideration.
The fact that human beings in non-western societies had been known to
associate themselves with various species of flora and fauna was an observation
that had fascinated anthropologists from the late nineteenth century onwards.
Durkheim, in his The Elementary Forms if the Religious Life ( 1 9 1 5), for example,
had commented on the manner in which aboriginal Australians had inscribed
snake and other symbols on stones and wood and henceforth regarded such
artefacts as 'sacred'. Other observers had noted how non-western peoples
sometimes referred to themselves as birds or bears or whatever and used
various species of animal as 'totems'. Until the publication of Levi-Strauss's
Totemism, the predominant question asked about such activities concerned
their meaning. What does it mean, for example, when human beings claim that
twins are like birds? The genius of Levi-Strauss was to side-step this kind of
question entirely. In his view it made no sense to ask for the meaning of such
associations but instead to ask how what was said was arranged. What were the
logical relations between things in a classificatory scheme? What was the
underlying structure based upon? How were one set of objects opposed to
another and a further set of objects allied with others? (You will recall that the
Chinese dictionary, for example, groups words relating to 'evil' together, imply­
ing that words relating to 'goodness' must also be grouped together in another
section of the dictionary.) In following through on such questions Levi-Strauss
came to view totemic systems as systems of taxonomy rather than as systems
of belief (or religion) , and in that respect to be directly comparable with the
taxonomies found in western society.
A focus on the arrangement of the words and sentences and things, instead
of on meaning, has much to recommend it. In a somewhat different theoreti­
cal context it was an idea taken up by Michel Foucault in his The Archaeology
if Knowledge ( 1 972) . This was a work in which Foucault sought to study the
nature of what he later referred to as 'discourse'. Things are both represented
in discourse and shaped and fashioned through discursive practices. In
Foucault's sense of the term a discourse expresses itself through statements and
sentences. Yet, what is of interest to social scientists is not the surface feature of
statements (about the objects of medicine, or grammar, or botany, or sexuality) ,

Copyrighted Material
26 USING D OC U M ENTS I N S O C I A L R E S EARCH

but the underlying rules and principles that bind such statements - and their
authors - together in a unifying matrix. Foucault saw discourses as having a
history - a beginning and an end - and he viewed one of his tasks as tracing
the starting and finishing points of specific systems of ideas. More importantly,
Foucault linked discourses and 'discursive regimes' to a world beyond the
text. 12 Thus, 'A statement must have substance, a support, a place, and a date'
( 1 972: 1 0 1 ) .13 By linking statements to a non-textual world (of people, roles,
places, buildings and institutions) Foucault opened up a route for discourse
analysis that moved beyond the world of linguistics and textual analysis, and
into the world of social practices. It is such a route that will be followed in this
book.

Conclusions and key points

Our concerns have been widespread and varied, yet the central arguments of
the chapter can be stated quite simply. I list them as follows:

• Documents form a 'field' for research in their own right, and should not be
considered as mere props to human action.
• Documents need to be considered as situated products, rather than as fixed and
stable 'things' in the world.
• Documents contain text, but text and documentation are not co-extensive.
• Writing is as significant as speech in social action and the medium through
which writing is carried should always be attended to. In everyday life, the
form, the list and the letter are, for example, as important as the verbal question,
the verbal answer and the command.
• Documents are produced in social settings and are always to be regarded as
collective (social) products.
• Determining how documents are consumed and used in organized settings -
that is, how they function - should form an important part of any social
scientific research project.
• Content is not the most important feature of a document.
• In approaching documents as a field for research we should forever keep in
mind the dynamic involved in the relationships between production, consump­
tion, and content.

In the next chapter we shall look at some facets of the production process. In
Chapters 3-5 we shall examine the use of documents in action. In Chapters
6-8, we shall turn to matters of content and the rhetoric of social research with
documents. Chapter 9 will bring the book to a close. It will do so by opening
up a further dimension for research into documents - that dimension relates
neither to production, nor consumption, nor even content, but rather to the
process of exchange.

Copyrighted Material
 BASIC THEMES 27

RESEARCH EXERCISES

Exercise 1.1

Mundane documents such as shopping lists, 'to-do' lists and appoint­

ment d iaries usually function in a number of ways. For example, they
can serve as a form of 'external memory' for an individual, they can
pattern the timing and order of future activity, and they can act as
simple records of things done. Consider identifying a small sample of
people (say, N 5) who draw up l ists and/or appointment diaries on a
=

regular basis and question them about how they use such documents.
Pay particular attention to the following issues. (1) Who it is that
authors the documents. (2) The extent to which the document reflects
and structures relationsh ips between the author and the user (and how
such structuring is achieved). (3) How the document functions in the
everyday l ife of the user (try to be as exhaustive as possible here).
(4) The extent to which the document may be said to act back on its
creator - noting, of course, exactly how this is done. Then consider
expanding the sample so as to gain coverage of additional kinds of
user or to explore elementary hypotheses that may have emerged from
working with the initial sample .

Exercise 1.2

As we shall note in Chapter 5, diaries are documents that can be used

to function in various ways. For example, they can serve as a record or
'log' of things done. They can also serve as an aide-memoire, as a
receptacle of a personal confession, as an aid to dieting (or quitting
smoking), as a legal record (as might be the case with someone
attem pting to document a d isability), and as social scientific research
tools. Using suitable WWW searches using the terms 'journals +
diaries' draw up an extensive list of functions for the diary/journal. On
the basis of your findings generate some hypotheses about the rela­
tionships between function and content.

Notes

1 The sorcerer's apprentice used his novice spells to get a broom to carry buckets of
water. Unfortunately the broom then acted independently of the apprentice's com­
mands - and flooded the sorcerer's house. It is something of a Mickey Mouse tale,
the complete version of which can be found in a 1 779 German poem by Goethe.
See, www.fln.vcu.edu/goethelzauber.html
2 Lincoln and Gubba (1 985) draw a distinction between a document and a record
(based on their functioning) . It is not a distinction that I intend to adopt herein for
a document can function in many ways. It seems somewhat invidious, therefore, to

Copyrighted Material
28 USING DOCUMENTS IN SOCIAL R E S EARCH

isolate just one such function. Other commentators try to classifY documents
according to whether they are 'private' or public' 'primary' or secondary, solicited
or unsolicited (see, for example, Burgess, 1 984; Finnegan, 1 996). However, it is
clearly conditions of consumption and use that will determine which of these
categories a document will belong to. The active document is usually too slippery
a creature to fall neatly into such classificatory traps. As the Torres Straits example
(in this chapter) shows, what is primary in one frame is often secondary in another. .
3 Dorothy Smith (1 990: 1 2 1 ) has also argued, though in a narrower frame than is
adopted here, for a move away from a study of the inert to a study of the 'active
text'.
4 It is, however, notable that in social scientific research, 'speech' extracts are always
mediated through writing - as, for example, in the work of those who undertake
conversation analysis.
5 The operation of which he saw as being 'based upon written documents ("the
files") ' . See Weber (1 979: 957).
6 I use the word text to refer to written (or printed) inscription. Consequently, text
is to be distinguished from talk. It is a simple point, yet it is one that needs con­
siderable emphasis: first, because in some elementary methods manuals text is often
loosely conflated with talk - see, for example, Burgess (1 982: 131) and Watson and
Seiler (1 992); secondly, and more importantly, because in the world of the semi­
oticians (see, for example, Barthes, 1 977; Derrida, 1 976), any and all things can be
regarded as 'text', and subsequently decoded according to the rules of semiotics.
Thus, in contemporary cultural studies as, say, exemplified by Hall (1 972), images,
sounds, talk and writing are often bundled together as 'text' that is subject to
encoding and decoding procedures. A key example for Hall would be television
discourse. In addition, some writers also refer to social action as text in order to
argue that the study of action can be approached in the same manner as one might
approach a written text (Ricoeur, 1 977). A further position is that adopted by
Geertz (1993: 452), who conflates text and culture, arguing that the latter is, essen­
tially, an ensemble of texts.
7 The concept of performance - borrowed from Goffinan (1 959) - has been more
recently adapted by Law (1 994) and Mol (1 999; 2000) to place emphasis on how
things are done. An emphasis on doing encourages us to avoid speculation about
private mental operations such as what people are 'thinking' and 'believing' and to
concentrate on the visible effects of activity.
8 An alternative way of looking at this problem is to use the concept of 'actant'
(Greimas, 1 987). Actants may be said to have functions and effects. For example,
the concept of author is suggestive of an 'actant' in so far as it gathers up all the
processes, activities and actors involved in the production of a text into a single
identifiable bundle. That bundle is usually a named person who is then viewed as
the executor of all such processes, activities, etc.
9 The concept of discourse is a tricky and complicated one. Van Dijk (1 997)
provides a good overview of possible meanings of the term, whilst Gill (2000) use­
fully discusses some meanings of the term ' discourse analysis' . For an indication of
the way in which this is to be used in this book see pp. 25-6.
10 Frankenstein created a monster that turned against its human creator, see Mary
Shelley's Frankenstein or the Modern Prometheus (1 996 first published, 1 818).
-

11 Even within the English language, however, dictionaries exhibit varying forms of
structure and classification. So although The Oxford English Dictionary lists words
alphabetically, the pattern of each entry is determined by chronology of use rather

Copyrighted Material
 BASIC THEMES 29

than by popularity of usage. So an archaic meaning of a word listed in, say, 1 646,
will always come before a more popular meaning listed as 1 990, or whatever.
12 The supposition of a realm beyond text constituted the basis for one of Foucault's
criticisms of Derrida's philosophy. The latter had claimed in Of Grammatology
(1 976: 227) that there is nothing beyond the text. The assertion was part of a larger
claim to the effect that texts have no external referent against which their truth or
validity may be assessed. According to Derrida (1 988), one can only assess a text
by factors internal to its composition. It is, as we shall note, a problematic position
to adopt.
13 Other attempts t o link text t o action under the umbrella term o f discourse have
been developed by a n,umber of Marxist theorists such as Ernesto Laclau and
Chantal Mouffe. For outlines of the central debates, see Torfing (1 999).

Copyrighted Material
2
Prod u c i ng Facts

From Paris :1.748 to Geneva :1.998 30

Death: a progress report 32
Enumerating neurotics 43
Conclusions 47
Notes 49

From Paris :1.748 to Geneva :1.998

In the Paris of 1 748 a policeman by the name ofJoseph d'Hemery set up a

series of fil es. He named them 'Historique des auteurs' (Damton, 1 984) .
D'Hemery was an early version of what, today, would be called a secret police­
man, and his auteurs were what were later to be referred to as intellectuals and
ideologues - though neither word existed in 1 748. His interest to us is that he
faced a set of problems common to all those who seek to report on the nature
of the world. For a start, he had to define an 'auteur' - a business no easier then
than now (see Foucault, 1 979). Was an auteur someone who wrote a book or
a play? Or was it anyone who wrote a pamphlet or even a few lines of prose?
(In fact some 67 of d' Hemery's auteurs wrote nothing at all) . How was he to
impose order on his files? What categories was he to use to organize the infor­
mation that he collected? Indeed, what information was he to collect as
relevant to his purpose, and what to ignore? (D'Hemery, for example, consid­
ered the physiognomy of a person as important as their ideas; thus he describes
Voltaire as 'Tall, dry and the bearing of satyr' as well as a 'bad subject'.) In gather­
ing data on his auteurs, d'Hemery very naturally gave vent to his own modes
of expression, and thus helped to create a new genre of document, a new form
of literality - the secret police file. In that respect he is an auteur in his own
right, and his writings provide invaluable insight into how a loyal state agent
viewed the enemies of the ancien regime that was pre-Republican France. (For
insight into a more recent secret police file see Ash, 1 997.)
Categorizing, defining, sorting, ordering, including, excluding and reporting
on the world: these are tasks that concern social scientists every bit as much as

Copyrighted Material
 P R O D U C I N G FA C T S 3:1.

secret police (see Kwasnik, 1 9 9 1 ) . But how, exactly, are such tasks achieved?
The processes that underpin the manufacture of documents are rarely made
visible or accountable. Those of us who handle published documents as
resources for research, for example, see only a finished product, an object ready
for use. Yet in order to produce that object various auteurs have had to call
upon a complex system of rules, conventions, organizational strategies and
conceptual schemes. Writing anything involves one in a system of production.
For example, in order to produce a statistical report, someone has to devise a
classificatory system, and operate rules that dictate how events and objects are
to be assigned to the specific classes - as to how, say, 'bad' and 'good' subjects
(and satyrs) are to be recognized and allocated. Someone has to devise rules of
precedence to cope with those instances in which an individual or a case has
more than one property - say, 'bad subjects' who are satyrs. On top of that,
someone has to devise rules about how numbers are to be allocated to objects -
how, for example, acts of disloyalty or treason are to be counted. These days, of
course, most of our counting is undertaken in terms of well-defined taxonomies
and rule-books. In d'Hemery's day it was devised and developed in a some­
what more rudimentary fashion. The essential principles of the construction
process are, from our standpoint, however, timeless.
As well as attending to classification and order in writing up their reports,
auteurs have, necessarily, to attend to other matters. They have to keep an audi­
ence in mind and decide on the purpose of their documentation. And so the
systems of relevance of users and readers also impinge on the production
process. Further, they have to decide how to locate themselves within the
document - whether to declare their presence as an 'auteur', or whether to
mask their presence behind the name of a committee, or an office of some kind
or invisible ' other'. In this last respect it seems as if d'Hemery was not a very
successful secret policeman by modern standards - simply because he declared
himself to be present in the files. Writing notes and reports as a committee, an
office, or as an anonymous functionary would ha\re been far more professional.
In order to get to grips with these issues of authorship, rule-based systems,
and the design and production of documents it is often useful to disassemble -
comma by full stop - the documents that we use as data sources. That is to say,
to investigate each stage of the process by means of which a document has
been put together, concentrating on how each component has entered into the
production process, and how the parts are eventually fitted together. Indeed,
looking at how documents are manufactured invariably provides insight into
how we assemble facts about the world in general. For, in many respects, the
procedures through which an auteur (such as d'Hemery) assembles a report on
the world is not wildly different from the way in which ordinary individuals
assemble accounts of everyday and routine interactions. After all, each and
every one of us has to devise and apply some kind of classificatory and cate­
gorizing system for dealing with and describing the circumstances and the
people that we daily encounter. (Harvey Sacks, who studied processes of the

Copyrighted Material
32 USING DOCUMENTS IN S O C I A L R E S EARCH

latter kind, coined the term 'membership categorization device' (MCD) to

assist with the description of the process - see Lepper, 2000; Silverman, 1 998) .
In this chapter we are going to unzip a number of what we might call quantita­
tive reports, in order to illustrate some key processes involved in their produc­
tion. We begin by looking at a World Health Organization (WHO) report on
health statistics, published in Geneva during 1 998. Following that, we shall
look at some of the issues that impinged on the production of a report on psychi­
atric disorder in Great Britain. Our aim is to demonstrate that 'like all knowl­
edge', statistical reports 'must be analysed as a product' and are 'never mere
givens' (Hindess, 1 973: 1 2) . We shall begin with the observation that, as pro­
ducts, they reflect both conceptual and organizational (technical) features of
their production. It is to the investigation of such matters that we now turn.

Death: a progress report

It is something of a paradox that one of the most useful and malleable of quan­
titative measures that is called upon to assess the 'health' of populations is the
death rate - or, more accurately, the mortality rate. Thus, the rate (per thousand
born) of babies who die in the first year of life (the infant mortality rate) , for
example, has long been used by agencies in the advanced industrial societies as
a key measure of both the health and quality of life of a population. Health
agencies are, of course, also interested in what people die from as well as how
many people die at any given age. Thus, WHO publishes, on an annual basis, a
manual of world health statistics (WHO, 1 998) . The manual provides data on
both the numbers of people who die in any one country during a given year,
and the cause of death of the individuals concerned. In that respect the
manual exists as a resource for social scientific or epidemiological study. One
can lift it off the library shelf and consult it for facts about mortality, or health,
or, if one wishes, transpose and integrate the data into a measure of the 'quality
of life' in Nigeria, Canada, the Ukraine or the USA. Furthermore, where
required, the facts within the manual can be plotted on graphs, slotted into
tables, or correlated one with another to reveal trends and patterns (see Unwin
et al. ( 1 997) for a review of possibilities).Yet, as with many reports of this kind,
it is often more revealing to look not so much at what can be derived from the
document, but at the building blocks of the document itself - at how the
report was put together in the first place.
The foundation stones of big reports - such as the WHO report - are often
designed and set at some distance from the final product. In the WHO case this
is so in both a bureaucratic and a geographical sense. One 'stone' is the death cer­
tificate. In most western societies a medical practitioner completes this certificate,
and it provides a cause of death for each deceased person. The certificates are then
processed through a series of local, regional and national agencies so as to com­
pile a picture of mortality in any given country. It is from such national pictures

Copyrighted Material
 PRODUCING FACTS 33

that the international (WHO) picture is derived. To illustrate key procedures I am

going to begin with the concept of a cause of death - and move upward.
We should be aware that deciding on the cause of death of any individual is
not a simple task (Bloor, 1 99 1 ; Prior, 1 989) . This is partly because people com­
monly die of many things at one and the same time, and it is not always easy to
disentangle one cause from another. A serious infection of the lung, for exam­
ple, may accompany a growth in the airways, leading to eventual heart failure.
A minor wound may encourage the development of septicaemia, or whatever.
In addition, we have to be aware that the physiological 'causes' of death are not
always easy to determine in the absence of autopsy. For instance, brain tumours
are particularly difficult to diagnose at death without an autopsy and it would
be impossible to determine the precise nature of a lung infection without
recourse to laboratory analysis. Despite this, only a minority of people are ever
autopsied. Indeed, in many cases, the physician or coroner or other individual
who certifies a death may not consider it worthwhile to find out exactly why
a person died. All that is important is to determine whether the individual died
'naturally' or from foul play. Further, and even if one feels able to overlook the
aforementioned (technical) obstacles, one is faced with the fact that, in the dis­
course of western medicine, only some causes of death are regarded as legiti­
mate in the first place. In that context, readers may be delighted to note that one
is simply not allowed to die of either 'old age' or 'poverty'. More significantly,
the exclusion of such causes suggests that modern western societies have very
specific vocabularies of causation as far as matters relating to death are
concerned, and a distinct image of what can and cannot cause a death.
So what can one die of? The answer to that question is buried (if that is not
an entirely inappropriate metaphor) in another WHO manual. This latter is
called The International Statistical Classification of Diseases and Related Health
Problems (WHO, 1 992) . It is often referred to in an abbreviated form as the
ICD.The current edition of the manual is the 1 0th, and so the abbreviation is,
more accurately, ICD- lO. lCD- 1 0 provides a list of all currently accepted
causes of death, and they are classified into ' chapters' . Thus, there are chapters
relating to diseases and disorders of the respiratory system, the circulatory
system, the nervous system and so on. In different decades different diseases and
causes of death are added and deleted from the manual. HIV / AIDS is an obvious
example of an addition and it appears as a cause of death only in lCD- 1 0,
whilst 'old age' as a cause of death was eliminated in ICD-6. In all cases, of
course, the conceptual architecture in terms of which death is comprehended
is structured around the human body (Foucault, 1 973: 3) .
As well as containing a long list of medical causes of death, the ICD also
contains rules about which causes of death are more important than others. So
when a person dies of many conditions, the people responsible for coding the
data on which the health statistics depend 'know' which cause to select as
the cause of death. Thus, diseases of the heart, for example, commonly take
precedence over diseases of any other organ, and cancers take precedence over

Copyrighted Material
34 USING DOCUMENTS IN SOCIAL R E S EARCH

infections and so on. These rules also change from decade to decade. Thus,
during the early part of the twentieth century diseases of the liver and lung
took precedence over disease of the heart (see Prior, 1 989) .
In the context in which it is here considered, the ICD-10 is an excellent
example of what we might call a generative document - a document that lays
down rules as to how other documents should be constructed. It contains both
the conceptual structure in terms of which any explanations have to be built,
and, in addition, rules for the building process. Generative documents come in
various forms. The ICD is, perhaps, one of the most important for getting to
grips with professional, 'expert', understandings of physical health and illness
and serves, in many ways, as a window into western culture (Bowker and Star,
1 999) . A related publication - The Diagnostic and Statistical Manual of Mental
Disorders (American Psychiatric Association, 2000) or DSM - is available for
the classification of psychiatric (mental) conditions. One might say that the
DSM provides the conceptual architecture in terms of which western culture
comprehends disorders of the mind. And once again, if a disorder is not listed
in the DSM then it is not regarded - in expert discourse - as a distinct
psychiatric condition. Post-traumatic stress disorder is, for example, recognized
as a disorder only in DSM-III (first published in 1 974) . The route by means of
which it attained inclusion is, in itself, a matter of some interest (see Young,
1 995) . As for stress-related disorders, generally, they tend to appear in the ICD
only from the 9th edition onward. (In other words, they were not regarded as
legitimate medical conditions before their inclusion, and certainly would not
have been enumerated by health agencies.)
It is already clear then that the WHO World Health Statistics Annual (1 998)
that was mentioned at the start of this section is a secondary document - its
production dependent on the existence of pre-given items. First, there is a con­
ceptual structure, developed over decades and reflecting fundamental assump­
tions about the nature of disease, death and the human body. It is best
encapsulated in the ICD. Next there are national statistics. These latter sum up
all the individual details of people who have died from HIVIAIDS, or lung
cancer, or pneumonia, or road accidents, or whatever, and form the basis for
the published tables. The national statistics, in turn, are produced partly on the
basis of the death certificates of individuals, and partly on the basis of the ICD .
rules and codes. So, producing a report on mortality clearly requires the develop­
ment and exploitation of a conceptual (theoretical) as well as a technical and
organizational structure. Indeed, to understand the fundamentals of the WHO
Annual, we probably need, once again, to step down a couple of notches. To
that end it would be as well for us to focus on a single topic within the WHO
publication. Any topic, such as the provision of heart disease statistics, cancer
statistics or AIDS statistics, would serve our purpose. Here, however, I intend
to focus on a set of numbers that are reported upon at the foot of each of the
national tables, namely suicide statistics.

Copyrighted Material
 PRODUCING FA C T S 35

12

10

g 8
o
o
o
-

; 6
Co
en
!
� 4

Year
I -+- Rates --- Log rate I
FIGURE 2 . 1 Suicide rates for Northern Ireland, 1968-98

Disassembling Durkheim and other users of statistical data

Suicidal behaviour has fascinated social scientists since the birth of the disci­
pline. Social scientific and mathematical interest in the matter arises from a
number of considerations. Above all, it is clear that although suicide is a
supremely individualistic act, the study of it at a population level seems to
exhibit distinctive social patterns. In fact, the earliest pattern to be noted by
social commentators (mainly early nineteenth-century mathematicians) was
the stability of the rate of suicide in each European society. A rate is normally
measured per hundred, or per thousand, or per ten thousand (the appearance
of rates, though common today, is essentially an invention of nineteenth­
century social science) .When one examines rates of suicide, say, per one hundred
thousand in a society such as France or the UK or the USA it is evident that
over some decades, the suicide trace remains reasonably static. Indeed, the
larger the denominator chosen the more stable the rate will appear. Thus
measuring instances of suicide per million or, if possible, per 1 0 million will
always make it appear as if trends in the phenomena being examined are pre­
dictable and stable. Using raw data, on the other hand, can often make it seem
as if the phenomenon under study is erratic in its occurrence. In Figure 2 . 1
I have plotted a chart showing the number o f suicides in just one small part
of the world - namely, Northern Ireland between 1 968 and 1 998. There are
two details to note. First, the rate varies over the time period. Secondly, the
trend is rather higher after 1 982 compared with the period 1 968-82. The

Copyrighted Material
36 USING DOC U M ENTS IN SOCIAL RESEARCH

shift is slightly more obvious in the lower of the two traces (the log rate or
the rates transposed into logarithmic scores) . The transformation of rates into
logs and other measures is frequently carried out by quantitative researchers
to get good 'fits' of data. Indeed, such manipulations are usually necessary for
the manufacture of statistical facts - an issue that, unfortunately, extends
beyond the aims of this book. Readers interested in common forms of trans­
formation and something of the reasoning behind them are best referred to
Tukey ( 1 977) .
It was undoubtedly the apparent stability of data patterns that encouraged
many nineteenth-century commentators to suggest that social life is governed
by laws every bit as deterministic as are the laws of physics in the natural world.
Thus, the English historian Buckle, for example, developed the notion that, in
each society, there must be some kind of physical force compelling constant
proportions of people to commit suicide year after year (see Hacking, 1 990) .
Such ideas belonged very much to a discourse of what is often called, nineteenth­
century positivism, and the French sociologist Durkheim (1 855-1 9 1 7) was
very much absorbed in the development of that discourse. Indeed, for him, the
study of suicide seemed to offer an excellent occasion for examining the nature
and dynamic of a 'social fact', and to elucidate on the nature of the laws that
might determine it. As one might guess, the starting point of Durkheim's
published analysis of suicide was in the statistical patterns.
A close study of Durkheim's 1 897 text (translated into English 1 95 1 ) always
pays a dividend. Even the casual reader cannot fail to note, for example, the
manner in which the author uses tables and maps to underpin his general
thesis. In fact the use of cartography - to show the distribution of suicide in France
between 1 878 and 1 887 mapped against the distribution of drunkenness, alco­
holic insanity, mean family size and so on - provides one of the earliest instances
of what was later termed the ecological fallacy. The latter refers to the fallacy of
drawing conclusions about individual behaviour from data that refer only to
collective behaviour. For example, Durkheim argued that if one studied maps
of 'La France' one would note that the regions with high rates of alcohol
consumption (say, Normandy and Brittany) did not coincide with areas showing
the highest suicide rates (The Paris Basin and the Cote d'Azur) . He therefore
concluded that the two phenomena were unrelated. Even in terms of a positivist
social science, however, such a conclusion is unjustified, since it does not preclude
the possibility of the two phenomena being linked, somehow, in the everyday
lives of the people who actually committed suicide. (All the suicides in Paris
could have been alcohol dependent after all . ) Such difficulties do not, of course,
prevent people from using maps to undertake social scientific and other forms
of reasoning. Maps of mortality, for example, are still plentiful enough in the
modern day, though what it is that we are supposed to read off them is never
made entirely clear by their authors (see, for example, Burgher, 1 997). As well as
developing the use of mapping techniques, Durkheim was also among the earliest
social scientists to use rates - per million and per thousand - as statistical props

Copyrighted Material
 PRODUCING FA C T S 37

to his arguments. Ratios also figured in his thinking, and although he would
have been unaware of it, he often provides what these days are called odds ratios
on, for example, patterns of male suicides (195 1 : 1 99).
We have no need to trouble ourselves too much with these statistical details.
We should merely note for the time being that, as far as Durkheim was con­
cerned, the suicide rate exhibited stability and that stability called for explana­
tion. In search of an 'explanation' Durkheim unravelled his empirical data
somewhat further, though, once again, we have no cause to delve into the
detail of the unravelling. Suffice it to say that his reasoning focused on social
variations in the rates. For example, he spotted a variation in the rates of suicide
committed by males as against females, married males as against unmarried
males, and members of the Catholic faith as against the Protestant faith, and so
on. From those observations Durkheim then proceeded to link 'facts' about the
lives of the married and the unmarried, of Protestants and Catholics and Jews,
to explain the observed variations. (That is, he sought to reason why there
should be a higher suicide rate among males than females, among the unmarried
than the married, among people associated with Protestant churches than those
associated with the Catholic or Jewish faith.) His theorizing is, in so many
ways, rather stunning, and the general drift of his empirical claims still stands
today as what are commonly regarded to be valid generalizations about the
social distribution of suicide (Chauvel, 1 997) .
We could, of course, usefully disassemble the component parts ofDurkheim's
analysis of suicide down to the dots and commas, and that would also pay
dividends. Some years ago Douglas (1 967) more or less did just that in his The
Social Meanings of Suicide. Indeed, in that book Douglas not only rakes through
the Durkheimian theory of suicide with a fine tooth comb, but also traces the
influence that Durkheim had on other social factor theories. That is, theories
that associated such things as gender and social status with the commission of
suicidal acts. Indeed, post-Durkheimian researchers in the positivist mould
eventually developed a long list of what would today be called 'risk factors' for
suicide on the basis of statistical analyses. One of Douglas's key conclusions,
however, was that the manufacture of statistical associations between suicide,
and factor 'X' (or 'Y' or 'Z'), more often than not, slid over a consideration of
the presence of what he termed situated meanings. That is to say, positivistic
social scientists tended to ignore the simple matter that suicide involves a
process of judgement and evaluation - judgements that involve, among other
things, the attribution of motives to the perpetrator of a suicide. Such attribu­
tions are, of course, most commonly made by family members, the police,
coroners, or the members of a coroner's jury. And they are necessary to reach­
ing a conclusion as to whether a particular death is a suicide rather than a
homicide or an accident. According to Douglas, positivistic researchers not
only ignored such meanings, but tended to impose their own (second-order)
interpretations onto the statistical data - thus committing, so to speak, a double­
strength methodological error.

Copyrighted Material
38 USING DOCUMENTS IN SOCIAL R E S EARCH

Situated meanings and the production of data

The claims ofJ.D. Douglas emanated from a style of social scientific work that
revolutionized the study of 'social facts' during the 1 960s. The revolution
emanated mainly from academics based on the West Coast of the USA. The
latter, in particular, sought to take issue with the manner in which social scientists
tended to reify the world about them. According to Garfinkel (1 967) , for
instance, acts of reification appeared at every level of conventional social
scientific analysis and explanation. Thus, when psychologists or sociologists
were asked to explain variations in, say, rates of criminal behaviour or of
suicide, they would reify the topic of study (crime, suicide or whatever), the
factors that they called upon to explain such variations (socio-economic status,
poverty, mental illness and the like), and even the very mechanisms that
they used to associate the object of study and the explanatory factors (such as
mathematically designed covariances) .
A particular bete noire of the ethnomethodologists were social scientists who
worked in the manner of the Durkheimians. For, to the ethnomethodologists,
the kind of work that Durkheim undertook is drenched in serious method­
ological flaws. Among such flaws one would have to emphasize his tendency
to take 'suicide' as an immediately recognizable and incontrovertible act. For
example, Cicourel (1 964) - another member of the ethnomethodological
camp - had highlighted how terms such as suicide,juvenile delinquency, crime
and so on were essentially linguistic categories, and the procedures in terms of
which events are assigned to such categories is what sociological study should
be all about. In that sense, the Durkheimians appeared to start at stage 2 of the
research process - counting and associating events such as suicide with events
such as 'alcoholic insanity' - when they should be starting at stage 1 - investi­
gating the procedures by means of which happenings in the world are assigned
to classes.
According to the ethnomethodologists, then, one major task of the social
scientific researcher is to study the manner in which ordinary people recognize
and impose order on events as they unfold in the everyday world. That is to
say, a study of the ways in which members of society make sense of the situa­
tions that they encounter, the ways in which they manage to classify them (and
the ways in which they consequently organize them as ongoing accomplish­
ments.) One of Cicourel's interests in this respect was the manner in which
'delinquents' and reports about delinquents were manufactured through
socially organized and socially sanctioned procedures of arresting officers, desk
sergeants and the like (Cicourel, 1 976) .
Making sense of situations that we encounter is, of course, heavily depen­
dent upon pattern recognition. This is as true of our routine, everyday work as
it is of social scientific work. So, acts of pattern recognition - of recognizing,
say, a greeting situation or a situation for apology - are in many respects similar

Copyrighted Material
 PRODUCING FACTS 39

to those that lay at the heart of social scientific research. In both instances we
are required to recognize the 'sameness' of events - such as, say, those instances
in an interpersonal exchange when apologies are always appropriate. Such a
process of recognizing sameness was referred to by Garfinkel ( 1 967) as the
'documentary method of interpretation' - a given instance of events is seen as
'documenting' the underlying category. The documentary method was said to
form the core of the practical or everyday reasoning process. And it is, perhaps,
already clear that a documentary method of interpretation comes into play
every time that we reach a decision as to whether a sudden death is 'natural' or
'unnatural' - that is, an accident, a suicide or a homicide - or, say, a particular
activity is or is not a 'crime'.

Suicide as narrative

Suicide is a sad and depressing business for all of those involved in its discovery.
It is rarely clear, however, whether any given death is as a result of personal
intent on behalf of the deceased or not. People are found dead. They are found
under the wheels of vehicles, and by the side of rail tracks. They are found lying
face down in rivers in the late afternoon. They are found in hotel bedrooms
with plastic bags over their head. They are found at home, dead in bed, shot
through the head, or in fume-filled garages. Yet others are washed up on
beaches. But few people leave written or verbal declarations of any intent to
kill themselves (and even if they have, such notes must be treated with caution).
Consequently, 'suicide' is always something of a problematic category. Indeed,
suicidal intent and motives have always to be read into the circumstances and
events in question. So suicidal motives are always imputed - that is, imputed to
the deceased by others.
Now one of the key insights of the ethnomethodologists was to focus on
the process whereby such imputations occurred, and a sudden death is trans­
lated into a suicide. In most societies it is a complex process. And the process
whereby relatives and friends, police investigators and coroners put together
a feasible narrative of death is not an easy one to research. A number of socio­
logists, however, variously attempted to trace the procedures, and in so doing
their work has generated some fascinating conclusions. Thus, J.M. Atkinson
( 1 978) focused on the reasoning processes of English coroners with respect
to suicide. Taylor (1 982) investigated the organizational processing of deaths
of people who had 'jumped' in front of London's trains. Instead of summa­
rizing those works, however, I am going to present some of my own data -
derived from a coroner's office - to illustrate the essential points of the
construction process. Some of the personal details have been altered so as to
disguise the identities of the deceased and I have added a few details (in
square brackets) to assist with the reading. The summaries are transcripts of
the coroner's 'findings'.

Copyrighted Material
40 USING D O C U M ENTS I N SOCIAL R E S EARCH

Case 2.1

Married m a l e . Age 28. Unemploye d .

C a u s e o f death: Poisoning b y alcohol, Val i u m and Dal mane.

In 1989 the deceased h ad been attacked and had received head

injuries. After that he suffered from head aches for which he took
tablets. He spent the night of June 9th alone, in his sister's house .
When the deceased failed to return home, his brother-in-law went to
the house where the deceased had stayed , and forced an entry. H e
found the deceased lying dead on the floor o f the sitting roo m . There
were 367 mg of alcohol per 100 ml in his b l ood . Its effects had been
increased by the use of the drugs Val i u m and Dalmane.

Case 2.2

Married Female. Age 58. No occupatio n .

Cause o f death : Poisoning b y Maproti line.

The deceased suffered from depression for 5 years for which she
had received hospital outpatient treatment. On 20th June 1990 her
h usband. was admitted to hospital with a chest complaint and she
visited him there on June 23'd . Later that day she was visited at
home by her grandson . The next d ay she failed to pay her customary
visit to her d aughter. Consequently her daughter called at her
mother's house and asked the pol ice to force an entry to the house
at about 16.00. She found her mother dead in bed with several
e mpty packets of Lud i o m i l [an anti-depressant] nearby.

Case 2.3

Si ngle Female. Age 20. Student.

Cause of death : Trichloroethane pOisoni ng.

On the 13th Novem ber the deceased was l iving in 32, Apple Street
with 4 other girls. She was known to be involved in substance abuse,
and h ad been advised to desist by her friends. On the evening of the
13th , a friend cal led to see her in her room at about 18.00, and about
19.00 she began to sniff 'Zoft' - a liquid used for removing plaster.
Around about 19.15 she began twitching and shouting and the 'Zoft'
was taken from her. Some moments l ater she collapsed on the floor.
An ambulance was called, and she was taken to the hospita l . She
was declared DOA at 20 .00.

Copyrighted Material
 PRODUCING FA C T S 41

Case 2.4

Married Male. Age 40. Salesman.

Cause of death : Alcohol and chloral hydrate poisoning.

The deceased had a 10 year h istory of depression. He had recently

received [ psyc h i atric] hospital treatment for his condition. At hospital
he had been prescribed ' Noctec' [the source of the c hloral hyd rate]
to h e l p h i m sleep. In the past he had often taken overdoses of pre­
scribed d rugs when in a low and confused state . He had a lso been
reckless in taking pills with alcohol. On the 17th of May h e returned
home after drinki ng, but was not drunk. He went to bed at 2 1.00 and
took 4 ' N octec' tablets. He was found dead the fol l owing morni ng.

These four narratives of death were written by coroners - the words within
square brackets are mine. The narratives are of considerable interest in them­
selves, not least for the manner in which they seek to describe the salient
history of an event and thereby single out some issues for mention whilst
ignoring others. For example, psychiatric histories, dates and times are men­
tioned, and so too are family relationships. On the other hand the financial
background of the individuals or, say, their religious beliefs are not mentioned.
(The use and application of terms such as 'history of depression', 'substance
abuse', 'reckless in taking pills' and so on forms potentially useful examples of
Sacks' MCDs - mentioned in the opening section of this chapter.)
In each of these narratives there is supposed to be sufficient information in
the narrative to enable any reasonable observer to form a judgement as to
whether the deceased intended to kill him- or herself, or whether death was
accidental. In that sense, such narratives function as what we might call meaning­
making devices.
Only one of the above deaths was coded as a suicide. The remaining three
were coded as accidental deaths. It would be natural to think that these deci­
sions were arrived at by detailed and considerable deliberation of a jury or
some such, and that the narratives provided above are only summary statements
that emerged from more complex analyses carried on elsewhere. However, it
was not so. The only information that the person coding the data had on these
events was as above. So the coder had to read into descriptions such as these
his or her own personal images of what a suicidal person might or might not
do. It would, of course, be useful if the coder had used simple rules about the
decision-making process - such as, for example, any mention of a psychiatric
history to be taken as being suggestive of a suicide. Once again, however, it was
not so. The decision-making process was, and remains, a messy and often inco­
herent one. Indeed, some time ago, Garfinkel ( 1 967) had indicated that no
matter how rule bound a coding system may be, the rules have always and in
every case to be interpreted by the coder. Such a process was referred to as 'ad

Copyrighted Material
42 USING DOCUM ENTS IN SOCIAL R ES EARCH

hoeing' - interpreting the information at hand in a manner that enabled the

coder to finish his or her task. The study of ad hoeing procedures forms a
potentially rich terrain for research into the construction of documents. (See,
for example, Benson and Hughes, 1 983; Reiner, 1 996.) For, we do know that
when we ask people (such as coders) to explain their reasoning they can and
do point towards rule-based systems. For example, they may call upon a rule
to the effect that 'several empty packets of a drug' is suggestive of suicidal
intent. Usually, however, such reasoning is post hoc and constructed for the
benefit of the listener. !
The significance of these detailed and somewhat concentrated deliberations
is that the suicide statistics that were referred to near the opening of this
chapter are, ultimately, assembled on the basis of such procedures. Indeed, the
examples that I have drawn upon are taken from work conducted in Belfast,
and it is on the basis of such narratives that the official statistics for the town
and the associated region are constructed. It is, however, important to realize
that suicide decisions have not always been executed in this manner. In fact,
during the period 1 968-82 there were a series of major changes in the law
concerning who could and could not bring in a verdict of ,suicide' (Prior,
1 989) . Mter 1 982 verdicts of suicide, accident and homicide were abolished
and so, in the strict legal sense, there are no suicide verdicts in Northern Ireland
at all. Deaths are categorized as suicide for the sole purpose of providing mor­
tality statistics (and this could be connected with the upward shift of the log
rate in Figure 2 . 1 ) .
A number o f important lessons concerning the production of statistical
reports can now be derived from our deliberations. First, the procedures
through which events in the world are enumerated constitute an important
'topic' for research in their own right. Most social scientists, of course, prefer to
gloss over these considerations and use social and economic statistics as a
'resource' - as if they reflected facts in the world unmediated by organizational
processes. (On the distinction between 'topic' and 'resource', see Zimmerman
and Pollner, 1 9 7 1 ) . We have seen that there are good reasons for refuting such
an argument. In fact, the claims that we have made above could be applied with
equal force and fervour to the construction of crime statistics, cost of living
statistics, birth statistics, marriage statistics, business statistics and any other
realm of enumeration that one might care to consider.2 Secondly, we have to
be aware that a consideration of the ways in which point prevalence rates (the
rate at which something occurs during a point in time) are constructed can be
extended to a consideration of the ways in which statistical trends are con­
structed. Thus in looking at the suicide rate we have seen how the production
of a graph - showing deaths over time - lulls both researcher and reader into
thinking that 'the same' facts are being reported upon. As we have briefly
noted, however, with respect to Northern Ireland statistics the agents who con­
struct those facts have changed markedly over time (they were coroners variously
with and without juries until 1 982 and then only coders) . Different agents call

Copyrighted Material
 PRODUCING FA C T S 43

TABLE 2.1 Prevalence of psychiatric disorders in private households by

gender. Rates per thousand of population in past week, GB 1995
Nature of disorder Females Males
Mixed anxiety and depressive disorder (MADD) 99 54
Generalized anxiety disorder (GAD) 34 28
Depressive episode 25 17
All phobias 14 7
Obsessive-compulsive disorder (OCD) 15 9
Panic disorder 9 8
Functional psychoses* 4 4
Alcohol dependence* 21 75
Drug dependence* 15 29
* Rates per thousand of population in past 12 months.
Source: Meltzer et al. (1995)

upon different reasoning processes and different MCDs in the deliberations

and therefore we can justifiably argue that the graph in Figure 2 . 1 provides a
false sense of unity to the phenomena we refer to as suicide rates. Durkheim's
tables did likewise. Finally, we have noted how the day-to-day activities of
those who manufacture documents work within well-established frameworks
of relevance and order: they use generative documents - documents that pro­
vide the conceptual framework in terms of which the world is reported upon.
In the above cases that framework was encoded in the I CD. In the following
section we turn to another framework - that of the DSM. Whatever the frame­
work, however, it is essential to underline the fact that our original document -
the WHO World Health Statistics Annual - is akin to a rather large Chinese box.
Open the lid and we find other boxes within - boxes containing conceptual
frames, boxes that contain operational rules, and boxes that contain situated
organizational decisions. The WHO Annual in that respect serves as little more
than a wrapper. It provides an image of a unified and independent object
(document), whilst, in fact, hiding a vast machinery of manufacture. The WHO
Annual is not alone in this respect and one essential (though rarely executed)
task of the social researcher bent on collecting 'facts about society' should,
therefore, involve removing the dust jackets of the documentary material that
he or she encounters, and to ask two very simple questions. How exactly, and
by whom, was this document assembled?

Enumerating neurotics

I am looking at a table of research results (Table 2 . 1 ) . It tells us about the com­

munity prevalence rates in Great Britain of certain types of what are sometimes
called 'neurotic' disorder, together with some estimates for the 'functional psy­
choses'. The table represents 'facts' about mental illness, and in line with what
we have noted above, it can be used, together with others related to it, as a

Copyrighted Material
44 USING DOCUMENTS I N SOCIAL RESEARCH

resource for researchers. Thus, we can, for example, refer to Table 2 . 1 as evidence
for our statement that about 1 6 per cent of people in any one week show
symptoms of a neurotic disorder. But how were this and other facts arrived at,
and how was the report put together?
As with the manufacture of crime or suicide or any other form of statistics,
the production of psychiatric statistics depends on the existence of a concep­
tual or theoretical scheme, combined with rules and technical instructions for
applying the concepts to a set of events and occurrences. As has already been
indicated, the conceptual scheme in terms of which mental illness is compre­
hended is that contained in The Diagnostic and Statistical Manual of the American
Psychiatric Association (the DSM-IV, 2000) . This contains a series of categories
relating to the various psychiatric maladies that people might be said to suffer
from, and it also contains diagnostic criteria for recognizing the distinct disor­
ders. Some of the disorders are listed in Table 2. 1 . We will return to the DSM
in a moment. For now let us focus on how the figures in Table 2 . 1 were
obtained.
The data in the table are derived from answers to a survey conducted in over
1 0,000 private households. The Appendix to the research report from which
these figures are derived provides the detail of the sample frame that was used
(Meltzer et al., 1 995) . It tells us the rules by means of which households and
the adults within them were selected. It also provides the questionnaire or
instrument by which mental illness was recognized. In this particular case the
instrument was called the Clinical Interview Schedule (Revised) or CIS-R.
The CIS-R is one of a variety of'instruments' that produce clinical and other
phenomena (see, for example, Bowling, 1 997) . In many respects documents
such as the CIS-R are like machine tools - tools for producing 'things' . Indeed,
phenomena such as 'disability', types of psychiatric illness and ' quality of life'
are conditions routinely manufactured by instruments of the kind referred to
here. In the case of the CIS-R the tool operates through a system of questions
and answers. For example, there are questions about people feeling fatigued or
feeling ill. One such question asks, 'During the past month, have you felt that
you've been lacking in energy?' Another question asks, 'have you had any sort
of ache or pain in the past month?' Respondents are required to answer 'Yes'
or 'No'.
In his Method and Measurement in Sociology, Cicourel (1 964) discussed the
status of questions such as these. As one might expect on the basis of what we
have already stated, Cicourel raised issues relating to the ways in which the
interview process, and the questions and answers provided within the inter­
view, are socially embedded. In particular, he became interested in the process
through which the interview - as a social event - can turn conversation into
social scientific data, pointing out that the use of instruments (such as, say, the
CIS) involved an act of measurement by fiat. That is to say, the instrument
imposes a commonality of meaning on questions and answers that are, in all
likelihood, variously understood - at different times and by different people

Copyrighted Material
 PRODUCING FA C T S 45

(see Houtkoup-Steenstra, 2000) . Indeed, the whole matter of measurement in

the social sciences is something of a Pandora's box, raising as it does problems
concerning whether or not the CIS (or a similar instrument) is valid for
measuring such things as depression and anxiety. Rather than concentrate on
measurement issues at the level of the interviewing process, however, it would
be more useful at this point to turn to an examination of the computer-assisted
procedures by means of which psychiatric diagnoses were made. That is to say,
to focus on the rule-based procedures by means of which the research managers
moved from the 'Yes' /'No' responses on the interview schedule to the cate­
gories of'depressive episode' , 'panic attack' and 'obsessive-compulsive disorder'
contained in the tables of the report.
The CIS-R depends on the use of 'lay' interviewers. In other words, the
people who ask subjects questions about mood and behaviour are not trained
psychiatrists. Consequently, it is impossible for them to diagnose the respon­
dents. Instead, on the return of the completed interview schedules to the
research team a set of algorithms were put into play. Many of the algorithms
involved scoring responses. For example, the questionnaire was divided into
sections and points were awarded in each section. On the section relating to
anxiety, for example, respondents would score 1 if they had been 'generally
anxious or nervous or tense' for four or more days in the past seven days. They
would score another 1 if they felt tense, nervous or anxious for more than three
hours in total in any one of the past seven days - and so on. As we move through
the questionnaire we add the points, and if they total more than 1 2 then the
subj ect is assumed to display symptoms of a psychiatric disorder.Which disorder
it is, is dependent on which sections of the instrument the respondent scores
within. So, no mention of the diagnostic conditions is made in the interview
schedule and certainly at no point were subjects asked whether they suffered
from obsessive-compulsive neuroses or depression or whatever.
Human beings, of course, rarely pick up disorders singly and sequentially and
in an easily labelled fashion. More likely they suffer from many things at once -
they are both depressed and psychotic, say. They indulge in substance abuse and
suffer from anxiety. So given that many individuals suffer from multiple
pathologies at one and the same time, the CIS-R provides precedence rules
that enable multiple disorders to be placed in a hierarchical sequence. For
example, depressive episodes always rank above phobias. In this respect, the
report follows the system prescribed by the DSM.
It was by the posing and coding of questions and answers, then, that diag­
noses of psychiatric disorder were arrived at. In fact, there is a sense in which
one could argue that the data assembling process - using interview schedule
and algorithms - manufactured the disorders. So what does this suggest about
our concept of psychiatric disorder? And what would happen if the research
managers had set the cut-off point on the schedule to 1 0 or 1 8 rather than 1 2?
Such questions are in many ways related. It is possible, for example, to select a
different cut-off point. Moving the point to, say, 1 0 would increase the prevalence

Copyrighted Material
46 USING DOC U M E NTS IN S O C I A L R E S EARCH

of mental illness in the community. Moving the point to 1 8 would decrease it.
So we can have as much or as little mental illness in the community as we
want. (It was once said of Poland that it was a country on wheels, seeing how
its borders were changed so frequently, and one might be inclined to take the
same view about the prevalence of psychiatric disorders.) There are, of course,
conventions about where the point should be, but the fact that the point is
movable tells us something about a particular type of discourse on mental ill­
ness in the late twentieth century. It is a discourse that argues that mental illness
is not something that is a qualitatively different category from sanity, but some­
thing of the same order, but which differs only in degree. The implication is,
then, that states of health and illness can be arranged along a continuum - a
continuum that runs from zero to infinity. Whether or not individuals are to be
deemed 'ill ' depends not on what they think, feel or do, but on the cut-off point
that we use for our classification. The level at which the cut-off is set is impor­
tant, even in research. terms, mainly because the impact of such things as genetic
or social factors on psychiatric disorder can be amplified or even 'eradicated' by
moving the point upward or downward. (See, for example, Brown, 1 98 1 .)
This vision of psychiatric disorder as a quantitative variation on normal behav­
iour expresses only one of a number of possible positions on the subject. It was,
for example, a vision that used to be contained within the DSM. However, the
contents and the theoretical ideas behind the DSM have changed markedly
between the appearance of the 1 st ( 1 952) and the 4th (1 994) edition. And one
consequence of the alteration of the conceptual scheme is that the conditions
that we are referring to in the 1 990s are not the same conditions as were referred
to in the 1 952 edition. For example, the word 'depression' occurs in both, but the
nature, course and origin of that depression have radically altered (Healy, 1 997).
Naturally, our table of statistics (Table 2 . 1 ) would not highlight this change, but
the changes are nevertheless embodied within it. (A table showing trends in the
prevalence of 'depression' over the later half of the twentieth century would,
however, be affected, in a fundamental way, by these alterations.)
By examining the history of the DSM, then, we can see how it is produced -
as with all forms of 'expert' documentation - in a politically structured space.
In fact, the DSM is a document that has been produced by a professional or
expert faction. Given the significance of the American Psychiatric Association
in the global network of expertise that faction has the power to decide what
is and what is not a psychiatric disorder and how that disorder is to be defined.
Above all, the DSM is (like the ICD) a machine tool - a tool that, when assem­
bled with others of the same ilk, can generate new products. Such generative
documents set out the boundaries in terms of which experts think and talk and
write. In the manufacture of data about psychiatric disorder, they are not, of
course, the only machine tools at hand. The CIS-R, the coding frame, the algo­
rithms to which we have referred, also serve in the workshop. Put together
such instruments produce, and how the production process unfolds is forever a
matter of legitimate social research.

Copyrighted Material
 PRODUCING FA C T S 47

Conclusions

Structures of literality (de Certeau, 1 984) are produced in a political and social
space. In this chapter we have made reference to some very different forms of
literality - the statistical report, the death certificate, the interview schedule, the
expert manual and the secret police file. How these various forms of literality
emerged and the social, political and economic contexts of their emergence are
themselves legitimate areas for research - though areas that are, perhaps, often
more appropriate to the historian than to the social scientist.
Our tack has led us to concentrate upon the enrolment (Pinch and Bijker,
1 989) and mobilization of some key generative documents in specific organi­
zational contexts. Generative documents, as I have stated, constitute the machine
tools by means of which other documents are produced. In particular they are
central to the manufacture ofsocial scientific data. The manner of their recruitment
and deployment for the creation of social facts is therefore crucial. During
recent decades a large body of work has developed on the role of documenta­
tion as tools in laboratory work and so it is clear that we need not restrict our­
selves to the manufacture of social scientific data in this respect. Thus Fujimura
( 1 996) , for example, has indicated how key laboratory manuals come to figure
as important tools in work in genetics, whilst Bowker and Star (1 999) have
illustrated how the ICD serves as a tool in medical work.
Some years ago Hindess ( 1 973) pointed out that statistical reports were always
produced in and through the operationalization of technical and conceptual
frames. In this chapter we have indicated the presence of both conceptual and
technical procedures in the manufacture of two types of report - a report on the
health of nations and a report on the results of a social survey. Our analysis could,
in theory, be extended to reports on economic affairs, crime statistics, quality of
life statistics, and any other genre of statistical summary that one might care to
mention. In that respect, the lessons of this chapter are uncomfortable for those
who routinely use statistical reports as sources of factual data. Indeed, had we
extended our analysis, we would not only have accumulated yet more examples
of how generative documents, forms of literality and social action intertwine to
manufacture 'facts' abou't the world, but also have seen how knowledge and
power interact. For the arrangement of knowledge involves, above all , the oper­
ation of power. Knowledge/power (the term is derived from Foucault, 1 991)
defines how things are to be arranged, and what is to be included and excluded
in the realm of what is known and what is knowable.3
Damton (1 984), with whom we opened this chapter, provides an excellent
illustration of the operation of power/knowledge in his consideration of the
Encyclopedie. The latter, a 1 7-volume work produced, among others, by Diderot in ·
the dawn of the age of enlightenment, contains on the face of it litde more than
a series of entries (listed from A to Z) on commonplace concepts, facts and ideas.
Treated as an anthropologically strange treatise, however, Damton sees within it a
world-view, an image of reality. Diderot, of course, was one of d'Hemery's

Copyrighted Material
48 USING D O C U M E NTS I N S O C I A L R E S EARCH

suspects, and unable to impose his vision of the world on anyone in particular.
Nevertheless, as Darnton points out, the EncyclopMie has been referred to as
'machine de guerre', a key item in the armoury of those who opposed the old
regime to its intellectual foundations. By the side of the EncyclopMie our examples
are, perhaps, puny. Nevertheless we can see within them similar acts of classifica­
tion and of inclusion and exclusion. It demonstrates that in every arrangement -
no matter how puny - there is a world-view to be studied and analysed.
Power/knowledge is not only contained and expressed within documents,
of course, but also activated in practice - by interviewers, coders, research man­
agers, 'auteurs'.With that in mind, I shall conclude by listing some simple ques­
tions that ought to be kept in mind when reports that enumerate (and even
those that do not) are scrutinized:

• What generative documents are called upon for the manufacture of the report?
(Be sure to examine the origin, design, conceptual architecture and modes of
instrumentation of any that are implicated.)
• Who (as Monsieur d'HI�mery would ask) are the auteurs reporting and how
does the author-function operate in the document's creation?
• What rules - of selection, coding and precedence - are used in the manufac­
ture of facts?
• How are the rules applied in practice? (The answer to this question will, of
course, demand a study of situated actions.)

Dismantling documents is not an easy task, but it is a worthwhile one, not least
because every document is packed tight with assumptions and concepts and
ideas that reflect on the agents who produced the document, and its intended
recipients, as much as upon the people and events reported upon. For what is
counted and how it is counted are expressive of specific and distinctive ways of
thinking, acting and organizing. In that sense, all documents serve as a two-way
mirror on aspects of human culture. That is precisely why Monsieur d'Hemery's
police files are every bit as instructive as the 1 7-volume EncyclopMie that was
produced by his suspects.

RESEARCH EXERCISE

Exercise 2.1

Use the questions listed i m m ediately above to deconstruct - as far as

you can - a set of official statistics on cri me , poverty or disability.
Some useful pointers to U K statistical data in these three areas (and
others) are provided i n Levitas and Guy ( 1996). For a next to compre­
hensive l isting of worldwide sources of official statistical data visit:
http://www2.auckland.ac.nz/lbr/stats/offstats/OFFSTATSmain.htm

Copyrighted Material
 PRODUCING FA C T S 49

Notes

In that sense 'reasons' for action, rather like 'motives', are not secret inner states of
private individuals, but rather drawn from culturally sculpted vocabularies (Gerth
and Mills, 1 953). That is, vocabularies that individuals can call upon and use so as
to give satisfactory and plausible accounts to interrogative others. It is, of course,
such processes that the ethnomethodologists regarded as forming the focus for their.
investigations drawing attention to processes such as 'ad hoeing', and the reflexivity
of accounts (see Garfinkel, 1 967, in particular) .
2 Bulmer (1 980), for example, has argued that the criticisms levelled against official
statistics apply only to a narrow range of suicide and crime statistics. That, of course,
is fundamentally to misunderstand what the constructionist argument - with
respect to statistics - is all about.
3 A much cruder, Marxist, argument about the role of power in the construction of
official statistics is provided by Miles and Irvine (1 979) . See also papers included in
the collection edited by Levitas and Guy ( 1 996), as well as the work of Coleman
and Moynihan (1 996) on crime data.

Copyrighted Material
3
Documents i n Action I . Docume nts
i n O rga n izatio n a l Setti ngs

An ethnomethodologist in the archive 50

A Thai village circa 1968 52
Talk and text in the clinic 54
Documents and organizational activity 60
The politics of fog 63
Conclusion: how to do things with documents 67
Notes 69

An ethnomethodologist in the archive

In the course of developing a research programme concerning the selection of

patients for treatment in a Los Angeles psychiatric clinic, the aforementioned
Garfinkel ( 1 967) and his colleagues sought to extract data from a set of clinic
folders. The latter were designed so as to contain information on key features
of the clients and their interactions with the clinic. So, for example, any one
folder might contain data on such things as the age, sex, religion, place of birth,
income and other features of the client, as well as the names of the clinic per­
sonnel with whom the client came in contact. Space was also allocated for the
recording of psychiatric diagnosis, previous psychiatric 'experience' and so on.
On examining folder contents, however, Garfinkel noted that many items of
routine data that should have been contained in the folders were missing. Thus,
age was absent from the folders in 5 . 5 per cent of cases, occupation was miss­
ing in over half of the folders, and place of birth was missing entirely. In a similar
way, reasons for the non-acceptance of patients was missing in 20 per cent
of the folders, whilst the names of the staff members in charge of the intake
conference were missing in just over 50 per cent of cases. Clearly these were
'bad' records, and Garfinkel turned to asking why such incomplete records
existed.
One part of the answer to his puzzlement related to what Garfinkel called
'normal, natural troubles' . For example, filling in questionnaires is a time­
consuming business for clinic staff. It is therefore not so surprising perhaps that

Copyrighted Material
 DOCUMENTS I N ACT I O N 51

such staff avoided completing those parts of the questionnaire that they
considered too costly in terms of time and effort, or that they considered irrel­
evant to the everyday work of the clinic. Such strategies are fully understand­
able to anyone who has to complete such routine tasks. Yet, these 'normal
natural troubles' constitute only one part of Garfinkel's story. The other part
concerns reference to what he called contractual issues within the clinical
encounter. And it is this second line of analysis that is of concern to us here.
By making reference to the fact that the folder contents implied contractual
relationships, Garfinkel was seeking to highlight how the folders were routinely
constructed with other than merely actuarial purposes in mind. In fact, sug­
gested Garfinkel, such documents were just as readily being assembled to
'show' or hint at what had happened in the clinical encounter and what might
happen to the patient. For example, clinic personnel assembled the folders
aware of the possibility that the detail contained within them might be called
on at some future stage to demonstrate that patients had always got the treat­
ment they deserved. So clinic folders were, if you like, being constructed in a
medico-legal framework such that it could always be shown that the 'right'
things were done to the 'right' person at the 'right' time. (Though exactly how
the records were to be read, and how the detail within was to be interpreted,
were always dependent on a reader's purpose at hand. In that sense the mean­
ing of the records altered according to circumstance.) Such a contractual reading
of folder contents explained why it was that basic items of data could be miss­
ing from the files on the one hand, whilst marginal notes and corrections and
additions to the folder contents could appear on the other. In short, it
accounted for why such bad records were, nevertheless, assiduously kept. (For
a study of similar themes with respect to the use of medical records in anaes­
thesia, see Harper et al. , 1 997) .
The title of Garfinkel's essay was ' "Good" organizational reasons for "bad"
clinical records' . It successfully demonstrated how records originally designed
for one set of purposes could be routinely used for quite different ends. In the
Los Angeles case, actuarial purposes were supplemented with medico-legal
considerations, but that in itself is of relatively little interest to us. More impor­
tant is the observation that one very important dimension of any document
rests in the manner of its use. What the document 'is', is specified by the way
in which it is integrated into routine activity. So one key lesson that we can
draw from the Garfinkel example is that people who use documents in their
research schemes need to look at how documents are picked up and manipu­
lated in situ, and not simply to focus on matters of content. Indeed, when one
does that, one can begin to see how documents can function to mediate social
relationships.
Garfinkel ( 1 967) and his colleagues often called upon the concept of reflexi­
vity to emphasize aspects of use. The notion of reflexivity implies that words
or texts not merely represent some aspect of the world, but that they are also
involved in making that world. In part, they constitute the world. With that

Copyrighted Material
52 USING DOCU M ENTS IN S O C I A L R E S EARCH

point in mind, I intend, in this chapter, to look at a variety of ways in which

documents are used so as to structure social relationships and social identities
in organizational settings - how they enter into what we might call the per­
formance of social life. In doing so it will become clear that in contrast to the
usual manner of approaching documents in social research, we will have to
subordinate a consideration of document content and focus instead on an
anthropology of use.

A Thai village circa 1968

In 1 968 - one year after Garfinkel published his essay on the Los Angeles clinic - the
anthropologist SJ. Tarnbiah published an essay concerned with literacy in a Thai
Buddhist village.The world that Tarnbiah describes is by this stage, no doubt, a lost
and distant world. (We need also to keep in mind the possibility that the world as
represented was Tarnbiah's world more than the world of the inhabitants that he
sought to describe.) Despite that, the insights thatTarnbiah (1 968) offers us remain
instructive - especially in so far as he demonstrates how documents and the script
that they contain can serve to structure the settings of everyday interaction and
help constitute the social relationships in which they are embedded.
Documents, as we have already noted in Chapter 1 , do not have to contain
script. However, in the Thai case they do. What is more, and as Tambiah points
out, there were at least three separate forms of script in use in the rural villages
of north-east Thailand during the 1 960s. Sacred or ritual literature was written
in Tham script (and so were some traditional medical texts) , whilst Lao script
was used for secular purposes, along with the modern form of literacy embodied
in Thai script.
The different scripts were associated, in Tambiah's day at least, with different
social roles and different social activities. For example, those people that we
might regard as the traditionally learned used documents written in Tham
script. (They also used other forms of script, but seemingly regarded Tham as
the most prestigious.) Consequently, the acham wat (the lay leader of the
Buddhist congregation), the mau khwan (or officiant at religious rites), the mau
ya (physician) and the mau du (the astrologer) used all three forms of scripts
(with, apparently, varying degrees of competence) . The mau lum singers of
-

the folk opera - had also to be able to read texts in all scripts, mainly so as to
gain access to the traditional songs and story lines of the opera.
In village life, of course, there are always varying layers of activity and not all
forms of it necessarily entail the existence of written discourse. So, for example,
knowledge relating to the cult of the spirits apparently took an oral form and
did not require literate participants. This latter type of activity was, Tambiah
implies, of lesser status. And in between these two poles were literate people
who worked almost entirely in forms of modern script, the most important
example being that of the village schoolteacher.

Copyrighted Material
 DOC U M ENTS I N ACT I O N I 53

As well as marking out social roles and activities, script was more directly
used to encode particular types of knowledge. Thus, religious and traditional
documents (sacred texts) were, naturally enough, in Tham script. They were
deposited in the village temple along with other texts containing stories about
the life of the Buddha, texts on the discipline of monks, the nature of sermons
and so on. Access to such texts and to the script that encoded them was via the
temple - Buddhist traditional script was taught within the temple. Govern­
ment documents, on the other hand, were written in Thai script, and would
have required of their readers a knowledge and use of such script - the script
of the village headman and of the schoolteacher.
So what does this resume of the Tambiah essay suggest? First, it highlights
how script can be used to mark out social roles and associated forms of
power/knowledge (Foucault, 1 99 1 ) . In rural Thailand, it seems, the various
forms of script marked out the roles of monk, teacher and traditional
singer/actor. And whilst there is no direct parallel to such a trio in contempo­
rary western society, there nevertheless remains an association between forms
of specialized script and specific activities. The professional (classical) musician,
for example, needs to be able to read modern musical notation, whilst the pro­
fessional mathematician and engineer need to read and use modern mathe­
matical and other forms of scientific notation. Access to the specialized scripts
provides access to specialized knowledge and the use of such scripts in practice
marks out the specialized roles. Similar considerations no doubt arise in rela­
tion to the manipulation and use of modern forms of script such as those, for
example, involved in computer programming languages. How these various
scripts are accessed and disseminated is clearly related to patterns of social
structuration (Giddens, 1 984), and throughout this chapter we shall follow
through on yet other examples of how a specialized script can enter into, and
structure, social interchange. Secondly, it is evident from what Tambiah tells us
that the actual use of a text can, in itself, serve to constitute social events and
relationships. Thus, the use of a prayer book in Tham script can be used to mark
out a 'sacred' - rather than a profane - moment, whilst the use of a form in
Thai script may be used to mark out a moment when official (state) business
is being conducted. The script serves to constitute the scene and mediates the
interactions within it.
One final point: we may note from Tambiah's description that knowledge
derived from script is often turned into an oral medium (translated) before
circulation. All text that is verbalized involves translation. In Tambiah's world
one very important medium of translation involved song, dance and chant (the
opera) . And the importance of song and performance in aspects of daily life is
not to be underestimated. I The significance of such matters, however, extends
way beyond the boundaries of this text, though the importance of translation
and of specialized script does not. In that vein, it will prove more interesting to
us to examine a translation process that moves in the opposite direction - from
talk to text. To that end I intend to zoom in a little closer on the use of

Copyrighted Material
54 USING DOCUM ENTS IN SOCIAL R E S EARCH

specialized script in organizational settings. The following examples are <\ll

drawn from studies of medical clinics.

Talk and text in the clinic

In his discussion of a psychiatric record, Hak ( 1 992) provides an example as to

how a professional psychiatrist translates items of patient talk and observed
behaviour into a written record. In so doing the psychiatrist - as note taker -
highlights the essential details of a patient's conversation, codes them into pro­
fessional language (of delusions, hallucinations, diagnostic terms, etc.) and
makes suggestions for future action (entry into a psychiatric unit or whatever) .
In Figure 3 . 1 we can see similar processes at work. Figure 3. 1 is a facsimile of
a page of nursing assessment notes that I came across in a psychiatric hospital
in the late 1 980s. It is clear from the notes that the members of the nursing
staff were concerned to categorize their patients in a variety of ways. The
latter included a one-word diagnosis of the patient's condition, an assessment
of'activities of daily living' (ADL) skills, brief notes concerning the level of co­
operation and hygiene exhibited by the patient and so forth. These assessments
were based on conversations and interchanges between nurses and patients and
among nurses alone. It is important to recall that the patients/clients - as with
all human beings - commonly indulged in a wide array of activities and behavi­
ours. For example, patients would talk to themselves, watch television, lend
each other cigarettes, shout, laugh, go to work in the hospital workshops and
so forth. Yet of such arrays of activity only a few are ever highlighted. Thus, in
the case of the first-named patient, it is the patient's 'schizophrenia' , poor ADL
skills, temper tantrums and quarrelsome behaviour that are highlighted. This
selectivity of focus would become even more evident were one to examine
other kinds of patient records. Thus, in the hospital to which I am currently
referring there were also psychiatric records (called 'charts'), and social work
records kept on each patient. The former were maintained by the medically
trained psychiatrists and contained other kinds of information, such as data on
whether the patients exhibited any 'first-rank symptoms' (of schizophrenia) ,
their medication and its effects, their 'history', items about family life, patient
delusions and so forth. In fact the psychiatric records looked very much like
those alluded to and reported upon by Hak ( 1 992) . Social work records were
also made up for each of the patients. These paid relatively little attention to
medical diagnoses and the effects of medication and referred more often to the
stability and maturity of the patient vis-a-vis relationships with others, the
nature and level of the patient's state benefits and the like. Considerable refer­
ence to the whereabouts, behaviours and opinions of other family members
was also made within social work files. Access to such records and the 'right' to
make entries in such records were more or less restricted to the members
of the individual professional groupings. In that respect the 'script' in each

Copyrighted Material
 Patient Admissions Diagnosis Problems/Constraints Medication Plans Assets
Name: No. of adm. = 1 Schizophrenia Long time in hospital Gavison tabs bd Maintain
Dob 1111 2/36 Date of last adm. Temper tantrums & bad Trifluoperazine 2 mg bd
Admin area: 1 7/06/1962 language Vit BPc 1 tab mane
East Activities of daily living poor
Dr: Yellow Recently quarrelsome with ASB
Ward: Blue
0
Name: No. of adm. = 5 Paranoid Loner. Poor motivation Piroxican 30 mg nocte Move to community ADL skills good 0
Dob 23/02/53 Date of last adm. Schizophrenia Wishes to stay in hospital Ranitidine 1 50 mg nocte (")
c
() North 29/09/1987 Failed RA (stay in residential Vit BPc am
�
0 Dr: Green accommodation) rn
"b
'<:: Ward: Blue Injury to right hip z
-I
�.
�
Name: No. of adm. = 3 Schizophrenia Poverty of thought Benxtropine 2 mg mane ADL activities Pleasant & co-operative en

m Dob 3/07/58 Date of last adm. Low lQ Withdrawn Thioridizine 75 mg tid Hygiene good
z
0.. West 1 7/05/1988 ADL skills limited Senna 2 tabs nocte
Dr: Green »
�
Childish and na"ive in manner
Ward: Blue (")
...... Mother has encouraged -I
CD dependence over the years
0
� but is now opting out
Poor road safety
z

Name: No. of adm. = 6 Dependent Multiple Somatic complaints Thyroxine 0.1 mg mane Maintain ADL good
Dob 10/08/60 Date of last adm. Personality Resistant to suggestions Nifedipine 10 mg bd
West 10/03/1988 Poor compliance tends to opt out Thioridizine 50 mg bd
Dr: Green Poor response to
antidepressant therapy
Ward: Blue Poor attendance at OT unit
Poor Hearing

FIGURE 3 . 1 Facsimile of a ward-based nursing assessment record (UK psychiatric hospital, 1989)
C1I
C1I
56 USING DOC U M E NTS IN S O C I A L R ES EARCH

document served, in part, to mark out the realm and expertise of the various
parties - in much the same way as did the various forms of script in Tambiah's
Thai village. Social work talk belonged in social work records,
psychiatric talk belonged in medical records, and nurse talk belonged in nursing
records.
Recorded observations on patients/clients are, then, highly selective. In the
case of public service agency files, such records often define the human beings
that they refer to in specific and particular ways. In so doing they call upon and
activate a whole series of what we referred to in Chapter 2 as membership cate­
gorization devices (Silverman, 1 998) . How a particular device comes to be
associated with any individual and how that categorization might be used and
called upon to account for and explain an individual's behaviour in specific
circumstances can form the occasion for important and fundamental socio­
logical research.
In my work on the psychiatric hospital referred to above, I was primarily
interested in how patients came to be classified in different ways through
routine procedures (see Prior, 1 993) . Naturally the use of notes and records
forms only one support for the identification system that surrounds patients.
Everyday conversation and casual interchanges form another, and in any
organization there will always be a constant interchange between talk and text.
Thus I provided in my 1 993 study a short extract of an exchange between
nurses In the ward office concerning the issue as to what was wrong with
'X' . Viz.

Nurse 1 : Does anyone know what's supposed t o b e wrong with X?

[ Blank looks and silence meet the question. ]
Nurse 2: Schizophre n i a , I su ppose .
Nurse 1: Hmm. I 've never seen any sign of it.
Nurse 3: Wel l , he's on chlorpromazin e , so he m ust be schizophrenic.

The MCD of 'schizophrenic' recorded in the notes is, then, sustained and
underlined in this case by means of a casual conversation - and especially the
reference to ex's' medication. (The use of medication to define a specific
psychiatric disorder, rather than the other way round, is not uncommon in
psychiatry.) If, however, there were any real doubt about 'what was wrong with
X', it would be the notes that would carry the day. So the researcher who
wishes to concentrate on the use of documents in action has to be constantly
aware as to how the written record is tied into and anchored within other
aspects of organizational life such as conversations at the nursing station.
Nevertheless, it is only when assessments are written down and can be pointed
to that they are used to form a foundation on which routine social actions are
built. Thus medical professionals can and do use 'the files' as a warrant for their
actions in relation to their patients, showing how what they do to patients is

Copyrighted Material
 DOCU M E NTS I N ACTION I 57

warranted by the information on the record(s) . Indeed, in the context of

psychiatry Barrett ( 1 996: 1 07) has underlined how ' clinical writing' not only
describes the treatment of patients, but also constitutes the treatment.
In a wider context, Bowker and Star (1 999) and Young ( 1 995) have also
pointed out how (documented) nomenclatures of disease are routinely tied
into the financial accounting mechanisms of hospital life. So, psychiatrists
routinely draw down a diagnostic category from the DSM, 'fit' it to a given
patient, and then justifY what was done for and to the patient/client in the light
of the category. (This, in the full knowledge that categories and patients rarely
make a 1 : 1 fit.) Indeed, the naming of diseases and disorders in modern medical
systems is often used more for purposes of financial reimbursement (to and
from insurance companies) , and other accounting and monitoring purposes,
than it is for fonning �ccurate descriptions of a given patient's condition.
Arguing along a similar path, Zerubavel (1 979: 45) had previously indicated
how notes written up by medical and nursing professionals were 'among the
main criteria used by their supervisors to evaluate their clinical competence',
as well as fonning the primary mechanism through which continuous super­
vision of patients was maintained. In fact Zerubavel's study highlights the
centrality of charts, graphs and records of all kinds in underpinning the routine
social organization of hospital life. Thus, printed schedules are used to organize
the patient/staff day; printouts of various kinds are routinely used to monitor
patients; and notes are written so as indicate how the 'hospital' cares for its
clients. (In American hospitals, of course, patient records are, as we have already
indicated, also used as a hook on which to hang financial costs and trans­
actions.) Above all perhaps, we see in the use of documents to monitor activity
an essential component of that self-reflexive capacity that is said by Bauman
( 1 9 9 1 ) to characterize modernity as a whole.
This capacity of medical records to mediate social relationships of all kinds
has been further researched by Berg (1 996; 1 997) who points out how hospi­
tal patients are both structured through records and accessed through records.
One important feature of patient existence that is emphasized by Berg is the
manner in which medical records are used so as to keep the case (and the
patient) 'on track'. Such structuring of patient trajectories through records is
achieved in numerous ways - planning and monitoring being two of them. In
this respect it is of interest to note the column headed 'Plans' in Figure 3 . 1 . In
the context of these records the most important plans concerned whether or
not the patient was ready for life in the community. (My own hospital study
was executed at that rather important cusp where psychiatric patients were
being moved out of hospitals and into 'the community'.) In most cases, the
patients were to be 'maintained' (kept in the hospital ward) . The detail hardly
concerns us here. What is important to note is that records of this type always
contain some rules for action. Other rules for action are contained in the
column relating to medication. Latour ( 1 987) has used the term 'action at a
distance' to indicate how decisions written down in one context and setting

Copyrighted Material
58 USING DOCU M E NTS IN SOCIAL RESEARCH

Part 2: Special Educational Needs.

( H ere set out the ch ild's special educatio n a l needs, in terms of the ch ild's learning
difficulties which call for special educational provision)

'Z' is a pleasant, likeable boy who has a severely delayed/disordered phonological

system and a mild delay in his expressive language. His speech is largely
unintelligible. Due to the severity of this delay he has difficulty in accessing the
curriculum. He was slow to adapt to the social routines of the school day. His
attention span is limited without adult support. He does show a high response in
one to one situations with a known adult. Interaction with his peers is improving.
Assessment of his non-verbal intellectual functioning is at the bottom end of the
high average range, however, verbal intelligence falls into the low average range.

This pupil's special educational needs have been identified as follows:

1 severe speech and language delay/disorder

2 poor social/interactive skills

The attached reports contain more detail about 'Z's ' specific attainments.

Part 3: Special Educational Provision.

Objectives. ( Here specify the objectives which the special educational provision
should aim to meet)

The objectives of the special educational provision to be made for this pupil
should be:

1 To improve his receptive and expressive language skills

2 To develop early literacy and numeracy skills
3 To improve social/interactive skills.

Educational provision to meet needs and objectives

FIGURE 3.2 Extracts from a 'Statement of Special Educational Needs ', Great
Brita i n , 1999

can carry implications for action in future settings. And it is indeed the case
that records often contain instructions for future organizational activity. On
that note I draw attention to Figure 3.2.
The figure contains extracts from a 'special educational needs assessment' of
a young boy. A psychologist executed the assessment. Note the categorization
devices used in the assessment - such as 'pleasant, likeable boy [with] a severely
delayed/disordered phonological system' - and how the concerns of the psycho­
logist structure the identity of the child. Note also that Part 3 of the assessment
contains instructions for future action. How such action was to be imple­
mented was the subject of a later section of the report that has not been repro­
duced here; nevertheless it is clear that the report is designed not simply to
categorize the child and his 'problems' but also so as to instigate future action­
at-a-distance. (Alternatively the written assessment was open to recruitment by
parents, teachers and others so as to demand more resources for the child even
in the event of the 'action' failing to take place.)
There is one final point that perhaps needs emphasis. It is clear that people
not only maintain records so as to legitimize action or to claim warrant for their

Copyrighted Material
 DOC U M ENTS I N ACT I O N I 59

actions, or, indeed, to classify events, objects and processes in the everyday
world, but also use records to sustain micro interactions. In this respect the
following extract is worth studying. It was obtained . in a fatigue clinic. The
physician (denoted by D) usually had difficulty in telling his patients (denoted
by P) that, despite extensive test procedures, no evident physical pathology
could be found to account for their fatigue. The alternative explanation was to
raise the possibility of their condition being 'psychological' (lines 70-76) - a
suggestion that few of the patients welcomed. In this extract the physician is
attempting to 'talk to his notes' and thereby use them as a key prop to his con­
sulting strategy. (Incidentally, the patient referred to below was seated in a
wheelchair at the time of the interchange.)

5 7 0 [ reading notes containing test results]: M R I scan tells us that you

haven't got MS . . . Blood tests,
fine . . . antibodies . . . fine/
60 P65: What is it then?
61 0: Wel l , we've checked out the
central nervous system and
there ' s noth ing u ntoward
there . M aybe we o ught to
check out the peripheral
system . [ Returns to his notes,
reading apparently to himself
whilst the patient talks to her
husband. ]

70 0: You were treated for depression

in . . . ?/
72 P65: The doctor [ i . e . the Primary
Care Practitioner], he just kept
on saying I
75 was depressed.
76 0: And were you?
77 P65: I ' m housebound, I can't get out
of bed i n the morni ngs . . . I
have to climb the
80 stairs on my bum . . .

I have explored the wider implications of these kinds of conversations in Banks

and Prior (200 1 ) . Here I am only interested in drawing attention to the use of
written documents as an ally in face-to-face interactions. In the above instance
the notes were used as a prop to dealing with a sensitive issue. In that vein it
might be useful to examine some further examples of documents in use, this
time in the context of homicide.

Copyrighted Material
60 USING DOCU M ENTS IN SOCIAL RESEARCH

Documents and organizational activity

In The Concept <if Mind (1 990), Gilbert Ryle uses an illustrative example of a
visitor to Oxford being shown around some college buildings and then asking,
'But where is the University?' The visitor, Ryle suggested, was making a cate­
gory mistake - confusing problems that can be posed in terms of 'where' ques­
tions with those that cannot.Yet, be that as it may, in common-sense usage, the
visitor's question still stands. 'Where', indeed, is the University?
One answer to such a question is that a university (any university) is in its docu­
ments rather than in its buildings. In the UK, for example, universities are estab­
lished by a charter. The charter - together with other documents - names the
university, provides warrant to award degrees, legitimizes the officers of the
university and so on. Naturally, a university has buildings and equipment and
lecturers and students, but none of those things are sufficient for the award of
university status. Only the charter can define the organization as a university, and
in that sense provide the one necessary condition for its existence. More impor­
tantly, it is documentation that invariably forms the basis for what Atkinson and
Coffey (1 997) refer to as 'documentary realities' - organizational features that are
created and sustained almost entirely in and through the documentation.
Universities are not alone in this respect. Indeed, in any organization it can be
quickly seen that it is documentation - rather than its artefacts or members -
that underpins the organizational presence. (Atkinson and Coffey ( 1 997) , for
example, cite a financial statement as a representation of an organizational exis­
tence.) On a more specific level we might take the post of university lecturer
(or accountant, or media relations officer, or research chair) . These posts only
exist in the written (documented) job descriptions that brought them into
existence. That is not to say that human beings always act in accordance with
job descriptions and fail to act in any other way than as described, or that work
can be reduced to job descriptions. On the contrary, we know that such things
are not possible. It is to say, however, that the warrant for the position and the
template against which human performance is to be measured are in the writing
rather than the doing. Posts, committees, and even organizational structures
themselves (such as departmental structures) exist and can be pointed to, only
in so far as they are documented. This is, of course, a point well made by
Dorothy Smith (1 984) in her paper on textually mediated social organization.
In the normal course of events documentation has a relatively low profile in
any organizational system. That is not to say that things, events and processes
are not documented, only that such documentation is regarded as routine and
thereby becomes invisible. At critical points, however (usually when things go
'wrong' , or when procedures are subject to an unusual degree of scrutiny or
monitoring) , documentation comes into its own (Zimmerman, 1 969) . Thus,
Bowker (1 994) points out, for example, how the written traces left by the
Schlumberger Company in its own archives changed radically in response to
the realization of possible forms of legal and political pressure.

Copyrighted Material
 DOCU M ENTS I N ACTI ON I 61

In this section I am going to concentrate on what we might call organiza­

tional failure. Such failure is usually only traceable and accountable through
the written record and therefore (from a research standpoint) crises in orga­
nizational life can provide considerable opportunities for investigation. Thus,
if someone claims to have a plan of action, then - when things go wrong -
they have to produce its written form in order to convince us that it existed.
If someone claims to have ' done the right thing in the circumstances'
then their case is only watertight if they can show that they acted in accor­
dance with some written protocol or other. If someone claims to have fore­
seen a disaster and alerted people to it, then their alarm should be evident in
the documentation. And if someone wishes to demonstrate that their work
and their proj ects were organized they must, once again, recruit some docu­
mentation to demonstrate systematization and orderliness. In all these
respects Dianne Vaughan's (1 996) study of the Challenger disaster provides
invaluable examples that require extensive study.2 In this section, however, I
am going to focus on disasters of another kind. They concern homicides in
the UK - homicides committed by people who are said to have psychiatric
problems. My main point of focus is on how the 'organization' of care only
comes into focus through documentation. In following through to that con­
clusion, however, we will be enticed to consider a series of other significant
Issues.
Let us establish at the start that people with psychiatric problems rarely
commit homicidal acts. Indeed, such individuals are far more likely to be a darIger
to themselves than to others. Nevertheless, for reasons that do not concern us
here, homicidal events involving people with psychiatric problems normally
receive a high media profile. In the UK such events are invariably marked by
the publication of a Mental Health Inquiry (MHI) Report. Such Reports first
appeared during the 1 980s. (By the end of the century there were just over
30 MHI Reports in publication.) It is to the nature of such reports that I now
wish to turn.
In 1 994 the National Health Service Executive (the body responsible for the
policy and planning of the UK health service) published a circular (NHS,
1 994) stating - among other things - what should be done 'If things go wrong'
in the organization of care for those with a serious psychiatric disorder. It stated
that in cases of homicide an MHI would always be required. (There is a sense
therefore in which the very presence of an MHI Report marks out a space -
a crisis area - where something has gone wrong.) And when things go wrong,
people always ask penetrating questions.
In most cases, MHI Reports open with a simple narrative of the events lead­
ing up to the murder or murders. The narrative links together things, people
and events that until that point had not been so linked. Thus, any and all inci­
dents in which the murderer had made verbal threats, or used violence, or had
been involved in unusual behaviour previous to any homicide are closely
recorded. For example, Sharon Campbell murdered her social worker in

Copyrighted Material
62 USING DOCUMENTS IN SOCIAL R E S EARCH

1 984. The previous year it was said that she had assaulted her vic"tim and made
threatening telephone calls to her. She had also been in possession of a knife in
her hospital ward. Christopher Clunis, who stabbed his victim to death in
1 992, had previously been involved in minor theft from a shop, abusive behav­
iour, sexually explicit behaviour and violence. The very recording of such
incidents in the MHI Report (Ritchie et al., 1 994) serves to implicate them in
the homicidal process. In fact, such events are usually recontextualized in the
Reports as 'warning signals' , and it is subsequently implied or suggested that
various professionals should have picked up the signal� and acted on them.
Indeed, among the many recommendations that the MHI Reports make, time
and again, are those that concern the fact that notable incidents be recorded
(see Reith, 1 998) . So Sharon Campbell's possession of a knife 'should have been
recorded' in her medical record. The various misdemeanours of Clunis should
also have been recorded - by nurses, social workers, the police - and so forth.
The absence of a record is in such contexts interpreted as a 'failure' . Indeed,
lack of documentation is commonly invoked to demonstrate lack of concern,
planning, foresight and organization.
Yet, it is also clear from the MHI Reports that recording, in itself, is usually
viewed as an insufficient demonstration of 'organized' care. Thus if we take the
case of S.A. Armstrong who raped and murdered 4-year-old Rosie Palmer
during the summer of 1 994, we see that whilst there were many notes, files and
dossiers on the man, none of them were linked. Indeed, Armstrong was only
ever considered a single, undifferentiated and identifiable case in retrospect -
that is, in the MHI Report (Freeman et al. , 1 996) . (The same, incidentally, was
true of Clunis as a 'case'.) So Armstrong, as a client and as a subject, was for­
mally described, accounted for and referred to in various registers - social
services registers, probation registers, registers concerning child sexual abuse,
police registers, in-patient and out-patient registers and so on.Yet, the exchange
of information on Armstrong as a single autonomous subject was limited to the
extent that no agency ever 'looked at the totality of the case' (1 996: 88) . Each
criminal and medical episode in his life was therefore investigated and dealt
with on an isolated basis ( 1 996: 88) , and there was little communication of
information between agencies. In other words, the organization of care was not
demonstrable in the documentation.
In line with this fragmentation of records, Armstrong had numerous histo­
ries - written up by different people working in different agencies. The MIH
Report's authors in fact criticized the paucity of history taking in both nursing
and social work. They were especially critical of the manner in which hospital
nurses had described Armstrong in terms of a specific ('Roper-Logan- Tierney')
model that had been designed for use in the hospital as a whole and which was
unsuitable for use with psychiatric patients. In other words, nurses had used the
wrong template when they recorded his detail. Somewhat interestingly, nursing
staffhad variously described Armstrong as '(1) pleasant and sociable . . . (9) tall with
glasses . . . (12) wearing Cuban heels . . . ( 1 3) slimy ( 1 4) a story teller' ( 1 996: 74) .

Copyrighted Material
 DOCUMENTS I N ACT I O N I 63

Not being in receipt of psychotropic medication, however, Armstrong was not

truly considered as a case of mental illness.
So the organization of community care - the 'seamless web' that is supposed
to surround psychiatric clients placed in the community - is j udged, at times
of crisis, not so much in terms of what goes on in a given location or institu­
tion, nor by street-level activity, but more in terms of written plans, registers
and monitoring procedures. Indeed, the actual organization of care can only be
made evident in plans, registers and written protocols. In the absence of rele­
vant documents a systematic and planned approach to professional practice
cannot be demonstrated. Indeed, we might say that 'care' is something that is,
essentially, 'performed' (Goffinan, 1 959; Mol, 1 999; 2000) . One important
component of such performance is in the documentation - for the documents
specifY how the performance is to be executed. Moreover, unlike physical
interactions, documents can extend across time and place and can be pointed
to and recruited as evidence in a way that transitory actions and talk ·cannot.
That is, perhaps, why 'performance indicators' usually take a documentary
form. For action turned into documentary outputs (figures, reports and so
forth) can be evaluated long after the physical performances, to which they
supposedly refer, have passed.
One final point; it is a point often made in the MHI Reports themselves:
namely, that the only time that the lives of the damaged individuals of whom
we speak is reported upon in a holistic manner is in an MHI Report. Outside
of the Reports such lives and identities remain partial and fragmented. We shall
follow through on the relationships between documents and identity in
Chapter 5. For now, however, I want to provide just one more example of the
ways in which documents can be used to constitute an event. On this occasion
the event is an environmental one rather than a personal one. Our focus, as
ever, is on how the document functions in the network of concerns that we
are about to unravel.

The politics of fog

In December 1 952 an anti-cyclone came to rest over the town of London.

A combination of warm air from the Gulf Stream and the slow-moving anti­
cyclone produced a temperature inversion over the city. The warm air trapped
beneath the cloud cover and the cold air above it were ideal conditions for fog.
Indeed, it was and is referred to as the Great Fog of 1952. The fog lasted from
about 4 December until 8 December. Visibility was reduced astoundingly -
theatregoers at Sadler's Wells, for example, were unable to see the stage on
which La Traviata was being performed. The performances were cancelled.
Londoners had difficulty in breathing and the number of people with respira­
tory problems admitted to hospital increased dramatically over the December
period. Indeed, an interim report on the fog suggested that some 1 2,000

Copyrighted Material
64 USING DOCUMENTS I N S O C I A L R E S EARCH

people died as a result of it.3 Fog, it would seem, has such consequences. The
official report on these events, published in 1 954 - Ministry of Health Report
on Mortality and Morbidity during the London Fog oj December 1 952 - later
reduced the estimate of the number of casualties to 8,000, simply by using a
cut-off point for deaths of 20 December. As we have seen in Chapter 2, quan­
titative analysts are allowed such tricks. Yet, the deviousness with numbers was,
perhaps, the least disturbing thing about these events.
A few years earlier in Donora, Pennsylvania, a similar set of circumstances
had arisen.4 The Donora Smog disaster of October 1 948 had claimed 18 to 20
lives. Some 7,000 citizens were hospitalized. The root of such disaster lay in the
presence of toxic levels of sulphur dioxide in the air. The poisonous gas had,
like the London fog, been trapped in a temperature inversion, on this occasion
an inversion in the Monongehela Valley. Unlike the London fog, however,
many individuals had an idea that the smog (a word denoting smoke mixed
with fog) was far from 'natural ' . For, on at least three previous occasions, the
American Steel and Wire Company - which owned a zinc plant in the valley -
had been cited in damage suits filed in the State Supreme Court for generat­
ing noxious fumes 'wilfully, wantonly and maliciously'. In fact, the 1 949 state
investigation report on the disaster argued that the events were caused neither
by smog nor by fog, but by 'air pollution' during unusual weather conditions.
Fog is a natural occurrence. Under specific conditions it occurs in valleys and
coastal sites. Air pollution, on the other hand, is manufactured. It is, if you like,
person-made. The description of the London pollution as 'fog' was, of course,
extremely useful to the government of the day. Fog was natural. Nothing could
be done to prevent it. (The government had, however, issued flimsy face-masks
to the population in order to be seen to be 'doing' something.) But air
pollution - the emphasis is on 'pollution' here - demands action. The Donora
disaster led to the adoption of state and federal laws to control levels of air
pollution. The London disaster resulted in few immediate changes.
Fog was known to be endemic in London. Towards the end of the
nineteenth century there had occurred a series of particularly dense fogs. The
twentieth century, however, had been relatively free from fog. Apart from
Donora, the most renowned case of air pollution had occurred in the Meuse
Valley (Belgium) rather than London. The base cause of such fogs lay in the
burning of fuels - particularly coal and petrol - and the consequent pollution
of the atmosphere with sulphur dioxide and other contaminants. In calm winter
weather of the type that affected London in December 1 952, the pollutants
became trapped below the cloud level and thickened up with each passing day.
The 1 952 fog was probably exacerbated by the fact that a number of very large
coal-burning power stations had been recently opened in and around London,
and together with the smoke from other industrial, domestic and traffic out­
puts the density of pollution was increased. (How and why the technological
style of the London power system required the construction and operation
of a large number of coal-fired power stations is of considerable interest in

Copyrighted Material
 DOC U M E NTS I N ACTI ON I 65

itself - see Hughes, 1 989.) The official report on the fog, however, side-stepped
these issues entirely. 'It is not within the scope of this report', stated the anony­
mous authors, 'to draw attention to measures necessary for the prevention of
atmospheric pollution' (Ministry of Health, 1 954: 3).
Instead, what the report concentrated on was deaths and the causes of death
during the month of December. There had clearly been an excess of deaths
over the average for the previous five years - though the size of that excess, as
. we have seen, depends on the cut-off point that is used to measure it. People,
of course, die naturally. Indeed, claimed the anonymous authors of the 1 954
report, the pattern of deaths from respiratory diseases was not all that different
in 1 952 from what it had been in previous years, but 'the fog hastened the
deaths of a number of patients' ( 1 954: 1 8) . Fog was a 'precipitating agent' on a
susceptible group of people whose life-expectancy must, in any case, 'have
been short' ( 1 954: 1 ) . So there were 'no deaths attributable to fog among those
who were previously healthy' ( 1 954: 38), and there was no evidence in anyone
of a 'smog lesion' ( 1 954: 25) . Finally, the committee responsible for the report
concluded that in the present state of knowledge it was impossible to state that
any one pollutant was the cause of death.
In restricting itself to a report on the causes of death, the conunittee thereby
confined itself to a focus on the anatomical sites of death. For, as we have seen
in Chapter 2, modern western culture understands death solely in terms of
anatomical sites and named diseases. In this respect the reference to a 'smog
lesion' is particularly instructive. The interpretation of disease as involving the
production of a lesion (a visible pathology in the body) has its origins in the
development of Paris medicine during the early nineteenth century. Foucault
( 1 973) has already described the significance of Paris medicine for western
understandings of disease and death, and there is no need for us to follow his
path from this particular junction. We need only add that during the very last
quarter of the nineteenth century a further theory of disease (let us call it the
'germ' theory) indicated how lesions could be caused by microbes. Henceforth,
there was a belief that for every lesion there was one and only one causative
agent. The tuberculosis (TB) lesions were produced by inhalation of the (TB)
bacillus. The spiral-shaped bacterium that travelled through the genito-urinary
system caused the lesions of the syphilitic - and so forth. So if 'smog' were a
cause of disease then, by implication, it too should have a lesion - but there was
none. The lesion that was evident upon post-mortem was similar to that
caused by the pneumococcus. That is to say, the dead had died from the pneu­
monias and especially bronchopneumonia. In the medical frame of the age,
then, the deaths were caused by things that had no relation to smog or fog.
If the fog were significant, then it was significant only as some exacerbating
agent - it affected the rate of death not the causes of death.
A similar kind of expert argument was widely used by other authorities
during the twentieth century. Cigarette smokers, for example, fell victim to
heart disease, lung cancer, respiratory conditions and other things that people

Copyrighted Material
66 USING DOCUMENTS IN SOCIAL R E S EARCH

died of 'naturally', so it would be foolish to claim that cigarettes caused death.

Coal miners and cotton workers as well as those who worked in asbestos
factories died similarly. So it would be equally mistaken to blame their deaths
on coal dust, or cotton dust, or asbestos dust. A focus on a lesion and its specific
cause is clearly a useful mechanism for distributing and redistributing responsi­
bility. Robert Proctor (1 995) demonstrates just how useful a mechanism it has
been in obscuring our knowledge about carcinogens in the modern world. This
is not to suggest sleight of hand on behalf of all authorities that argued in this
manner. Those who affirmed the medical frame of the age were unable to inte­
grate things such as cigarette smoke and smog into their causal schemes. People
died of diseases and diseases were caused by specific agents - mostly microbes -
whose presence was visible in the (dead) human frame. Neither smog nor any
of its component elements were known to cause a 'smog disease' .
We can begin t o see, then, how the report as a document enters as a n actor
into a network of actions. It serves as a repository of ,expert' knowledge on a
problem. In fact, it defines what the problem is (people are dying of respiratory
diseases) . It collects evidence in accordance with its definition of that problem.
Within its realm of expertise, it hints at possible solutions (people with respi­
ratory diseases should stay indoors during fogs) . It functions to provide ade­
quate, expert, accounts and explanations of what has happened (it narrates the
biography of a disaster from within a specialist, technical frame) . It segments
the world into that of which one can speak rationally, and that which lies
beyond the realm of expertise (issues relating to atmospheric pollution) . It
maps out what, if any, political actions may have to flow from the findings. The
fact that the authors of the report chose to remain absolutely silent on what
many might regard as the single most important thing that was relevant to the
events (namely, atmospheric pollution), clearly meant that no decisions had to
be made or taken on measures to control air quality. Power, and other indus­
trial facilities could continue to belch out smoke, vehicle emissions required no
control, and householders could continue to stoke their coal-burning fires. The
expert report could, when required, be appropriately recruited by active
parties to support a laissez-faire attitude towards pollution control. The experts
had spoken and had failed to identify pollution as of any primary significance.
As we have seen with the MHI Reports, the report on the London fog may
be analysed in terms of its content or of its use. These are not mutually exclu­
sive options, but it is truly the active rather than the inert document that should
form the field for the social scientific investigator. In that frame what we need
to do is to look at how the text is used by social actors in the course of their
everyday activities, and how the text itself can become an agent in the various
social networks in which it becomes embedded. Who recruits the report as an
ally, and who is arraigned against it? What is enrolled within the report and
what is excluded? And how does the text (report) itself become an agent in a
network of action?

Copyrighted Material
 DOCUMENTS I N ACTION 67

Conclusion: how to do things with documents

In his justly renowned How to do things with words, J.L. Austin (1 962) refers to
performatives. The sentence 'I promise' is an example of a performative. For, in
speaking the sentence the speaker also does something (acts) . It is an idea that
can be usefully transferred to the study of documents, which also 'do' things by
the very condition of their existence. For example, the common or garden
office memo not only carries (mundane) information, but also gives expression
to a set of (power) relations within an organizational setting. Indeed, memos
give concrete expression to systems of hierarchy - of superordination and
subordination - and certainly serve to define social networks. Further, by con­
taining demands and instructions, the memo becomes engaged in a system of
action-at-a-distance. (The memo written in one co-ordinate of time and place
triggers activity in another.) The same might be said of summonses, invoices
and contracts of all kinds (see, for example, Hughes and Griffiths, 1 999) . Indeed
texts routinely have such 'structuring' effects (Smith, 1 984) and are central to
the patterning and organizing of everyday activities. So the manner in which
documents circulate and are accessed serves to mark off social groupings and
organizational positions.
The content of memos, registers, contracts, blackboards and other 'inscrip­
tion devices' (Latour, 1 987) are, then, of considerable significance. Yet, a focus
on content to the exclusion of the manner in which a document is used could
easily lead the social scientific researcher astray. Thus, a study of legislation
(such as was embodied in, say, the Constitution of the old Soviet Union) and the
history and manner of its production would provide a thin picture of the legal
process unless we also studied how such legislation was used in action. That is,
how it was drawn down, referred to, interpreted, recruited by opposing parties
in the courts and so forth. And the same might be said of any legal system.
Documents as inert matter offer a very different field of study from documents
as agents.
It is for such reasons that I have attempted, in this chapter, to focus on docu­
ments in organizational settings. I have focused, in particular, on how docu­
ments can both mediate and structure episodes of social interaction. Above and
beyond that, I have also shown how documents can be recruited into alliances
of interests so as to develop and underpin particular visions of the world and
the things and events within that world. Thus we have examined, for example,
how the.identities of 'patients', 'clients' and the criminally insane are structured
through documentation; how forms of documentation can be used as warrants
for action, or as props in interaction; and how 'organization' is made evident -
'performed' - through the written record. Finally we have seen how an inves­
tigative report can both structure a view of events and be enrolled so as to
justify political and administrative (in)action with respect to environmental
problems. In that light, it is very clear that, ultimately, documents can damage

Copyrighted Material
68 USING DOC U M E NTS I N SOC IAL R E SEARCH

your health! If there is one single lesson to be learnt from these diverse examples,
however, it is that documents serve to constitute the events of which they form
a part. In that respect they deserve parity of esteem with talk and behaviour in
the execution of the research process.
So just to summarize the key points:

• People do things with documents.

• The task of the researcher should therefore be to follow a document in use.

Following documents in use requires situated (empirical) studies. In particular,

we need to pay attention to the manner in which:

• A document is enrolled into routine activity, who enrols it and who opposes it.
• A document functions, and whether its function alters with context.
• A document serves to constitute an event or phenomenon of which it is itself
part.

RESEARCH EXERCISES

Exercise 3.1

Locate an official (local or national, state or federal) government report

on a set of events where 'something has gone wrong' - such as a
report on an environmental tragedy, an economic or political fai l u re or
a fai l u re of personal d uty. (1) Examine the l i m itations and boundaries
set by the terms of the report. (2) Try to determine whether any
parties attempted to broaden or narrow the scope of the report.
(3) Examine any key assumptions contained within the report and their
i m pl ications for the report's findings. (4) Try to determine the ways in
which any results h ave been us.ed , and by whom. (5) Note particularly
the individuals who used the report to establish claims about issues
that lay outside of the report's field of concern.

Exercise 3.2

Examine exactly how published crime statistics are used and recruited
by members of political parties for various ends. Make a special note
of d ifferences of ' reading' with respect to published results. In the U K
case, Reiner (1996) provides some useful leads o n this issue. See
also Coleman and Moynihan (1996).

Exercise 3.3

Using a selection of school reports on individuals, identify an array of

characteristic MCDs used in the reports. Examine how the MCDs are

Copyrighted Material
 DOC U M E NTS I N ACT I O N I 69

used so as to construct an identity for the people being reported upon.

Analyse and identify how the reports may i mplicate others i n action-at­
a-distance .

Exercise 3 .4

Identify and study a small sample of e,ma i l communications and try to

determine how such mailings h ave been used to dema rcate and under­
pin a social network.

I n executing work for exercises of the type mentioned i n 3.3 and 3.4
above, think carefully about the ethical implications of your i nvestiga­
tion before undertaking it. In executing work for exercises of the type
mentioned in 1 and 2 searches of electron i c newspaper databases
and WWW pages are indispensable.

Notes

1 For example, in his analysis of the street songs and common conversation of
pre-Revolutionary France, Darnton (2000) emphasizes how stories, songs, jokes
and pamphlets belong to a dense web of communication that are sometimes used
to further important political ends. Thus in the world of pre-Revolutionary France
the songs and literature of the street tended to emphasize the scandals of the French
court and the sexual life of the king. The libelles and the chroniques scandaleuses
embodied in written documentation echoed in the notes of the chansonnier in the
cafes. A modern example of translations between talk and text is provided by Jonsson
and Linell (1991).
2 For another example of organizational failure involving documentation see
Vaughan (1 983), and Vaughan (1999).
3 At the time of writing, information on the Great London Smog of 1 952 was avail­
able on the following web pages: http://www.docm.mmu.ac.uk/aric/eae/Air_
Quality/OlderiGreaCLondon_Smog.html and http: //www.met-office.gov. uk/
education/historic/smog. html
4 http: //www.dep.state.pa.us/dep/rachel_carson/donora.htm

Copyrighted Material
4
Documents i n Action I I . M a ki ng t h i ngs vis i b l e

The tree of knowledge 70

Of fungi and fish 72
Making disease visible 73
Making risk visible 79
Conclusions: making work visible 86
Notes 88

The tree of knowledge

People think with things as well as with words (Prior, 1 997) . How they arrange
and organize things in the world is important, not least because - as was shown
in Chapter 2 - the organization of things provides insight into the most funda­
mental aspects of human culture. In literate cultures, of course, the organiza­
tion of things is anchored in writing as well as in three-dimensional space. As
a result the social researcher of such cultures has a wealth of data available for
analysis that is simply not open to researchers in cultures where writing is
absent - though access to other symbolic systems may well be available (see,
for example, Poole, 1 969) .
For a panoramic view of the ways in which people arrange things in the
world there are few better obj ects for study than encyclopaedias. Nowadays
encyclopaedias are common enough items, accepting that the weighty tomes
of the past have been replaced by electronic rather than written media. Yet the
world has not always been brimming with encyclopaedias. In fact, it seems fair
to assert that a set of volumes that claim to provide a comprehensive overview
of the world and the contents within is essentially a feature of modern western
culture.
There is probably little point in entering into debates about the first instance
of any one thing. Yet, it seems fair to suggest that one of the earliest of all
encyclopaedias was that published in Paris during the period 1 75 1-72. The editors
of the vast work sought, above all, to provide a map of available 'knowledge'
and of all the by-ways within it. In short, to impose an encyclopaedic order

Copyrighted Material
 DOCUM ENTS I N ACT I O N II 71

on the world. In so doing they used a number of simple ordering devices.

Foremost among these was the use of an alphabetic . sequence - so the entries
in the encyclopaedia were arranged in dictionary order. 1 (Indeed, in the one set
of volumes we see the emergence of two rather impressive forms of literality -
encyclopaedias and dictionaries.) Alphabetic sequence, was not, however, the
only ordering device that was used by the encyclopaedia's editors. They also
used an ordering metaphor to organize knowledge. That metaphor was of a
tree. The image of a tree of knowledge has, as Darnton ( 1 984) points out, a
much older pedigree than that of the Encyclopedie. It is also of some interest to
note that the only image of evolution provided in Darwin's Origin oj Species is
of a tree, and that the image of a branching tree continues to play an impor­
tant symbolic role in western science generally - see Gould ( 1 989) . In the case
of Diderot and d' Alembert (the editors of the Paris encyclopaedia) there were
really three trees - named, respectively, Memory, Reason and Imagination.
Within each tree there were many branches. So, for example, the tree of
Memory had branches leading to 'nature' and its many divisions. Within that
there were divisions relating to the 'uniformity of nature', and the 'irregulari­
ties of nature'. Among the latter, we find in turn articles on monstrous
vegetables, unusual meteors, wonders of the earth - and so forth.
Monsters are not to be laughed at, and dividing or segmenting the world
into the routine and the monstrous is a serious matter. Above all, i,t is a form of
exercise that provides telling insight into how societies think (see, for example,
Douglas, 1 966; Ritvo, 1 997) . Indeed, one might say that the study of
taxonomic structure (or forms of classification) is key to the study of any
culture precisely because it makes 'thought' visible. That is, in so far as it can
reveal a map of concepts and of the links and associations that are made
between concepts. With that point in mind it is easy to understand why the
authors of the Paris encyclopaedia often referred to their work as . a
'Mappemonde' . Yet, whenever things are divided, much more is implicated
than human thought and culture. For, associated with each and every classifi­
catory system is a set of practices. And it is truly in the links between human
practices and forms of taxonomy that a space for social research opens up.
In fact, if we reconceptualize Diderot's Encyclopedie as an information stor­
age and retrieval system - rather than as a mere dictionary - we will be led to
pose questions not merely about the content of the system, but also about how
the system was accessed and used and modified and challenged. That is, how
the system - as a technology - is nested within a web of activities (Bijker et al.,
1 987) . In recent decades, investigations into the nature of information storage
systems and their relationship' to organizational life have emerged as a distinct
subj ect area in itself (see, for example, Berg, 1 997; Bowker and Star, 1 999) .
And as an introduction to just a few of the issues contained within this newly
emergent field it would, perhaps, be useful for us to consider yet another
encyclopaedia.

Copyrighted Material
72 USING DOCUMENTS IN SOCIAL R E S EARCH

Key C Basidiomycoti na
( G i l l bea ring and pore-bea ring forms)

1 a Hymen i u m consisting of g i l ls
1 b Hymenium consisting of tubes opening by pores

24a Cap ca mpanu late or co nical

24b Cap not campa n u l ate o r conical
25a Cap greasy o r viscid, brightly co l o u red Hygrocybe
25b Cap not with this combination
26a S m e l l i n g disti nctly of fish
26b Not smelling of fish

30a O n wood M icromphale

30b On cones Baeospora, Stro b i l u ru s

FIGURE 4 . 1 The identification of fungi (adapted from M . Jordan, the Encyclopaedia

of Fungi of Britain and Europe, 1995, London: David and Charles)

Of fungi and fish

Consider Figure 4. 1 . It is an extract from a table concerned with the manner in

which we classify what, in England, are often referred to as 'toadstools' . The table
is to be found in a modern encyclopaedia Gordan, 1 995) . It is an encyclopaedia
of fungi, and it is designed to enable users to identify specimens found 'in the
field'. In an anthropological sense the table facilitates the task of identification
in an interesting manner. Thus, one can see that the table is based on a little
more than a series of binary divisions. So, either something is, or it is not. For
example, the fungi either smell distinctly of fish or they do not. They either have
rounded caps or they do not. This manner of operating in binary pairs - some­
times referred to as an Aristotelian system of classification - was regarded by
Levi-Strauss (1 969) as fundamental to all human cultures. Indeed, Levi-Strauss
tried to decode myths and totems in terms of such logic, pointing out how we
should concentrate on how what is said is ordered, rather than with what it
'means'. (For further comments on the approach see Chapters 1 and 7.)
Now the content of any classificatory system is important - as is this one.
Consequently, the order on the page requires analysis. However, what many
structuralists in the literary tradition tend to overlook is that order on the page
is invariably tied into forms of social order and it is the connection between
the two that demands investigation. For example, the use and manipulation of
classificatory systems, as tools, invariably serves to mark out the limits of social
groupmgs.

Copyrighted Material
 DOC U M E NTS I N ACTION II 73

In this light, were we to examine the fungi encyclopaedia a little more

closely, it would quickly become evident that the frame within which the clas­
sification is embedded marks out a certain kind of user. In fact, it is an encyclo­
paedia aimed at 'amateurs' (p. 7) - designed for use by the amateur in the field,
rather than, say, bench scientists in the laboratory. So, from the outset, it provides
the reader with a specific identity for reading (and doing) . At a more detailed
level we would also be able to determine that the text urges people to act in
specific directions. In particular, it encourages people to 'perform' classification
in a defined manner. Thus it suggests that readers smell, taste, look, pick and
preserve fungi in order to identify and distinguish one from the other using 'a
logical and progressive sequence of checks' (p. 7) - such checks as are outlined
in Figure 4. 1 . So as well as providing the reader with the identity of an 'amateur' ,
the system of classification also provides a script for doing. In that sense, the text
orders its readers as much as it orders 'things' (fungi) in the world.
In the same way, the alphabetic ordering of entries in the great Paris
encyclopaedia also orders its readers as much as it orders the entries - it orders
them to search the text by letter order rather than by entry size, or date, or
author, or topic link. (It is, of course, an ordering process that would be con­
sidered by most people to be one that is both efficient and convenient.) In any
event, one of the things that is happening with our fungi encyclopaedia (in use)
is that it is acting back on its readers and structuring them as amateur field
scientists. There are, of course, other ways of ordering fungi, and other classifi­
catory systems. They, however, would be linked into different technological
systems and thereby into different social groupings. Thus, a classification of
fungi according to, say, the features of the genetic (DNA) code would require
a very different group of readers - with different skills, different ways of working
and alternative arrays of technological hardware.
How documents place things, how they make things visible, and how such
systems of visibility are tied into social practices are the dominant themes of
this chapter. Our aim is to look at various ways in which graphical and similar
forms of representation can materialize things (phenomena) that would other­
wise remain opaque and diffuse. Naturally, in order to make things visible,
human actors and agents have to translate ideas into images and traces. Such
processes of translation are various. Indeed as we shall note the 'same' obj ect
can be translated (Serres, 1 995) into a number of alternative forms. How the
forms relate one to the other and how they act back on their creators are,
however, always a matter open to empirical research.

Making disease visible

The diseased brain

In his Art and Artifact in Laboratory Science, Lynch (1 985: 1 53) points out how,
'Documents are integrally a part of the work of doing science'. Documents

Copyrighted Material
74 USING DOC U ME NTS I N S O C I A L R E S EARCH

serve not simply as a source of ideas and information but also as an integral
component of bench work itself (in the form of notes, recipes, instructions and
so forth) . One type of document that appears prominently in Lynch's study is
the pencil and paper diagram. In most cases such diagrams portray structural
aspects of a particular part of rats' brains (the hippocampus) . Such drawings are
themselves based on other images - derived, for example, from laboratory slides
of sliced sections of rat brain, and electron microscope enlargements of brain
tissue. The preparation of slides, the staining or 'labelling' of tissue and the
microscopic enlargements serve to make the hippocampus and its functioning
processes visible. In the interaction between organic materials, technical
processes and human activity, of course, the possibility arises that some of the
things that are visible are not 'natural'. In the language of the lab scientist these
latter are referred to as artefacts. For example, it may be that the staining
process or the cutting process alluded to above produces marks, blotches,
appearances that are a product of the experimental interventions rather than
ordinary features of the hippocampus. Exactly how lab scientists distinguished
between artefactual and natural effects was one of the issues that Lynch set to
examining. This was especially difficult in those cases where the artefact was an
absence rather than a presence - that is, in those cases where something that
should have appeared on a slide or an enlargement failed to appear. (Note also
references in this chapter to Figure 4.2.)
This relationship between organic matter, pencil and paper images, electron
microscope images, staining techniques and so forth is central to the produc­
tion of many kinds of fact. As other science analysts have indicated (Bastide,
1 990; Fujimura, 1 996; Latour and Woolgar, 1 979; Lynch, 1 990; Myers, 1 990;
Rapp, 2000) the world of nature is never immediately visible but has to be
made and manufactured in order to be seen - Rapp (2000) refers to 'imagistic
knowledge' in this context. Among the procedures for making things visible,
photographic imaging techniques have an important place. (Since the comple­
tion of Lynch's work, of course, other imaging strategies and technologies have
come to the fore, some of which will be mentioned below.) The graphic image
thus presents only one of many ways of reading or seeing 'things'. Indeed, as
the following example illustrates, the same 'thing' can usually be seen or read
in many different ways, some of which are graphical and some of which are,
shall we say, numerical, and some of which are behavioural. The relationship
between ways of seeing (documenting) things and forms of professional prac­
tice constitutes a potentially important field for social scientific research. In the
first part of the discussion that follows we shall focus on ways of documenting
what is known to be a common form of disease in older people - Alzheimer's.
Alzheimer's disease is a well-known form of dementia. The key features of
dementia are memory loss, confusion, mood and behavioural disturbance. In
most cases people simply lose ordinary skills of daily living. Dementia is most
frequently found in older people (it affects around 20 per cent of people aged
80 years or more) , but in rare cases it can occur in people of half that age. In

Copyrighted Material
 DOC U M E NTS I N ACT I O N II 75

the literature of the UK Alzheimer's Disease Society, the disorder was described
as follows during 200 1 :

Alzheimer's is a physical disease which attacks brain cells (where we store memory) and
brain nerves and transmitters (which carry instructions around the brain) . Production of
a chemical messenger acetylcholine is disrupted, nerve ends are attacked and cells die. The
brain shrinks as gaps develop in the temporal lobe and hippocampus, important for
receiving and storing new information. The ability to remember, speak, think and make
decisions is disrupted. After death, tangles and plaques made from protein fragments,
dying cells and nerve ends are discovered in the brain. This confirms the diagnosis.
(http:// www.alzheimers.org.uk)

The Alzheimer's Disease Society literature also lists a range of specialists who
can be involved in the diagnosis and the management of the disease - these
include psychiatrists, neuro-psychiatrists, psychologists, nurses and social workers.
The list is in itself instructive because it suggests that what the disease 'is' is
likely to differ according to the forms of professional practice that surround it.
In particular, it seems likely that different professional groupings have different
images of the disease in practice. Note that, according to the passage above, the
'disease' can only be confirmed at autopsy (that is, after death) .
Making Alzheimer's visible in living people (as opposed to making it visible
in the dead brain) is a difficult process and as with so many things, it can be
done in different ways. Thus diagnosing people on what are called clinical
grounds is possible and desirable, but clinical diagnosis and diagnosis at autopsy
do not always match. Thus it is sometimes said that clinical diagnosis is only up
to 80 to 90 per cent ' accurate', a point that serves to highlight how different
forms of visibility provide different answers to our problems.
Clinically, the condition is determined by examining how the patient
behaves and talks and reacts. In fact, clinicians usually refer to three realms of
evidence in the determination of a diagnosis. These. are the symptoms that the
patient brings to the consultation (say, a report of'not feeling right') ; the signs
that the doctor discovers in the patient (such as memory loss and confusion);
and evidence available from investigations (such as brain scans) . So there is
always an ensemble of traces - of visibilities and translations - used to deter­
mine whether a person is or is not dementing (Harding and Palfrey, 1 997), and
in the early stages of the disorder the decision may be far from clear cut. (There
is a sense in which the diagnosis is always and forever a matter for negotiation /
Even were we to concentrate solely on visual data - from brain scans - we
would still be faced with alternative pictures of what is apparently the same
disease. Thus, there are several types of brain scan available. For example, CT or
CAT (computerized axial tomography) scans are a way of taking pictures of
the brain using X-rays and a computer. MRI (magnetic resonance imaging)
scans also use a computer to create an image of the brain but, instead of
X-rays, they use radio signals produced by the body in response to the effects
of a very strong magnet contained within the scanner. SPECT (single photon

Copyrighted Material
76 USING DOCUMENTS IN SOCIAL R E S EARCH

emission computerized tomography) scans look at the blood flow through the
brain rather than at its structure, whilst electroencephalography (EEG) traces
would provide yet another means of imaging brain dysfunction. Indeed, each
type of scan provides different kinds of visibility and evidence for brain 'patho­
logy', though none of them on their own would suffice to provide a diagnosis
of Alzheimer's. Indeed more important than the evidence derived from scans
would be evidence gathered from an examination of a patient's routine behavi­
our. All in all, then, and to use a term that we have called upon previously, we
can say that there are radically different performances (Mol, 1 999; 2000) of
Alzheimer's - neurological, psychological and behavioural. How the one form
is translated and then meshed into the other is a matter of interest in itself,
though here we shall focus simply on the role of documentation in providing
translations .
As an adjunct to a doctor looking at and talking to the patient, it is possible
for him or her to use a standardized set of questions in the shape of what is
usually referred to as an outcome measure. (Note how the very term 'outcome'
gives emphasis to the notion of performance.) Outcome measures usually take
the form of a series of observations or questions about behaviour, the answers
to which can be scored. For example, in looking at dementia, an outcome mea­
sure is likely to include questions about whether a person forgets names, and
dates, and events; about whether they lose their way around familiar places, or
lose things; or whether they have lost ordinary daily skills - of dressing, washing
and self-care. To each question there is attached a numerical score. Scores can
then be added up.As we saw with the CIS-R in Chapter 2, the results can then
be translated in terms of a numerical scale (say, from 0 to 1 00) so as to form a
judgement about the presence or absence of Alzheimer's. (For an example of
such a measure see the references to the Mini Mental State Examination on
http : // www.alzheimers.org.uk)
As just stated, Alzheimer's usually affects older people. In some families,
however, it can have what is called an 'early onset'. That is to say, it can appear
in people aged from 35 to 60 years. So, for example, there are about 1 7,000
affected people in this age group within the UK. One group of people who
tend to suffer from an early onset of Alzheimer's are those with Down's
syndrome. Down's syndrome is a genetic disorder which usually results in
people having learning difficulties. In conjunction with the learning difficulties,
there is often a characteristic facial appearance. (More often than not, it is
the facial appearance that people focus on. One might say that that is how the
syndrome is made visible in lay culture.) Naturally, the severity of the learning
difficulties and the accentuation of the facial features vary considerably in the
Down's population.
Autopsy studies have shown brain changes (neuropathology) similar to that
ofAlzheimer's disease in almost all adults with Down's syndrome over the age of
40 years. It is an observation that is supported by neuroimaging findings - of
the type derived from the scanning procedures noted above. Neuroimaging is

Copyrighted Material
 DOC U M ENTS I N ACT I O N " 77

a relatively new kind of technology. As a form of documentation it provides us

with new translations of what Down's syndrome is. In the absence of neuro­
images, however, professionals tend to restrict themselves to looking at the
behavioural dimensions of the disorder.Yet, by looking at behaviour, dementia
can be diagnosed in only around 9 per cent of adults with Down's syndrome
aged 40-49 years, and around 55 per cent of those aged 60-69 years. In short,
although one might be able to 'see' the presence of Alzheimer's disease in a
human brain (if one had the tools) , one cannot easily detect it in daily life.
In a sociological sense therefore we might say that there are at least three dif­
ferent sites in which Alzheimer's can be made visible - namely, in the autopsy
room, in the neuroimaging suite and in the clinic. Different forms of text and
visual image are associated with the different kinds of setting, and one of the
tasks of the social scientific researcher should be to examine how the various
forms of documentary evidence available to human actors are woven into
specific forms of translation. This is because what the disease 'is' changes
according to the availability of the images. Or, to put things another way, the
visibility of the disease is dependent on the use of different kinds of inscription
device (Latour, 1 987) " by different kinds of professional.
As we have stated, an important set of such devices are the outcome measures.
Yet, there are various kinds of such measures, and they in turn give different
images of the disorder. For example, one such instrument is the Mini Mental
State Examination (MMSE) (Folstein et al. , 1 975) . Another is the 'Dementia
Questionnaire for Mentally Retarded Persons' (DMR) (Evenhuis et al. , 1 992) .
Yet a third is the 'Dementia Scale for Down's Syndrome' (DSDS) (Gedye,
1 995) . In one clinic, where all three measures were applied to a rather mixed
group of people, it became evident that of 30 subjects who had an MMSE
score of less than 24 (which is taken to be the usual cut-off for the detection
of possible dementia) , 23 did not have a diagnosis of dementia according to any
other criteria. In other words, the instrument produced dementia where it was
not otherwise evident - indeed, it seemingly overestimated prevalence by
about 75 per cent.
So the upshot of all this is that the outcome measures very often give
different answers to the same basic question, 'is this person dementing?' In so
far as that is the case, we might say that whether a living person has Alzheimer's
or not is often a matter of documentation rather than of biology. It is docu­
ments that make the disorder visible, and it is documents that define its sever­
ity. In terms of situated social scientific enquiries, of course, it would be crucial
to note how the various means for making Alzheimer's visible are called upon
and manipulated by different kinds of people. And how those tools are used in
different kinds of organizational setting - hospitals, care homes, day centres and
so forth. Rather than pursue the Alzheimer's example any further, however, I
will turn to a published example in order to illustrate how forms ·of classifica­
tion and scaling are linked into professional practice, and how they often serve
to define professional boundaries.

Copyrighted Material
78 USING DOCUMENTS IN SOCIAL R E S EARCH

Viewing pathology

I n an intriguing essay on the classification of the lymphomas (tumours) and

leukaemias, Keating and Cambrosio (2000) indicate how classificatory order
has changed during the last 30 years and how such changes have been linked
into forms of professional activity. For example, until the 1 970s the classifica­
tion of the lymphomas and leukaemias was essentially based on the study of
cell morphology (the shape and size of cells) . Using microscopes and staining
techniques, lymphomas could be seen and differentiated according their gross
characteristics. During the 1 970s, however, a different form of classification
came into operation. It involved a study not so much of cell morphology but
of cell function. In particular, the new system depended on the identification
of different types of cells called 'T' and 'B' cells, which (at that time) were not
distinguishable under the microscope. The new classification led to a new
understanding of what lymphoma was. Unfortunately, the classification was a
product of (laboratory-based) pathologists and not deemed to be useful to
clinicians (the doctors who actually speak to patients) . Above all, the latter
want to be able to 'see' a disease in the patients that they deal with, and to
make predictions about eventual outcome. Clinicians are not particularly
interested in new ways of classifying things - especially if this does not help
them to deal with patients. So although new ways of seeing had been intro­
duced into the medical world, not all professional groupings were happy to
adopt the new vantage points. In fact, as Keating and Cambrosio point out, as
far as the classification of lymphomas is concerned there is a constant tension
between what can be seen and discovered in the laboratory and what can be
used in the clinic.
Such tension can itself be turned to productive use. In terms of the example
that we are discussing here, the tension between clinicians and pathologists was
in part responsible for further developments in classificatory techniques, for
during the 1 980s a specific way of seeing 'T' and 'B' cells was developed. The
new method involved treating the cells with fluorescent antibodies and then
running them through a machine called a flow cytometer.
The arrival of the flow cytometer meant that cells could be counted and
quantified in various ways. Clinicians like to quantify things. For, as in the
Alzheimer's case, once things can be counted, they can be scored, and once
scored they can be used to classify anew. Unfortunately, despite the fact that the
flow cytometer technology had found a way of making cells fluorescent and
thereby amenable to quantification, the results were still of little direct use to
the clinicians. Many clinicians therefore tended to hang on to the old (morpho­
logical) ways of seeing the lymphomas and leukaemias. One consequence of all
this is that the modern pathology reports that move out of the laboratory and
into the clinic contain many visions of the same thing. Thus, pathology reports
commonly contain diagrams illustrating gross pathology, flow cytometer counts,
and more. Ways of seeing and ways of classifying proliferate.

Copyrighted Material
 DOCUMENTS I N ACTION II 79

The pathology reports, then, have to make disease visible in different ways.
They have to take account of the fact that clinicians have been trained in
different ways and educated into various forms of classification. I n that respect
such reports might still be thought to serve merely as conduits for the trans­
mission of data between one person and another. What we need to note is that
as well as serving as a conduit of information, the reports also mark out spheres
and boundaries of professional influence. Indeed, systems of classification can
be seen to act as 'boundary objects' (Star, 1 989; Star and Griesemer, 1 989) . In
short, the world as ordered on the page (in text) is linked to other forms of
order in social practice.3 In this instance, forms of classification are related to
different ways of doing analysis, to different ways of observing the world, and
to different sites of professional practice. In this light it is interesting to note the
claims of Myers ( 1 990) who compares scientific illustrations in popular and
professional scientific journals. In the former case, he suggests, illustrations are
tied into a mode of presentation that emphasizes the independence of things
(or a narrative of nature as he refers to it). In the latter case the illustrations are
tied into what scientists do - to a narrative of science.
In this subsection I have focused almost exclusively on classification and
medical practice. However, it is evident from the published work that these
same issues arise in other fields of human endeavour - including, for example,
business practice (see Desrosieres, 1 994) . In the following section, however, I
intend to stick with medical practice as a source of my examples. This time I
am going to consider how risk, rather than disease, is to be made visible. In
particular I am going to consider the notion of genetic risk. The clinic that is
to be referred to was the site for a research project that sought to discover how
'risk' is assembled by medical professionals.4

Making risk visible

Risk looms large in contemporary medical discourse (Gabe, 1 995) . Indeed,

Skolbekken ( 1 995) has spoken of a risk epidemic and has documented the
marked rise of risk-related publications during the 1 967-9 1 period (see
Chapter 8) . It is of considerable interest in itself as to why the language of risk
came to dominate medicine in the last quarter of the twentieth century.
Whatever the reason or reasons, however, it is quite clear that being-at-risk is
now a major focal point for medical research and clinical work. One of the
consequences of such a focus is that it draws healthy as well as unhealthy
people into the orbit of medical investigation. This is simply because being­
at-risk necessarily predates being diseased. The language of risk can therefore
be used quite easily to extend the domain of medical knowledge and of
medical practice almost without limit. In the world of clinical genetics and
genomics, especially, this capacity to be 'at risk' is central to the entire medical

Copyrighted Material
80 USING DOCU M ENTS IN SOCIAL RESEARCH

enterprise. That is because the clinical emphasis is on the detection and

prevention of pathologies rather than on 'cure'.
The new genetics is very much a product of the late twentieth century. In
fact, during the mid point of the twentieth century, scientific knowledge of
genetic inheritance underwent a revolution (Keller, 2000) . Until then, it was
recognized that genetic properties were carried in chromosomes, but the inte­
rior structure of chromosomes remained a mystery.s In 1 95 1 , however, Crick
and Watson published an important paper on the helical structure of DNA and
discussed its chemical components. Bazerman ( 1 988) both reproduces and
examines the paper in terms of the scientific rhetoric of the age.
Crick and Watson argued that the long strands of DNA that were contained
within chromosomes were composed offour different chemicals. These chemi­
cals are often referred to as C, A, G and T. (There is no need here to specify
what these letters represent.) Put very bluntly, these four chemicals combine in
various ways to make proteins, and it is out of such proteins that human physi­
ology is structured (Pilnick, 2002) . In some people, however, the sequences in
terms of which the four chemicals are replicated are abnormal. Perhaps, for
example, key sequences of CAGT are missing (truncated) , or perhaps they are
duplicated a few too many times (repeated) . In yet other instances the sequences
will be jumbled. When such things occur, there is a possibility that the abnor­
mality indicates a 'mutation', and in many cases mutations are associated with
bodily pathology.
For example, the two genes associated with breast cancer (BRCAl and 2)
are located on chromosomes 1 7 q and 1 3q respectively. If some of the chemi­
cals in the BRCA sequences on those chromosomes appear in abnormal
proportions, or are missing altogether, then there is a high probability that the
individual concerned will develop a cancer. (More than 300 mutations are
currently identified for BRCA 1 .) Mismatches or mutations are carried in
many families and individuals, but whether an individual person may be said
to be at 'low', 'medium' or 'high' risk of developing an inherited pathology is
always a tricky and awkward business (the mismatches have to be of the right
kind) . In the UK, finding mismatches of the right kind is a business that
involves a considerable amount of documentation. In fact, documents are
central to the entire process through which risk assessments are manufactured.
An estimate of being at risk of breast or of any other cancer is commonly
calculated as an average lifetime risk (currently around 1 in 1 2 for female breast
cancer) . That is to say, risk is calculated in terms of populatio!'!sYet for any indi­
vidual the risk may be higher or lower than the average. This element of uncer­
tainty about personal risk requires management in many different spheres,
some of which belong to the world of lay people and others of which are
firmly lodged in the world of professionals.
In order to identify people who are at a potentially high risk of an inherited
cancer, geneticists have devised their own rules for distinguishing between
people (and families) at different levels of risk. These rules are usually drawn up

Copyrighted Material
 DOC U MENTS I N ACT I O N II 81

as a product of expert meetings and discussions, consultation of published

research studies and so forth. In an individual clinic they are usually contained
in a 'protocol'. A protocol serves to classify the kinds of people that a clinician
is most likely to be interested in focusing on. For example, as far as breast
cancer is concerned, the protocol for inclusion in the genetics service being
referred to here required at least one of the following criteria to be met by a
man or woman seeking help. One first-degree female aged under 40 affected
by breast cancer; two first-degree relatives (on the same side of the family)
affected at 60 years of age or less, three first- or second-degree relatives of any
age (on the same side of the family) ; one first-degree relative with bilateral
breast cancer; one first-degree male breast cancer. Different criteria (the
Amsterdam criteria) would be used for colorectal cancers, whilst for ovarian
cancers the requirement would be for two or more such cancers within the
family with at least one first-degree relative affected. In practice, of course,
these criteria are open to modification in the light of various factors - some
'social' and others 'scientific' - and it is an intriguing problem in itself to look
at how the protocol is applied in everyday clinical work. In what follows clinical
professionals are attempting to deterrrune the risk of breast cancer for a specific
individual. The extract, in itself, provides intriguing data on how people
'follow' rules in action. You will also note that reference is made in the talk
to a computer program called 'Cyrillic'. We will examine the nature of that
program shortly.

CG1: So the proband at 39. Mum [cancer of] breast 38. Dad is alive and
wel l . The only other fam ily history which we are trying to find is a
mum's s i ster who allegedly had breast cancer at the age of 49, but she
knows very l ittle about her, and hasn't kind of got back to us, but
c learly quite young. So the breast was at 38 and the bowel was at 63.
CG2: She is moderate for breast isn't she?
CG3: She is automatically moderate for two under 60. And one is under 40
isn't she?
CG1: Yeah one is 38. Does that make her h igh with one under 40?
CG3: Wel l she might end up being a 24 per cent or a 26 per cent.
CG1: On Cyri l l i c .
CG3: On Cyri l l ic . W h i c h will switch h e r from o n e t o t'other. G u t feel ing i s
moderate.
CG2: M m . She has got a sister here who is . . .
CG3: OK.
CG2: And she is fine. And there is nothing else here now. So she is probably
moderate. S he will fit the moderate criteria.
CG1: She is coming up for 40 [ the age at which mammography is offered for
women who are at moderate risk].
CG2: She is 40 anyway. Yea h .
CG3: Wel l one under 40 would automatica l ly put her into moderate. So that
is why I ' m wondering whether we should thi n k h igh with another one

Copyrighted Material
82 U5ING DOCU MENT5 IN S0C IAL R E S EARC H

over here. [Pointing to the maternal aunt.] The only other thing is
because of having a second primary i s therefore . . . if she was at 27
per cent on Cyri llic and mum had a second primary I would be more
inclined to accept that as high risk.
CG1: So do you want me to run Cyri l l ic?
CG3: Yeah.
CG1: OK. And so therefore if she is above 25 per cent it takes her into high?
CG3: Yeah.
CG1: And so below 25 per cent is moderate? Sorry [CG1 writes in patient's
file] writing it a l l down .
CG2: It won't take in the bowel though will it, Cyril lic?
CG3: No. And I don't think it should either. It's just that we know that if you
have got a breast cancer . . .
CG1: So if it's dead on 25 per cent?
CG3: Come and speak to me. [Laughs.]

Clearly age and previous family history are all important in the determination
of risk. Family history is gathered by using a questionnaire on the medical
history of a person's relatives. The questionnaire seeks information of dates of
birth (and death) of blood relatives, dates at which any cancers appeared, types
of cancer, patterns of marriage and reproduction and so forth (information that
is not always available - as we can deduce from the early stage of the extract
above) . When this information is put together clinic professionals talk of having
a pedigree - of the kind indicated in Figure 4.2.
Producing pedigrees - as with the production of DNA sequences - is a
matter of inscription. How that inscription is executed often differs between
cultures. Thus, Nukuga (2002), and Nukuga and Cambrosio ( 1 997) offer
examples of how medical pedigrees are assembled in Japanese and Canadian
contexts. As Nukuga and Cambrosio point out, the results look very different.
They look different because pedigrees are not simply transcriptions of how
things really are, but in large part cultural representations of human relation­
ships. (!We referred to a further example of this representative power of family
trees in Chapter 1 .)
In the clinic to which I am referring the pedigrees were put together through
a complex process of talk, investigation and transcription. Information provided
by patients, for example, was often checked and rechecked against various
sources. Thus professionals sometimes sought data from a cancer registry in
order to complete or to confirm a pedigree. A cancer registry is, of course, yet
' another information storage and retrieval system - and one worthy of detailed
investigation. It registers all cases of cancer morbidity in a given region and is
often crucial to the work of geneticists and epidemiologists seeking evidence
for familial cancers. Naturally, other blood relatives could also serve as a source
of data on family illnesses. But whether from a registry, or a hospital, or from
relatives, such information was sifted through talk and technology not only so
as to represent a family tree, but also so as to constitute the nature of the
family under investigation. This capacity of documentation to constitute the

Copyrighted Material
 DOC U M E NTS I N ACTION I I 83

Condition: BrCa - Breast cancer susceptibility Drawn by: Fiona Wood

1:4 1:3 1:1 1:2

Paul Marie Stephen Louise
Fabricated Fabricated Jones Jones
1 Apr 1910 7 Apr 1920 1 Jan 1900 3 Apr 1922
80 79 55 44
ca breast

11:1 11:2 11:3 11:4

Mark Mary John Ann
Fabricated Fabricated (Jones) Jones Jones
2 Feb 1940 5 May 1940 11 Nov 1950 6 Jul 1945
61 50 51 66
ca ovary ca colon

For individual: Jane Fabricated (111: 1 )

)f 1l1:1 111:2 111:3
Calculated risk of cancer at age 32
Heterozygote risk: 19.2 per cent
Jane Jim Sarah
10 year risk: 4.1 per cent
Fabricated Fabricated Fabricated
20 year risk: 9.5 per cent
1 Feb 1969 3 Mar 1965 4 Apr 1966 30 year risk: 14.5 per cent
32 36 35 Risk to age 85: 28.1 per cent

FIGURE 4.2 Cancer in the family: the fabricated family's pedigree

obj ect of which it is supposedly nothing more than a representation has, of

course, appeared in some of the examples drawn upon in earlier chapters.
In large part, the assembling of a pedigree is a matter of paper and pencil
technology - and pedigrees are normally stored in paper files within the clinic.
To translate the data in the files into a risk assessment, however, requires
further productive work. For clinical purposes this is partly achieved through
the use of a computer program known as Cyrillic (Chapman, 1 997) . (You will
note that Cyrillic is referred to a number of times in the extract above.)
Cyrillic is a program that draws a family tree and provides a numerical risk
estimate for an individual. It is an inscription device (Latour and Woolgar, 1 979)
par excellence. It makes risk visible. It provides an explicit trace of hereditary
influences, and it constitutes objects for medical scrutiny afresh. It is always
worth taking inscription devices apart. In the case of Cyrillic the act of dis­
assembling is particularly instructive. For what the program is ultimately depen­
dent upon is a statistical model, and that in turn is linked to surveys concerning
the incidence of breast cancers in specific communities. Two of the more widely
used models are the Gail model (Spiegelman et al. , 1 994), and the Claus
model (Claus et al., 1991). The role of family history - and of other factors - is
somewhat different in each model and so a given woman's risk of cancer is

Copyrighted Material
84 USING DOCUMENTS IN S O C I A L R E S EARCH

dependent not simply on items drawn from her personal biography, but also on
the relative weight that is given to different factors. For example, risk alters
according to such things as age, age at menarche, etc. Thus a 50-year-old woman
who has not had breast cancer has a lifetime risk of 1 1 per cent instead of 1 2
per cent (for she has lived through most of her risk period and been free of
cancer) . Since the different models allocate different weights to each of the
factors they can often produce widely different estimates of risk for the same
human being.
Thus, in a recent comparison of risk assessments applied to 200 UK women
attending a breast cancer clinic (Tischkowitz et al. , 2000) , it was noted that the
proportion of such women allocated to a high-risk (of hereditary breast cancer)
category varied markedly - from 0 .27 using one method, as against 0.53 for a
second method. A third method allocated only 0 . 1 4 to the high-risk category.
So there are some women for whom risk is systematically 'underestimated' by
the very nature of the models. This is so because the populations on which the
risk models are based are themselves biased. For example, they contain only
(North American) women; women who predominantly work in the professions;
and women with documented family histories. Further, the samples under­
represent ethnic groups known to be at high risk - such as women ofAshkenazi
Jewish descent.
Such biases in the computerized inscription devices naturally create diffi­
culties for clinicians, and they serve to underline the problems inherent in the
primary translation of risk from populations to individuals. As a result, the data
from the paper and pencil exercises, the discussions of professionals, the data
derived from Cyrillic are always open to expert interpretation. In fact, the
documents of which we speak always need to be mobilized as evidence for or
against the further investigation of risk. In those instances where a high risk is
conferred on an individual (known as the index case or proband) , the option
to undertake DNA testing arises. And when patients opt to be tested it is, in
theory, possible to produce laboratory traces of risk.
As with so many other things in the world of science, there are different
ways of making DNA structures visible. Indeed, laboratory scientists routinely
manufacture images of mutations through the use of centrifuges, dyes, gels,
electrical currents, and ad hoc laboratory techniques, in different combinations.
For example, Figure 4.3 provides an image of a DNA sequence (relating the
breast cancer gene) that was produced in a dyed gel. Using this technique,
DNA is obtained from a blood s'ample, mixed with a fluorescent tag and
inserted into 'wells' in a gel. The mixture is then 'filtered' through the gel by
means of electric currents. In this instance the figure contains DNA sequences
from a number of samples (samples obtained from individuals) . Thus there are
samples from person 'B l ' , 'B6', 'C6', 'D6' and so forth. I have emphasized two
such samples with vertical arrows along the top of the image. We can also see
that there are a number of bright and not so bright bands running (horizon­
tally) across the diagram. I have attempted to highlight two such bands with
horizontal arrows in the left hand margin of the figure.

Copyrighted Material
 DOC U M E NTS I N ACTION II 85

FIGURE 4.3 Visualizing mutations through gels (UK laboratory, 2000)

In many ways the appearance of the bright (fluorescent) and dark bands ought
to be similar for each individual. We can see, however, that they are not. For
example, the pattern of bands for the person highlighted with the first (broken)
vertical arrow seems to be somewhat out of sequence with other samples in
terms of where the bright bands are occurring, whilst the column highlighted
by the rightmost vertical arrow seems to have hardly any bright bands in it at all .
Now, the fact that the columns are different - one from the other - emphasizes
the fact of human variability in genetic structure. In short, the gene sequence
for individuals can be very different even when we look at what is supposedly
the same gene (in this case BRCA 1 ) . However, whether such differences are to
be regarded as significant or not is a matter for expert interpretation. For example,
the expert reading the gel represented in Figure 4.3 has to decide whether a
visual difference in a sequence indicates the presence of a mutation or what is
called a 'polymorphism' . In the latter case there is said to be an abnormality in
a person's gene sequence, but one that is, perhaps, relatively common and
therefore assumed to be unimportant. In the former case, the abnormality is
assumed to be significant.
In the case of Figure 4.3 all of the variations were considered by the labo­
ratory scientists to indicate the presence of a polymorphism. (The column
highlighted by the rightmost vertical arrow represents a technical aberration, a
failure - it is, if you like, an artefact of the production process.) Consequently,
none of these samples were considered for further laboratory investigation.
However, had a mutation been suspected then a further stage of investigation
would have been warranted.
Figures 4.4 and 4.5 provide a second means of imaging DNA sequences.
Here the graphical representation is presented in four different colours - each
colour provides a trace of one of the four key chemicals C, A, G, T. In Figure 4.4
a trace of a normal fragment of BRCAl is illustrated. To the professional eye
it looks smooth and regular, and the sequence as traced between the two
arrows is known to correspond to a normal pattern for this particular segment.
Figure 4.5 should in many ways look rather similar to 4.4, but it does not.

Copyrighted Material
86 USING DOC U M ENTS I N SOCIAL RESEARCH

Between the points marked as 50 and 60 all is well, and then the trace goes
haywire. Comparing the two diagrams one can easily see irregularities from the
second row onward in Figure 4.5. Colours get mixed and the ambiguities of
colour lead to the appearance of the letter 'N' in the sequence. The laboratory
scientist explaining this fragment to a social scientist spoke as follows:

L5: This one starts off the same - let me find it - so it goes along it is pretty
m uch the same and then you hit this point [pointing to the second row of
the sequence] yeah?
55: Yeah.
L5: And it looks l i ke it' s a bloody mess. And the reason is if you have a look
at it, it means that the base can either be a T or an A and that one can
be a G or a n A. That one is an A on its own. That i s a T. That i s a G or
an A. If you write the sequence out, one of them will have a normal
sequence in it, and this ( Figure 4.5) i s a big 40 base pair deletion. So at
40 bases along the other one - it' s very difficult to show you without the
sequence in front of me, but you can see that. Let's get the fi le out. This
i s fragment 1294. It goes AGTGATGAAC, and then it goes n uts . The next
one is a G or an A - the normal is a G . . . [ Figure 4.5 shows a T. ]

In short, then, the figures are used t o demonstrate abnormality. One can 'see'
the mutation (if one knows how to read the diagrams) . Such traces can be
used to make risk visible in any number of contexts - including consultations
with the affected person. Indeed, genetic risk is made manifest through
various media: talk, machinery and documentation. From the standpoint of
the social researcher, however, it is important to look at how these items and
elements are worked up in practice, at how the documentation is produced
and at how it is mobilized and enrolled by various parties to practical ends
(see for example Barley and Bechky, 1 994) . In clinical genetics (and genomics
in general) making things visible is of paramount significance. Luminosity (as
we have noted) is central. Often such luminosity is achieved through what are
regarded as high-tech fixes. On other occasions, however, it is achieved in
remarkably ordinary ways using nothing more than pencil and paper traces.
Nevertheless, it is clear from what we have said so far that assessments of
genetic risk call upon a complex ensemble of technologies, and in the
midst of such technologies it is documentation that makes the world - and
the risk within - perceptible and palpable (for further examples see Prior
et al. , 2002) .

Conclusions: making work visible

In Sorting things out, Bowker and Star ( 1 999) pay considerable attention to a
classificatory system referred to as NIC. NIC stands for Nursing Interventions
Classification. Among other things the NIC provides lists of activities that -

Copyrighted Material
Ci G ct N N H N N N N N H G G N l'C N H I'f N N H N N N H H N N H N N H N r T i t lt T
HlI 28

.. .

r G � G r C ( CAC r A. I. T AG c .. e C A. IT CA au A. G 1 T t A. I co A. G T

FIGURE 4.4 A normal DNA sequence (UK laboratory, 2000)

Copyrighted Material
 (. ! A A T Il. G C A C CIl T r CA t A UC T a AT
l� )� � ft

" 1; T G (; T r T r c CA (I A � Ii T G " T (i ,., A C T CO A. T (i ( C r 1 N T K A C Nil. N r CG <

11 1 " N NeGG G A( r A r AN H A G G n "1 N e e A A 1 N 1 r (" N N G N T It 1

ge l ee HI 1 .6

no 1 M!

FIGURE 4.5 Abnormalities in the DNA sequence (UK laboratory, 2000)

Copyrighted Material
 DOCUMENTS I N ACTION II 87

nurses do. For example, nurses help patients with respiratory problems to
breathe more easily. So in the NIC there is a section on 'airways management'.
Yet again, nurses might also be expected to provide 'spiritual support' for their
patients and so the NIC provides a list of spiritual support activities. In like
manner, whatever nurses 'do' is documented in the NIC.
There are numerous j ustifications for developing an instrument such as the
NIC. One of the major ones, however, is that it enables nurse work to be made
visible and (in a literal sense) accountable. Thus by allocating numerical codes
to the kinds of activities just referred to it becomes possible to computerize the
billing of nurse time. This, as we have noted before, is especially important in
the health care systems of North America. In fact, as the authors of the NIC
have argued, once one has developed categories of action one can bill them,
time them, control them, monitor them, research them, teach them and mani­
pulate them in e�dless ways. In that respect many forms of what might first
appear to be nothing more than forms of technical documentation can be used
to exercise both surveillance and power in organizational life (Foucault, 1 99 1 ) .
I n making work and other phenomena visible, documents play a key role. We
have noted in this chapter how both diseases (Alzheimer's) and processes (risk
assessments) can be made visible through the manipulation of pencil and paper
traces. Making things visible is not, of course, simply a process that occurs in
medical settings. Thus, Porter (1 994) and Desrosieres (1 994) have both indicated
how the measurement of'value' and quality in economic settings depends, in the
first instance, on the means by which we make such entities visible. And the exam­
ple of the NIC (above) extends the process of making things visible into new
dimensions. In fact, it does so for a number of reasons, not least because it serves
to highlight how documents (computerized categories and classifications) can be
accessed and used in a multiplicity of ways. This multiplicity of use arises from the
fact that documentation of the kind that we have referred to is always embedded
in contexts. The prime issues for the social scientific researcher is therefore to
investigate how documentation functions in situated contexts, not least because
such documentation is invariably used in ways undreamed of by its creators.
Here are some key points from our analysis:

• Documents make 'things' visible and .traceable.

• Since the same things are often made visible (translated) in different ways, we
need to look at how forms of visibility and forms of documentation are linked.
• Since different agents use different tools for making things visible, we need also
to look at how the agents and the tools are linked - how documents function
as boundary objects (Star, 1989) .
• In any system of translations there will be controversies and disagreements
about what is being made visible. How documents are used in such controver­
sies should be central to the social scientific enquirer.
• Never look at documents in organizational settings as isolated tools, but seek to
discover how a document is linked into the wider information storage and
retrieval system of which it will form a part.

Copyrighted Material
88 USING DOCUMENTS IN SOCIAL R E S EARCH

RESEARCH EXERCISES

Exercise 4.1

Ideally, studies of the ways in which 'things' are made visible are best
executed through forms of ethnographic investigation. H owever, it is
possible to study images as representations in various ways. E m mison
and Smith (2000) provide a series of excellent examples. The following
suggestions are merely illustrative of some possibilities.
C lassification of 'things' often occurs i n space and place as wel l
as o n the inert page. Art gal leries a n d m useums of a l l kinds offer
two kinds of setting in which things are classified and arranged. By
concentrating on just one or, at most, two galleries of exhibited
artefacts it should be possible to carry out the following tasks.
(1) M a ke a n ote of criteria used to classify and organize the items
contained in the galleries. (2) M a ke a note of key concepts that are
contained i n the written i nscriptions connected to the a rtefacts.
(3) Locate the organization of the things studied in the wider
classificatory scheme of the museum as a whole, and (4) note how
the physical organization of the exhibits structures and guides the
experience of the visitor. To assist with a n a lysis of findings the essay
by Lidchi (1997) , cited i n the bibliography, should prove useful .

Exercise 4.2

For those who prefer to save their soles, examine the ways in which
social class was m ade visible i n social research d uring the twentieth
century. For a starting point consult D. Rose ( 1995) 'Social classifica­
tions in the UK' at http://www.soc. surrey. ac.uk/sru/Srug.html

Notes

1 In this respect we can see how radically different is an electronic text from an
eighteenth-century printed text. In our example, the latter orders by alphabetic
sequence. In the former the use of hypertext links allows for nested and web-like
arrangement of content.
2 The International Psychogeriatric Association also provides a guide to the diagno­
sis and assessment of Alzheimer's disease on its web pages. Search for 'International
Psychogeriatric Association', and then for 'A1zheimers'.
3 Another suitable example would be that relating to the classification of viruses. See
Murphy et al. (1995) .
4 The research work was funded by the UK's Economic and Social Research Council
under its innovative health technology programme, grant number L21 8252046.
Ethical approval for the work was given by the MREC for Wales.
5 Rapp (2000) describes how chromosomes are, these days, made visible by modern
laboratory practice, arguing that chromosomes produced by cell cultures are 'not
natural objects but cultural ones' (p. 2 1 3).

Copyrighted Material
5
Texts , Authors , Identities

Authors. What use are they? 89

A dossier manufactures its subject 92
Autobiography as a technology of self 95
Sybil. A network of text and action 98
Identity as performed through writing - and reading 1.01.
Conclusions 1.03
Notes 1.05

Authors. What use are they?

In 1 76 1 Rousseau's fictional work, La Nouvelle Heloise, appeared. By 1 800 more

than 70 editions of the book had been published (Damton, 1 984) . The book's
full title is Julie or the New Heloi"se. Letters if two lovers, residents if a small town
at the foot of the Alps'. For a twenty-first century reader the novel has little to
attract it. As a book, it is overlong, it has no discernible plot, it rambles persis­
tently, and it contains no explicit reference whatsoever to sex or violence. Yet,
as Damton ( 1 984) suggests, and as the single statistic cited above would sup­
port, it had considerable audience appeal in its day. Indeed, the book seems to
have generated feelings of intimacy towards Rousseau on behalf of the readers
and to have resulted in considerable quantities of what, these days, would be
called fan mail. In short, readers came to regard the author of'Julie' as a friend
and confidant, and corresponded with him as such.
This brief reference to an eighteenth-century author and his readers is
instructive to us on a number of grounds. First, it suggests that the markedly
under-researched process of reading documents is as significant as the process
of writing (see, for example, Jackson, 200 1 ) . Certainly it shows that reading is
not quite the personal and individualistic activity that, at first consideration, it
appears to be, but rather a task undertaken within, and fashioned by, a certain
cultural milieu. 1 (Indeed, in some cultures we know that acts of reading are
highly regulated, as is often the case in societies that encompass a guru system.)
Secondly, our example suggests that both a text and its author are, to a consi,­
derable degree, constituted by readers. (Rousseau, was henceforth viewed as a

Copyrighted Material
90 USING DOCUMENTS IN SOCIAL R E S EARCH

man of great sensitivity, a 'new Socrates' and an 'apostle of virtue'.) And thirdly,
it suggests that the manner in which a document is integrated into a 'lifeworld',
or network of everyday action, forms a worthy topic for investigation in itself.
In this chapter I intend to pick up on a few of these themes (together with
some related issues) and explore them heuristically. Let us start with the author.
In the introductory chapter oblique references were made to the desire on
the part of writers such as Barthes ( 1 977; 1 990) and Foucault ( 1 979) to
de-centre the author. That is to say, a wish to displace onto other agents (such
as readers, intellectual associates and publishers) the creative and productive
functions that we normally associate with the author (or the author as subject) .
Questioning the reality of authors is a dangerous business - especially when
one is writing a book. Yet, the theme of de-centring does have virtue. This, not
least because it encourages us to consider the author-subject as a canopy that
is used to cover the background procedures of social interaction on which the
production of documents depends. The creative author is in that sense a con­
venient fiction, a fiction that enables us to abbreviate, or to short-circuit, an
examination of the essentially social nature of documents - see Chapters 1 and 2.
In the language of Latour (1 987) we can refer to the author as a 'black box'.
And it is a box that needs to be opened up and closely examined if we wish
to reveal anything at all about how documents are assembled, used and function
in everyday life.
It may also be as well to note that challenges to authorial magnificence are
only a part of a deeper and more general project. The latter is one that seeks
to dispossess the active, conscious 'subject' in general - much loved by human­
istic sociology, psychology and anthropology - of many of its supposedly pri­
vate 'inner' capacities. Indeed, some commentators have argued that the image
of a social scientific subject is in itself nothing more than a cultural configura­
tion. Thus, suggests Foucault ( 1 970: xxiii) , 'Man [is] no more than a kind of rift
in the order of things . . . a recent invention . . . a new wrinkle in our knowl­
edge' - and, we may add, a transitory presence. Indeed, with the encourage­
ment of Foucault, we have every entitlement to view the social scientific
'subject' as a de-centred phenomenon.
The de-centred author and the de-centred subject in general - whose
capacities are dispersed to other forms of agency - is an intriguing phenome­
non. We have glimpsed, for example, how it is possible to get a much fuller grip
on personal motives, intentions, reasons and explanations in the light of such a
creature. That is to say, we have seen how it is possible to view what are com­
monly regarded as properties of a secret, individual and personal consciousness,
as collective properties. For example, in Chapter 2 (in the context of suicide)
we looked, very briefly, at the manner in which it is possible to speak of voca­
bularies of motive and vocabularies of causation, pointing out how human
beings assemble - in writing and in speech - an account of their actions (and
the actions of others) from a range of culturally available 'good' reasons, 'worthy'
motives and 'sound and acceptable' explanations. In brief, motives, reasons,

Copyrighted Material
 TEXTS , AUTH O R S , I DENTITIES 91

impulses and dispositions emerge from a socially sanctioned store of such entities,
rather than from an inner store of psychological qualities.
The same might be said of beliefs (see, for example, Prior et al. , 2000) . Thus
what people relate to interviewers and other enquirers (when asked about
what they believe in) are not so much secret, inner states of knowing or believ­
ing, but socially structured accounts,2 and statements that they consider will fit
the investigator's bill. In a parallel vein one can argue that scientific 'beliefs' are
more properly viewed as being contained in scientific documents than inside the
heads of scientists. (This is an issue that we will return to in Chapters 6 and 7.)
Indeed, in all of these instances, it is plausible to assert that human dispositions
and qualities are more likely than not carried along by agents and processes
external to any individual, rather than within people's heads.
This is also the case with identities. And one form of media by means of
which identities are carried (and dispersed) is that of the written document (in
a myriad of forms) . In this chapter I hope to demonstrate how documents
actually serve in the processes through which subjects, subjectivity and identi­
ties are created and stabilized. In fact, and in line with the general tenor of this
book, I shall base my analyses on the premise that documents function not
merely as simple repositories of facts and detail (about subjects), but actively
structure the nature of subjects. Remember: documents are never inert, but
enter int'\ projects as independent agents. This, no doubt, sounds a most pecu­
liar claim to those brought up with the notion that only humans can be actors
or agents. So perhaps I should point out that it is an idea that I have borrowed
from the work of European sociologists such as Callon (1 989), Latour (1 983)
and Law ( 1 994) . In fact, writers of the latter kind have developed a mode of
analysis that is built around the notion of an actor-network. This is not per­
haps the place to expound on the nature or limitations of actor-networks nor
of actor network theory - ANT (see, for example, Law, 1 994; Law and Hassard,
1 999) . Nevertheless, it is the case that ANT is suggestive of a number of
possibilities that are pertinent to our purpose.
One such possibility concerns a focus on the function that a document,
thing or person plays in a network of activities. For example, in his analysis of
the attempts of automobile manufacturer Renault, to engineer an electric
vehicle, Callon ( 1 989) refers to an ensemble of entities in a network such that
it is the linkages between things that form the actor rather than any single item
within the ensemble. Thus, in Callon's example, the fuel cell, lead accumula­
tors, the French middle class, city councillors and so forth all entwine so as to
comprise an actor. In the framework of our concerns, this manner of viewing
things implies that non-human components can serve as active agents in an
actor-network. Such agents may, for example, impose restrictions on, or facil­
itate possibilities for, other agents - that is, structure the action of others. They
can also serve as agents in so far as they (as inanimate obj ects) are open to
recruitment by a network as an ally, or even expulsion from it as a traitor. Such
a reconceptualization of actors points towards the possibility of regarding

Copyrighted Material
92 USING DOCUMENTS IN SOCIAL R E S EARCH

documents as agents that can restrict and facilitate, serve as allies or foes,
become involved in systems of domination and subjection, make and unmake
the nature of the material world. In short, documents can have effects and in
so far as they have effects then they can be researched as part of a field - as
dynamic rather than inert phenomena.
In what follows I intend to investigate some of the ways in which human
subjectivity and human identity are tied up with documentation. I shall begin
by looking at the way in which subjectivity and identity might be contained
in and structured through records. I shall then move on to examine how per­
sonal identity can and has been shaped through the use of biographical tech­
nology. Following that I shall consider how identities truly belong to networks
of action. I shall end the chapter with a consideration of some of the ways in
which identities can be performed through writing.

A dossier manufactures its subject

Sometime during the m�rning of 3 June, 1 835, in the village of la Faucterie,

in the district of Calvados (northern France) , a young man by the name of
Pierre Riviere grabbed a pruning hook and used it to slaughter his mother, his
8-year-old brother and his teenage sister as they busied themselves with their
daily affairs. Some five years after these events, at the age of 25, Riviere was
found hanged in his cell in the prison at Beaulieu. Over a century later, a team
based at the College de France assembled the documents surrounding this
gruesome case. They were published during the 1 970s (Foucault, 1 978) . The
collected documents comprise a dossier. That is to say, a collection of state­
ments drawn from various sources - the courts, medical practitioners, a con­
fession written by Riviere himself, newspaper cuttings and prison receipt
books. These statements not only circle the events of murder but also seek to
define how the central characters relate to it.
In a stark and very direct way the papers that emerged from the unusual
happenings that occurred at la Faucterie mark out, 'a case, an affair, an event,
that provided the intersection of discourses that differed in origin, form,
organisation, and function' (Foucault, 1 978: x) . In short, the dossier serves to
mirror a number of mid nineteenth-century ideas about sanity and madness,
the normal and the abnormal, the reasonable and the unreasonable, and the
fathomable and the unfathomable. Indeed, the dossier and the documents
within it lay at the cusp of a period in which legal definitions of'responsibility'
and 'insanity' were being refashioned across the western world.
As a mirror on events, the dossier provides us with an outline of other
worlds and other ways of thinking. Initially, however, the dossier functioned
otherwise. It functioned in a network of forensic actions. For example, much
of it was written so as to serve as 'evidence' in a court of law. The Cantonal
judge thus took care to describe the scene of the murder on the morning of

Copyrighted Material
 TEXTS, AUTHORS, I DENTITIES 93

the 3rd, a description that was intended to stand in place of the physical reality
it described. The death certificates of the doctors who examined the bodies
were to stand in place of the corpses. The recorded statements of the parish
priest, and other character witnesses who spoke about Riviere, were to stand
in place of his persona, and so forth. The preliminary and subsequent inter­
rogation of the accused (which contains the record of the questions posed to
him and the answers obtained from him) took on a special status. For these
recorded statements - that, for example, the accused acted in accordance with
God's Providence - are awarded a power and a strength over and above that of
any verbal statements that Riviere might have made elsewhere. Indeed, the
'admission' to the murders is lodged in the dossier (there exists a signed con­
fession - the signature somehow confirming a truth that might otherwise be
considered provisional3) . Put together, the documents generate a narrative of
the events leading up to the murders, and define them as such (rather than acci­
dents or suicides) . One might argue that the dossier as a whole features as an
actant,4 and that the individual records within it function to implicate, accuse,
explain and identifY a network of human actor/agents involved in the murder.
Now, it is clear that things happened in la Faucterie independent of any
documentation - in the same way that billions of things and events occur in
the universe independently of human consciousness and reflection. Yet, the
documents surrounding the aforementioned murders add something to those
happenings. In a sense, they order the events and produce them anew. They
thereby give shape to phenomena that were (necessarily) disarranged. We
might even say that they translate occurrences into a 'murder' . In addition, they
serve to create the persona that was Pierre Riviere, and to define him as a sub­
ject of legal action (a murderer and a criminal) . This, particularly in so far as
they are directed towards the decision as to whether Riviere was sane or insane -
whether, indeed, he was, in the language of the day, a 'monomaniac' .
In such a context, i t is, perhaps, useful t o point t o a comparison between the
attribution of madness to Riviere and the recognition of mental illness in
Dorothy Smith's 'K' (Smith, 1 978) . In both instances insanity was read, by
observers, into quite ordinary and mundane activities that were variously
undertaken by these individuals. Riviere, for example, was once seen to attack
some cabbages, and to have buried a churning instrument and a Jay in a field.
He also liked frightening children. (Not such a bad man, perhaps!) 'K', on the
other hand, was observed to swim up and down a pool rather too exuberantly,
and to have talked in whispers when there was 'no need' . In all instances it was
the incorporation of these events into a series of observer narratives that manu­
factured abnormality (madness) . Naturally, both 'K' and Riviere acted in multi­
farious ways. They did a thousand things, but of those thousand things only a
few that indicate the presence of madness were highlighted and subsequently
recorded. On the basis of such selectivity both lay and expert diagnoses were
constructed. Indeed, biographical work of that kind is far from being restricted
to the manufacture of madness, and is ubiquitous in contemporary societies.

Copyrighted Material
94 USING DOCUMENTS IN SOCIAL RESEARCH

Thus public records in schools, hospitals, courts and welfare agencies can all
serve as basic materials for the construction of personal biography (Gubrium
et al. , 1 994), and, indeed, for the construction of identities.
Let us stop, then, to summarize how the documents within the Riviere
dossier actually functioned, and how they were integrated into a web of social
action. First, the documents comprise a readable space, they mark out the
limits of concern and attention for the various agents involved in the happen­
ings of which we have spoken - they mark out the boundaries of events and,
to a certain extent, boundaries of professional competence. Many documents,
of course, stood in place of events (the report of the Cantonal judge, for
example) . Others stood in place of bodies (the death certificates) . Some served
to define the nature of the murderer (the character witness statements) . Others
stood as a confession (the record of interrogation) . Individually and collectively,
the documents provided a structured narrative of what happened, albeit a
narrative of multiple authorship. The various author-subjects rewrote (fash­
ioned) the events and, at one time, their documents and oral statements func­
tioned as evidence in a court of law (for an interesting comparison, see Jonsson
and Linell, 1 99 1 ) . Further, the documents allocate and define responsibility for
the acts; they accuse. They also reflect back on their creators so as to define the
role of doctor, magistrate, witness - for 'this' is how doctors write and 'this' is
what they write about, and this is how magistrates/witnesses/jailers write and
so forth. Finally, they define the identity of the perpetrator (a deranged man
not driven by malice, nor witchcraft, but by madness) . These days, of course the
only reality of the murders was, and is, in the documents, and they thereby
enter into the world as 'records'. Should anyone ever wish to challenge the facts
of Riviere's guilt, they would have to challenge and counter the documenta­
tion. Indeed, at this distance all that we have is the documentation, but be
that as it may, once written statements are put into the world, they have a
solidity - in any culture - that is not easy to overcome nor to ignore. They
have, forever, to be accounted for, interrogated, recontextualized (for example,
as forgeries, as false statements, as forced confessions or whatever) , and subse­
quently integrated into new schemes of social action. Such is the destiny of any
dossier.5
We can see then that the identity of Riviere is intimately connected to
documentation. Who he 'is' - or was - is to be found as much in the docu­
mentation as it ever was in the live person. His human capacities and qualities
are dispersed across the written spaces. Indeed, his 'monomania' could only
ever be lodged in documentation - it was certainly never in his head - and so
too his other attributes. This capacity of documents to structure identities and
bestow attributes on human subjects was, of course, also evident in some of our
examples contained in Chapter 3. Those examples suggest that identities are
commonly dispersed across records held in schools, hospitals, prisons and other
organizations. However, it is seldom that such records are ever brought together
to provide a unified image of a single person. (Our references to Mental Health

Copyrighted Material
 TEXT S , A U T H O R S , I DENTITIES 95

Inquiry Reports in Chapter 3 provided a rare insight into such a practice.) Yet,
western culture does have a mechanism for manufacturing coherent and uni­
fied identities, and it is to a consideration of such mechanisms that we now turn.

Autobiography as a technology of self

Whilst in prison Riviere wrote a 'Memoir'. It is full of grammatical errors,

errors of punctuation, spelling and so forth, and was printed as such in 1 835.
Indeed, it is interesting to speculate as to why such errors were left in the
manuscript by the original printers, and why Foucault ( 1 978) saw fit to 'cor­
rect' them at a later date. Was it that in the first case the printed page conveyed
an image of an uneducated peasant scribbler (a brute) , and in the second case
an image of a reasoning being? It is not for us to decide on such issues, but
merely to note how the 'I' of Riviere has been instantly mediated through the
work of others. Indeed, more relevant from our standpoint is the fact that the
Memoir provides an autobiographical account of a peasant life,6 and the narrative
opens by Riviere defining himself in terms of his murderous acts, thus:

I, Pierre Riviere, having slaughtered my mother, my sister, and my brother, and wishing
to make known the motives which led me to this deed, have written down the whole of
the life which my father and my mother led together since their marriage. (Foucault,
1 978: 54-5)

This use of an autobiographical memoir to unfold the personal history of an

'I' is itself a specialized form of literality (de Certeau, 1 984) . And it is, perhaps,
notable that the ' I ' (ego) of the philosopher, the self-portrait of the painter, and
the entire genre of biography and autobiography all seem to emerge - albeit
gradually - out of a specifically western culture from the sixteenth century
onward. These days, of course, everyone is expected to construct some form of
autobiography - to indulge in identity creation and to specify who one 'is'
through documentation - even if it is only at the level of a curriculum vitae.
Indeed, Castells (1 997) argues that identity creation and maintenance figure as
key strategies in what he refers to as the network society, though it is only in
a culture that is conscious of self, subjectivity, and individuality that autobio­
graphy has any place and purpose.7
The substantive contents of biographies and autobiographies are, naturally, of
considerable interest, but what I wish to do here is focus, briefly, on how bio­
graphy and autobiography function, rather than with what they contain. In order
to follow that process let us return to the first line of Riviere's Memoir, above.
We can see at once that the Memoir is, in part, intended to provide a sense
of unity to the writer's past (he refers to the 'whole' of a life). It is also designed
so as to reveal the workings of a secret, inner self (his motives and dispositions) ,
and executed so as to provide an ordered narrative of events (that led up to the

Copyrighted Material
96 USING DOCUMENTS IN S O C I A L R E S EARCH

murders) . Such narratives usually unfold in terms of a linear sequence - a

chronology - though we cannot, of course, deduce that from the initial
sentence. However, we can make a reasonable guess that Riviere's tale is highly
selective in its focus. The 'I', as ever, is situated so that the author emerges as
that person that he elects to write about and nothing more. (He is, from begin­
ning to end, a person that slaughtered his mother, his sister and his brother.)
Clearly, Riviere's Memoir (autobiography) is a work of construction. It
involves the exercise of what has been called biographical work so as to ensure
the 'storying' of a life (Bruner, 1 987) . In fact, autobiography is indispensable to
the construction of a unitary self (Sacks, 1 986: 105) . For where else might one
find such unity? That question is one that was posed and answered - rhetori­
cally - by Bruner (1 993: 38) and Barthes ( 1 977) . Such biographical work need
not, of course, be as detailed as that undertaken by Riviere. Indeed, it might not
stretch far beyond a brief c.v. Or, for actors and singers, a brief career outline in
some programme notes, or even a brief verbal account of whom one 'is'.
Verbal accounts of identity are plausible over the shorter term, but to rein­
vent oneself thoroughly, documentation is essential, whether it be through the
production of identity cards, passports, degree certificates, electronic number
systems, or whatever. For we know of many cases where a documented iden­
tity can take precedence over a lived identity. (For example, in the case of trans­
sexuals where the sex stated on a birth certi.ficate can override the gendered
identity of the person in life - or even in death.) 'False' and stolen identities,
of course, are almost entirely dependent on documentation.
Providing a sense of unity of self, constructing or reconstructing an identity,
situating oneself in an unfolding narrative of events, are only some of the things
that can be achieved by an autobiographical account. For autobiography (and
even biography) , in the modern world, form part of the technology of self
(Foucault, 1 988) , a technology that enables one to unburden oneself of inner
conflicts (Rousseau's Confessiol1s, 2000); justify one's thoughts and actions
(Marshall Zhukov's Memoirs, 1 97 1 ) ; construct 'identity' anew (Wilkomirski, 1 996;
Machler, 200 1 ) ; or manufacture a philosophy of life O.s. Mill's Autobiography,
1 989) . Yet, no matter how sophisticated the purpose, it is also the case that iden­
tity always depends on the existence of some 'other' . The other for whom an
autobiography is written is rarely immediately evident, but nevertheless always
present. In the latter sense it is, perhaps, biography rather than autobiography that
makes visible the social relations on which human identities are dependent. And
biography is certainly more interesting in that respect. For a biography is an
account of a life as manufactured through an interactive process between writer
and subject. (In some rather nimble cases, of course, the writer explicitly con­
structs him- or herself in the process of constructing another - as did Boswell
(1 976) in recording the life of Dr Johnson.) The emergence of an identity
through a network of action is something that we will consider in a later section
of this chapter. For now we need only note that autobiographical writing is an
expression of a cultural rather than a personal process.

Copyrighted Material
 TEXTS, AUTHORS, I D ENTITIES 97

The creation of identity through documentation is, of course, something

that social science has itself exploited. In that respect, Foucault's recruitment of
Riviere's Memoir serves to emphasize a strong trend in twentieth-century
social science that has consistently made extensive use of such accounts to further
its investigative aims. In fact, the 'life story' as a social scientific resource has
appeared in numerous sociological and anthropological contexts.
For example, researchers in the so-called Chicago tradition (or school) of
sociology gave voice to many human actors (Platt, 1 996) , though, as in the
Foucault case, 'delinquents' and criminals figured largely in such accounts (see,
for example, Shaw, 1 93 1 ; Sutherland, 1 937) . Anthropologists on the other hand
normally sought out autobiographies of those who had played more traditional
roles. Thus Simmons (1 942) , for example, elected to publish the autobiography
of Don Talayesa, a Hopi Indian. And according to Kluckhohn's ( 1 945) review
of life history method in anthropology it was fairly typical of the genre - a
focus on just one person (usually male and usually over 50 years of age) in just
one tribe. Such themes persisted well after the 1 940s - as with Casagrande's
(1 960) omnibus of around 20 anthropological life histories of 'native' life.
Unlike the Memoir of Riviere, however, the twentieth-century accounts have
often been directly coaxed, encouraged and thereby filtered through the inter­
ests of social scientists themselves, which is why the term 'life history' rather
than biography is used in anthropology (Langness, 1 965) . Indeed, in the hands
of social scientists autobiographical and biographical sequence is allowed to
unfold only through a conceptual grid of the expert observer. Thus a personal
reflection is turned into an autobiography of'a professional thief' (Sutherland,
1 937) , or of a Hopi Indian (Simmons, 1 942) , a hermaphrodite (Garfinkel,
1 967) , a Polish emigre to Chicago (Thomas and Znaniecki, 1 958) , all of whom
become the object of social scientific puzzlement. These 'others' invariably
emerge only through the categorizing language of the social scientists - con­
cerned as they are with gender, ethnicity, class or life style (and sometimes
through concepts of colonization and oppression) . In fact, the identities and
selves that fascinate social scientists are routinely filtered through the changing
interests, concepts and concerns of the age (see, for example, Smith and Watson,
1 992; and Stanley and Morgan, 1 993, on gender) .
Further, although life reports are normally presented as first-hand and indi­
vidual accounts of personal experiences, it is clear that (as was pointed out in
Chapter 1 ) such texts are always the product of collective action - of editors,
data gatherers, publishers, translators and so forth. In some cases, of course, the
researcher as data gatherer is openly aware of the role that he or she has played
in creating a life. Thus Crapanzano (1 980) in his study ofTuhami points out
how he became, 'an active participant in his [Tuhami's] life history, even though
I rarely appear directly in his recitations' (p. 1 1) . In other accounts, however,
the researcher often claims to be nothing more than the instrument of the bio­
graphical subject. In this context, for example, it is interesting to read Burgos­
Debray's (1 984) claim that, ' I became what I really was: Rigoberta's listener. I

Copyrighted Material
98 USING DOCU M ENTS IN SOCIAL RESEARCH

allowed her to speak and then became her instrument . . . by allowing her to
make the transition from the spoken to the written word.' (1 984: xx) .
Rigoberta was billed by Burgos-Debray as an ' Indian Woman in Guatemala',
and the biographical text is entided, 'I, Rigoberta Menchu'.Yet, just who that
'I' is, is a matter for some debate - the final text was edited and translated by
Burgos-Debray. (On the methodological issues surrounding the Rigoberta
Menchu story and how the account can be variously used, see Beverley, 2000.)
A similar question about the 'I' of autobiography is also raised by Klockars
( 1 975) in his account of how he obtained textual information from his fence,
Vincent.
As with ethnographic reports8 (see Chapter 7), it is usually of interest to ask
questions as to how the reporting subject is structured by the documents
through which they speak. For example, in the case of Riviere (and to a lesser
extent Rigoberta Menchu) the subject is structured as an isolated and untar­
nished reporter on self - as an active narrator rather than as a witness or infor­
mant. In the case ofTuharni the subject is recognized to be a construction of
the subject plus researcher. In social science, of course, one can always find rules
and rule-books about how researcher and researched ought to position them­
selves in their enquiries. And ever since the life story was objectified as a dis­
tinct social scientific 'method of enquiry' there has appeared a stream of
publications on how the method should be operationalized and used. Many of
the issues here discussed were explored in a special issue of Sociology (Stanley
and Morgan, 1 993) , and other statements on the method are to be found in
Atkinson (1 998), Denzin (1 989) and Watson and Watson-Franke ( 1 985) .
Plummer (200 1 ) provides an excellent overview of the history of the method
and its source materials.

Sybil. A network of text and action

The examples mentioned above give vent to another important issue in matters
of identity. In particular, they point to the possibility of viewing identity as a
relational property rather than a personal one. That is, a property that emerges
out of the relations between actors rather than one that emerges out of the indi­
vidual alone. In this section we need to examine that possibility a litde more
closely. So, in what follows we are set to investigate how a specific identity was
brought into the world through a network of action and text (of various kinds) .
In 1 974 F.R. Schreiber published a book entided Sybil. It concerned a
woman who claimed to have 1 6 personalities. Sybil was not, of course, the first
'case' of multiple personality. Hacking (1 995) lists a number of precursors,
though probably the best known must be that nice Dr Jekyll, and that rather
nasty Mr Hyde (a story first published in 1 886) .9 Sybil's notoriety, however, lay
in the fact that she was the first person in which the emergence of multiple
personality was linked to sexual abuse in childhood (Hacking, 1 995).

Copyrighted Material
 TEXTS, AUTHORS, IDENTITIES 99

Sybil was born in the mid-west USA during the 1 930s. She had an unusual
childhood, but no memory of any form of sexual abuse as 'herself'. When
hypnotized by her psychoanalyst - Cornelia Wilbur - however, she recalled
episodes of abuse by her mother. During the 1 1 years and 2,354 psychoanalytic
sessions in which Dr Wilbur treated Sybil, some 1 6 different personalities
were revealed - and, eventually, a 1 7th appeared when all of the previous selves
were fused into one and Sybil was 'cured'. The case was written up by
Schreiber (1 974) and became a best-selling book. It was also turned into a film
drama in 1 976. Following the Schreiber book a number of others appeared on
the theme of multiple personality. The appearance of multiple personality dis­
order (MPD) as a disorder in DSM-III (see Chapter 2) in 1 980 is far from
being unconnected with this narrative. (MPD does not, however, appear in
DSM-IV, though a condition referred to as dissociative identity disorder does
appear.)
So the Schreiber biography not only created an identity - in Sybil - but also
gave the concept of identity a new dimension. For previous to the Sybil case,
the notion of a multiple identity was rare, and certainly did not figure in the
caseloads of most psychoanalysts - as the absence of the condition from the
DSM testifies. (Indeed, and even at a common-sense level, one wonders about
the meaning of the word 'personality' when it has to share space with some 1 6
others.) The Sybil case also provided a fillip to those who wished t o trace adult
psychological problems back to childhood sexual experiences. In the wake of
the Sybil case thousands of Americans were diagnosed with MPD. Most of
these individuals also managed to 'recover' lost memories of abuse. 1 0
One of the analysts who met and treated Sybil - Dr Herbert Spiegel - was,
however, not so convinced about the veracity of MPD. Spiegel (1 997) has
reported that, based on his experiences, he found Sybil (who was at one stage
diagnosed as 'schizophrenic') to be highly suggestible - easy to hypnotize and
easy to influence. He also reports that Sybil had read a book on a case of mul­
tiple personality called The Three Faces <if Eve (Thigpen and Cleckley, 1 957), and
had been deeply impressed by it. She had also seen the 1 957 (Nunnally Johnson)
film of the same name. In fact, states Spiegel, Sybil had already encountered a
template of the disorder before she displayed the symptoms of it. In addition, he
claims that Sybil's analyst had probably suggested to Sybil that she was a multi­
ple. (At one stage, Sybil wrote a letter to Wilbur refuting the facts of any child­
hood abuse or of any multiple personality, but Sybil's plea was regarded in true
Freudian fashion as 'denial' (Schreiber, 1 974: 374) 1 1 .) Indeed, Spiegel argues
that, in psychotherapy 'we become engaged in storytelling and we impose our
hypothesis on the patient by the way we ask our questions. Highly suggestible
people will of course respond in a way that can please the doctors' ( 1 997: 62).
So MPD, in his view, was an invention of the psychoanalytic interaction
between these two particular individuals, and Spiegel has gone further to argue
that patients, generally, have been trained (rather than diagnosed) by their thera­
pists to become multiples. (Hacking (1 995) advances a similar argument.)

Copyrighted Material
100 USING DOC U M E NTS I N SOC I AL R E S EARC H

Whether Spiegel is correct or not, it is clear that what we see here is a

process of identity creation. It begins with patient and analyst, and involves
them as readers (of various texts including The Three Faces of Eve, the DSM,
professional journal articles and so on) , and, possibly, as moviegoers. In fact, the
identity of MPD and of Sybil emerges out of and becomes lodged in a net­
work. Books, films and people are progressively recruited into the network and
with each act of recruitment the identity becomes stronger.
We can also see how the circuit of recruitment expands the network so as
to incorporate a new book (on Sybil) , a new film (Daniel Petrie, 1 977) and a
new disorder, and that circuit, in turn, provides an impetus for other books and
accounts of multiples. Our interest, of course, is truly in the role that docu­
mentation plays in the network and circuit. In that light one is inclined to
argue that MPD is truly in the documents rather than in any human being.
That is where it is created and structured, and it is the documents (books and
films and manuals) that provide the template and foundations for the condi­
tion. Even the experience of MPD is measured through a document - the dis­
sociative experience scale (Hacking, 1 995: 96-1 1 2) . Like our examples in
Chapter 4, the scale is an instrument that makes 'experience' visible. Indeed,
the disorder is woven through text in every way. So individual human beings
provide only instances (with peculiar variations) of the pathology. 'Subjects' are,
in this sense, not to be confused with individuals. The former are manufactured
in documents - in text - as cases; individuals (human beings) are subsequently
recruited to fit the templates in situ - as Sybil was.
It is somewhat paradoxical, then, that having made a claim in the previous
section to the effect that autobiographies and biographies unify a sense of self,
we have ended with a narrative of many selves in one human being. The Sybil
case, however, serves to emphasize how subjects and subjectivity are manufac­
tured through collective processes - including processes of documentation.
They emerge out of networks of action. Thus, Sybil's 'personality' was con­
structed by others, and it was garlanded with features and activities that even­
tually comprised a 'case'. That case was then open to recruitment and
manipulation in various contexts - as is often so in medical practice. This is at
its clearest in matters of psychiatry. Indeed, psychiatric and other illness identi­
ties and their personal histories are, as we have seen, routinely created in medical
notes - Sybil, you will have noted, was once identified 'as a schizophrenic'
(Spiegel, 1 997: 6 1 ) . 1 2
One final point: we can also begin to see in the Sybil case how an identity
is a matter of performance (Goffinan, 1 959) . That is, a matter of following,
interpreting and acting out a script. In that sense we might say that Sybil
became a person with MPD in much the same way as Becker ( 1 953-4) had
suggested that one could 'become' a marijuana user. Naturally, the script for
such performances is lodged in various media and not just in writing. In the
next section, however, we shall see how a performance can indeed be lodged
in script alone.

Copyrighted Material
 TEXTS , A U T H O R S , IDENTITIES 101

Identity as performed through writing - and reading

In a paper published in 1 993, Miller and Morgan examined the social nature
of the academic c.v. In that paper they indicated how such c.v.s can be under­
stood as a form of autobiographical practice - a form of practice that is centrally
concerned with the presentation of self in occupational settings. Using
Goffinan's aforementioned ( 1 959) work they seek to demonstrate how the
academic c.v. involves matters of presentation of front and manufacture of
self. Above all, however, they point out how the academic c.v. is a matter of
performance.
Performances of the kind just referred to are, of course, executed solely
through writing. In that respect they are of special concern to us. It is also of
interest to note that Miller and Morgan indicate how the construction of a c. v.
can itself engage the use of generative documents o( the kind mentioned in
Chapter 2. In this instance the generative document referred to was the
Guidelines for the Presentation if a Curriculum Vitae of Manchester University.
Rather than look at how human actors can use and manufacture documenta­
tion for the presentation of the self, however, I propose in this section to look
at how documents can structure their readers. That is, to look at how they
might guide readers to perform in specific ways.
In this context it is helpful to note Iser's (1 989) claims about how the reader's
role in the work of English fiction altered between the eighteenth
century and the twentieth. We need not attend to the detail, but simply note
examples of the manner in which texts provide the reader (and the writer)
with specific scripts for performance. For example, Iser points out how
fictional texts have structured the reader as a passive recipient of a narrative or -
as with Thackeray'S Vanity Fair - as an observer of life's variety and display. With
Charles Dickens, on the other hand, the reader is almost invariably structured as
a judge of moral character. This, nowhere more so, perhaps, than with Oliver
Twist, a text in which we are inveigled into taking up a moral stance on poverty
and exploitation, subservience and domination, whilst the writer remains posi­
tioned as a dispassionate teller of a tale. That stance of reader and of writer is, of
course, quite different from, say, the one adopted in Dashiel Hammet's The
Maltese Falcon, a story that features the unforgettably amoral Sam Spade.
Naturally, our concern is not with works of fiction but with routine forms
of documentation. In that respect, we have already glimpsed, in Chapter 4, at
how a work of non-fiction (the encyclopaedia of fungi) can structure its read­
ers. There we noted how the reader was structured as an amateur and how that
amateur was directed to identifY fungi in particular ways. But even with every­
day documents readers and writers are structured and allocated to specific roles
(see, for example, Abelen et al., 1 993) . To this end I intend to examine, albeit
briefly, a form of documentation that is provided to people in the UK when
they collect medicines from a pharmacy. In each case where that happens the
client should be provided with a Patient Information Leaflet (appropriately

Copyrighted Material
:1.02 USING DOC U MENTS IN SOCIAL RESEARCH

abbreviated to PIL) . For each form of medication there is also a 'data sheet' for
pharmacists. It will serve our purposes to analyse the text for both forms
of document. We will do so with respect to leaflets that relate to one and the
same drug.
The PIL to which I am going to refer is that for ziduvodine. The latter is
commonly prescribed (under a trademarked name) for people who show symp­
toms of HIV infection and AIDS. The PIL for ziduvodine opens by highlight­
ing the possibility of the taker of the medication having doubts and worries.
Should the 'patient' have any doubts it is suggested that he or she consult his
or her doctor. And this pattern of the doctor as an expert (with full knowledge,
but without doubt and worry) and the consumer as in need of advice is repli­
cated throughout the leaflet. Consequently, the doctor is presented as someone
who will assuage doubt and worry, someone who can provide advice on all
aspects of medication, and as someone who will give instructions that ought to
be followed.

If you forget to take a dose, don't worry . . . If you take a larger dose than prescribed . . .
you should let your doctor know as soon as possible if this happens . . . I f someone else
takes your medicine by mistake, tell your doctor at once . . . You should not stop treatment
unless your doctor tells you to. (ABPI 1 996)

In line with the structuring of the patient as a non-expert, the leaflet also pro­
vides a lay narrative of the manner in which the drug acts on the body. The
narrative indicates how the anti-viral agent that is ziduvodine serves to delay
the progression of HIV and AIDS but does not provide a 'cure'. Instead it
'fights' against HIV For, according to the leaflet, HIV - left unchecked - will
enter into a group of cells called CD4 cells and turn them into a 'mini factory'
for infection. But in the great struggle against HIV, the anti-viral agent also
seeks to invade the CD4 cells and stop the factory producing viral agent. (It is,
if you like, an heroic narrative.)
In the respective data sheet for pharmacists, however, a somewhat different
narrative is provided. There one can read of the fact that ziduvodine is 'phos­
phorylated to the monophosphate derivative by cellular thymydine kinase',
that it is 'catalysed', and that it acts as 'an inhibitor of, and substrate for, the viral
reverse transcriptase' . Naturally, one would both hope and expect the data
sheet for pharmacists to read somewhat differently from that for patients - so
as to provide more complex information, for example. But what is important
from our standpoint is that in using this complex lexicon of chemistry, rather
than metaphors of mini factories, the texts are also structuring experts and
non-experts. And this structuring of expertise is also evident in more direct
ways. Thus the pharmacist's data sheet talks, for example, about 'the manage­
ment of patients', and how patients ought to be cautioned, advised, monitored
and so forth. All in all, then, the PIL structures the reader as an individual subject
to the direction of others (who are experts) , whilst the data sheet structures the

Copyrighted Material
 TEXTS , AUTHORS, I DENTITIES 103

reader as an expert in control of patients. None of this tells us, of course, how
such information might be used in practice - but that is a topic that lies well
beyond the boundaries of this particular chapter. 13

Conclusions

Some decades ago Harvey Sacks ( 1 984; 1 992) referred to the work involved in
'being ordinary' . Being ordinary demands work. For, to be ordinary is to be
unexceptional. That in turn demands a detailed knowledge about the routine
and the unusual - in forms of behaviour, speech, manner, dress and so forth.
And using such knowledge, we all undertake work (and repair work) on
managing and constructing our identities throughout our lives. In our termi­
nology, of course, being ordinary is clearly a matter of performance - as is the
matter of constructing autobiographies and biographies.
With that last claim in mind, I have attempted, in this chapter, to indicate
how a large part of identity work, as performance, involves documentation. In
some cases that documentation may take the form of full-length books -
biographies and autobiographies. More likely it will take less notah>le forms -
assembling and writing a curriculum vitae, or filling out a job application, a
census form, an application for a driver's licence or a passport. And exactly how
people use and manipulate such routine forms of documentation to do iden­
tity work is a legitimate field for research in itself.
Yet we have also noted that identities are not simply and necessarily con­
structed by individuals. Rather, they are assembled through networks of action.
And as in the cases of Sybil and Riviere, both human and non-human agents
are normally involved in such assembly work. This is so for all of us. Other
humans write reports on us, they translate our wishes, desires and motives - in
police reports, in hospital files, in school and welfare agency records, in social
scientific studies. They also allocate us to categories - good mothers, responsible
citizens, devious so and sos, or whatever.
Finally, we have noted that, as interesting as the reports of others may be,
more interesting still is the manner in which documents can themselves struc­
ture identities. For, documents invariably structure their readers. They do so by
offering fields into which such readers can slot themselves as moral beings, or
meticulous individuals, or seriously ill people. In that sense we can view docu­
mentation as offering forms of identity that are forever open to colonization.
Our specific examples drew on work relating to MPD and other illnesses, but
there are of course endless types of identity awaiting colonization. Indeed in the
modern world we can, perhaps, begin to talk about modern forms of identity
in terms of'cyborgs'. That is, as a hybrid of humanistic and informatic qualities,
where skin and bone, PIN numbers, electronic passwords, DNA sequences and
blood samples are enrolled as a single actant - text and body merged into one.

Copyrighted Material
104 USING DOC U M E NTS IN SOC IAL R E S EARCH

In his rather overlong and detailed account of life in a TB sanatorium (The

Magic Mountain) ,Thomas Mann presented an account of the ways in which the
category of , invalid' might be created. Mann points out how documentation of
all kinds plays its part in such identity construction (X-ray images, temperature
charts, GafIky scale scores and so forth), but he also notes that identity is a
matter of time and place. In that sense, perhaps, Mann's novel serves to remind
us that ultimately it is the role of documentation in context that should really
concern us, rather than the role of documentation in the abstract. How people
use documents to create and sustain identities in concert, rather than with how
documents contain IDs, is the central issue for social research.
So let me sum up the key points of this chapter:

• Authorship of documents is best viewed in terms of effects rather than in terms

of subjective 'identity'. How the author functions is the question that matters,
rather than who he or she 'is'.
• Identity is a matter of dynamic performance rather than of inert, personal quali­
ties. Documents have a role in that performance and, as such, should always be
considered as more than mere containers of facts about identity.
• Documents structure performance (identity) - this is so for both writers and
readers.
• How exactly a document structures a performance - how it is used in every­
day practice - is a key question for social research.

With respect to the above, it is the anthropology of use, more than the literary
study of content, that should guide the social scientist in matters of research
into 'documents of life'.

RESEARCH EXERCISES

Exercise 5.1

Documentation is crucial to the presentation of self. The c.v. (and in the

USA, the resume) are key tools for the m a nufacture of a professional
identity. Execute a web search using 'curriculum vitae' as a search
term. Locate pages of rules about how to assemble a c.v. (university
pages are often useful for thi s kind of exercise - see, for example,
http://www.columbia.edu/). Note what items of i nformation are recom­
mended for inclusion in the c.v. and what kinds are suggested for exclu­
sion. Then check your findings against personal details provided in a
' lonely hearts' column of a magazine. Recall that 'compare and
contrast' i s a useful strategy for analysing small samples of data such
as are likely to be gathered here. M a ke an assessment as to what the
results tell us about the nature of identity in the modern world.

Copyrighted Material
 TEXTS, AUTHORS, IDENTITIES 105

Exercise 5.2

Constructing identity is often undertaken for fraudulent reasons, and

identity theft has emerged as a growing twenty-tirst-century problem.
Consult the US government web pages about identity theft at
http://www. usdoj.gov/criminaljfraud/idtheft.html. Fol lowing that,
consult some non-governmental pages on the problem. Draw up a l i st
of documents and document content that identity fraud commonly
involves. Then draw some conclusions about what the l ist tells us
about the nature of 'doing' identity in the third mi llennium.

Notes

1 Hall (1 972) illustrated how what he refers to as 'texts' often encode for a culturally
'preferred reading', and that acts of ' decoding' (reading) are also restrained and
fashioned by the imperatives of the reader's culture.
2 One of the earliest discussions of 'accounts' was provided by Scott and Lyman
(1 968). They claimed that accounts were invoked when a social actor was called
upon to explain untoward behaviour. In this book, however, it is argued that any
call for an explanation or reasoning is a call for an account.
3 For some insights into the significance of signatures see Derrida (1 977).
4 A concept borrowed from Greimas (1 987), though used somewhat differently here.
An actant is a construct to which we can attribute functions and effects. One and
the same actant can be composed from various actors - so the author-subject
(actant) is, for example, built up out of many individual actors and their actions.
Greimas ( 1 987: 1 07) suggests that the converse - one actor being composed of
numerous actants - should also be considered.
5 Nowadays, of course, there are numerous other media that can hold and structure
confessions, statements, details and judgements. Indeed, for the first time in history
the twentieth century witnessed the human voice unhinged from the human body,
and lodged on mechanical and electronic tapes and discs. And so audiotape or disc,
as well as film, may serve as a 'dossier' on events, and thereby enter into fields of
action as actants - to function as accusers, witnesses, confessors, creators of narra­
tive and recorders of events.
6 The confessional form adopted by Riviere is often referred to these days as 'testi­
monio' (Beverley, 2000). The status of such testimonio as first-hand accounts is
often complicated by the fact that - as with Riviere's statement - they have to be
edited and/or translated by professional others. They subsequently raise a whole
series of fundamental questions about the nature of authorship.
7 Hardman (2000), for example, indicates the emphasis that is placed on relational
properties of identity as against the individual and personal in at least one non­
western culture.
8 Ethnography and autobiography are sometimes interlinked into what is sometimes
referred to as 'autoethnography' (Ellis and Bochner, 2000).
9 Hacking ( 1 995: 5) suggests that the first person recognized as a multiple appeared
in Charcot's Paris clinic in 1 885.

Copyrighted Material
106 USING DOC U M E NTS I N SOCIAL R ES EARCH

10 http://www.bbc.co.uk/horizon/mpd_script.shtml
11 Such recontextualization of client language is a common feature of professional
discourse and, among other things, serves to underpin the power of professionals -
see, for example, Sarangi ( 1 998) .
12 O n the relationship between MPD and schizophrenia i n the history o f psychiatry,
see Hacking ( 1 995) .
13 For a report o n a study o f drug users as lay experts and the manner i n which
chemical compounds are talked about in everyday life see Monaghan (200 1 ) .

Copyrighted Material
6
Content, Mean i ng and Refe rence

Mind, meaning and interpretation 107

Meaning and modernism 109
Referencing death 113
Referencing metaphors 115
Accounting for fatigue 116
Conclusions 122
Notes 123

Mind, meaning and interpretation

In the summer of 1 880 a 2 1 -year-old resident of Vienna - later referred to by

the pseudonym 'Anna 0' - began to exhibit a range of physical and psycho­
logical symptoms of a quite stunning variety. When he reflected on the case,
Ernest Jones (1 964: 202) rather aptly described her as presenting a 'museum of
symptoms'. The museum included headaches, squints, vision problems, paraly­
sis of the limbs and neck muscles, hallucinations, marked swings of mood and
self-starvation. Anna would variously refuse to speak in her native (German)
language (using English instead) , indulge in self-hypnosis, threaten suicide, and
exhibit bizarre patterns of behaviour. By December of 1 880, she had become
so ill that she 'took to her bed', permanently, and one of the doctors who
became involved in her treatment was a certain Dr Breuer. The latter was a
colleague of Freud, and the two men eventually wrote up the case of Anna 0
(together with case material relating to four other women) in a publication
entitled Studies on Hysteria (1 895) .
Our interest in Studies is multiple. For a start it is a book that is often recog­
nized as the very first psychoanalytic treatise. Secondly, it is a book in which a
number of key Freudian terms make their initial appearance. For example, the
first psychoanalytic use of the unconscious occurs in Studies (see Breuer and
Freud, 1 974: 1 00) . The terms abreaction, catharsis ( 1 974: 59) and repression
( 1 974: 6 1 ) also see their debut. Thirdly - and closely related to our last point ­
Studies redraws the geography of the mind so as to include among its elements
not simply conscious, but also subconscious thoughts and beliefs. And in the

Copyrighted Material
108 USING DOCUM ENTS I N SOCIAL RESEARCH

theoretical prolegomena of Freud and Breuer's Studies, it is argued that

subconscious tensions, left unresolved, become capable of producing symptoms
of anxiety and hysteria in individuals. In short, it is argued that deep within the
individual human psyche are repressed energies that can, under certain condi­
tions, manifest themselves as physical illnesses. The paralysis, vision problems
and generalized anxiety of Anna 0 were, for example, deemed to be traceable
to mental operations. In the Freudian universe, therefore, the task of the psycho­
analyst became one of unravelling the associations between inner thoughts and
external symptoms - of interpreting the meaning of signs and symbols so as to
effect a cure. Indeed, some years after writing Studies, Freud was to remark that
the patient's or (analysand's) weapon against neurosis was talk, whilst the ana­
lyst's weapon was interpretation (Gay, 1 988: 298) .
We have here, then, a whole series of terms and corresponding issues to
unravel. Interpretation is one such. Understanding other minds is another. The
practices of reading and referencing in general also necessarily figure. In fact
this vignette of Anna O's problems and their therapy ties up in one bundle
almost all of the essential issues that accompany those who seek to study the
meaning of texts and human actions. (For our purposes, Anna O's symptoms
can, for the time being, be viewed as a text.) In that light we can ask certain
questions about the narrative that I have just constructed, such as questions
about the 'meaning' ofAnna O's squints, headaches and other symptoms; ques­
tions about where such meaning might be found (is it to be found inside
Anna's head or in Freud's head?); questions about how we are to interpret such
meaning (is there, for example, a rule-book that links symptoms to mental
problems?) ; and questions about whether there is, in any case, a fixed meaning
to symptoms and signs, or whether meaning is necessarily fluid and flexible.
Above all, perhaps, we can ask questions about whether we should bother with
meaning at all.
Formal questions relating to the meaning of text and how it is to be known
have been regarded as fundamental for some centuries. Such questions are
commonly associated with the science of hermeneutics. ' Unfortunately, the
answers to hermeneutic problems are by no means easy to formulate. Nor,
from the point of view of the empirical researcher, are the answers that useful
once they have been located. In some circles, of course, it has been argued that
a focus on the meaning of text should be ignored in favour of some other charac­
teristic.Thus,Wittgenstein ( 1 958: 69) , for example, always argued (in relation to
words) that rather than ask for the meaning of a word we ought, rather, to ask
how it is used in ordinary language. That, as we have seen, is a productive prin­
ciple to adopt, and I shall return to it below. In the structuralist corner, of
course, writers such as Levi-Strauss (1 969: 84) consistently argued that the
focus ought always to be on how what is said is arranged, rather than what it
means. (Though structuralists, in general, remained as beguiled by the notion
of meaning as were the interpretationists2 - see Dosse, 1 997a, 1 997b.) Never­
theless, both of these alternatives have useful and practical implications for

, Copyrighted Material
 CONTENT, MEANING AND REFERENCE 109

social scientific research on documents. Hermeneutics, by contrast, offers little

in the way of practical guidance. Given the significance of hermeneutics to the
study of text, however, it would be amiss of us not to consider some of the basic
implications that the study of meaning has for the social sciences. To that end,
in the section that follows I shall sketch out the various ways in which the
problem of meaning has been approached in social scientific discourse.

Meaning and modernism

Let us begin by considering two quotations. The first is taken from E Scott
Fitzgerald's The Great Gatsby ( 1 925), the second from Leo Tolstoy's War and
Peace ( 1 889) .

Daisy began to sing with the music in a husky, rhythmic whisper, bringing out a meaning
in each word that it had never had before and would never have again.

'The road to Warsaw perhaps', Prince lppolit said loudly . . . Everybody looked at him, at
a loss to guess what he meant . . . He had no more notion than other people what he
meant by his words. In the course of his diplomatic career he had more than once noticed
that words suddenly uttered in that way were accepted as highly diverting, and on every
occasion he uttered in that way the first words that chanced to come to his tongue. 'May
be it will come out alright', he thought, 'and if it doesn't they will know how to give
some turn to it'.

The two quotations emphasize the importance of meaning in human affairs,

but they do so in different ways. Thus, the first provides a succinct example of
the modernist stance - the notion that meaning is an inner psychological pro­
perty of the author/speaker, and of subjects in general. The quote is also sug­
gestive of a somewhat uni-dimensional image of meaning, uni-dimensional in
the sense that it hints at but one true source and interpretation of meaning -
Daisy's. The second quotation - written over half a century before the first -
implies, on the other hand, that meaning is something bestowed on words (and
actions) by a variety of others, and that the endowment of such meaning is
variable or multi-dimensional. (In that respect it is characteristically 'post­
modern' .) Indeed, we can begin to see by means of Tolstoy's example how
'meaning' (like text, in general) is a situated product.
The tension expressed through these examples, between the author/subject
as the source of meaning and the situated 'other' as a source of meaning, has
formed the battleground for numerous debates within the social sciences. This
is not an appropriate point to deal with such debates in great detail. However,
we may as well note immediately that since the beginning of the twentieth
century it is, generally speaking, the situated 'other' - as reader and interpreter -
that has gained the upper hand. Thus de Certeau (1 984: 1 70) , for example,
states, 'Whether it is a question of newspapers or Proust, the text has a meaning

Copyrighted Material
110 USING DOCU MENTS IN SOCIAL R ES EARCH

only through its readers; it changes along with them; it is ordered in accordance
with codes of perception that it does not control' . Tracing the routes through
which the reader came to displace the author as the fount of meaning is an
issue of some considerable interest - though one for cultural historians rather
than ourselves. We have, of course, already noted the claim of Barthes ( 1 977)
that the 'birth of the reader must be at the cost of the death of the author'. It
is a distinctly late twentieth-century claim, and is one that is echoed in the
work of many other thinkers - Foucault, Derrida and Kristeva among them
(see Dosse, 1 997a, 1 997b) . Indeed, the ambiguity of meaning formed a critical
point of focus for the latter. Derrida, for example, argued that since all texts
contain ambiguities they can be read in different ways (/a difference) , and since
this is so, any ultimate interpretation of a text must forever be deferred (/a
differance) . In a similar manner the literary theorist Iser (1 989) has referred to
the 'indeterminacy of meaning' - an indeterminacy that allows space for the
reader to interpolate him- or herself into the text. (Unfortunately, as other com­
mentators have pointed out, the claim that language cannot make unambigu­
ous claims contains a fatal and damning paradox.)
Rather than base our discussions in the work of literary theorists, such as the
above, it would be more appropriate for us to ground our thinking in social
scientific debates. To that end I shall, in the paragraphs that follow, highlight
some markers that have been left behind in the long trek to get to grips with
the nature of , meaning' in social scientific contexts. The aim, of course, is not
to provide an exhaustive analysis of the issues, but rather to provide a sketch of
the journey to date.
One of the most appropriate of all possible starting points is with the ideas
of Wilhelm Dilthey (1 833-1 9 1 1 ) . Dilthey was among the very first of modern
philosophers to argue that there was a fundamental difference between what
he called the cultural and the natural (physical) sciences. The latter dealt with
an exploration of facts and the causal laws that explained such facts. The former
dealt with meaning. Meaning necessitated understanding ( Verstehen) rather
than 'explanation'. Indeed, argued Dilthey, since the subject matter of the two
sciences differed then their methodologies must differ likewise.3 Like many
who were to follow this path, Dilthey took the presence of meaning to be
self-evident.
Dilthey's arguments, together with many others, were later picked up by
Max Weber (1 864-1 920) and developed during the earlier part of the twentieth
century. Weber also held meaning to be at the core of social scientific investi­
gation, and, like Dilthey, he claimed that such meaning had to be 'understood'.
Social science therefore needed a method of Verstehen in order to grasp the
significance of an action to the actor. For Weber, as with Dilthey, meaning was
believed to reside in the minds (or intentions) of acting subjects. In that respect
they both subscribed to a form of what we might call mentalism - the notion
that entities reside in individual minds. Naturally, there were important differ­
ences between the ways in which the two enquirers conceptualized meaning,

Copyrighted Material
 C O N T E N T, M EA N I N G AN D R E F E R E N C E 111

and they had very different ideas about how meaning was to be accessed,
Dilthey arguing for a form of empathetic understanding and Weber for a more
distanced, structured approach that involved the construction of typical ways of
acting in given circumstances. Both authors were, however, thoroughly modern­
ist in their desire to locate meaning as a property of subjects - as a feature of
human psychology (see Outhwaite, 1 975). Some of the problems faced by
Weber's theory of meaning have, of course, already been touched upon in
Chapter 1 . Here we need only note the significance that Weber attached to the
search for meaning through the method of Verstehen.
It is a matter for historians to decide how Weber influenced twentieth­
century American sociology. What is clear is that the concern with the meaning
of social action figured largely in such sociology. This was especially true of
work executed in the symbolic interactionist tradition. The symbolic inter­
actionists, however, added a new twist to the theory of meaning. They socio­
logized it. Thus, G.H. Mead ( 1 863- 1 93 1 ) , for example, argued that the
meaning of objects, symbols, gestures and so forth were not so much contained
in thoughts, but forged in and through human action. In stating that claim he
divorced meaning from mentalism. (,Meaning is not to be conceived as a state
of consciousness' (Mead, 1 934: 78) .) He located meaning, instead, in fields of
social action. It was a critical shift, and it freed the concept of meaning from
the strictures of folk psychology.
This sociologized theory of meaning was further developed in the work of
the phenomenologist Alfred Schutz. Schutz argued that interpretation of
meaning is the fundamental activity of social life. In that sense Verstehen is not
some special method of the social sciences but the method by which the every­
day, intersubjective world (Lebenswelt) is constituted. So, determining how
people interpret and obj ectify the world is the central problem for social
theory. Thus Schutz was interested, for example, in how people categorized the
world into recognizable and meaningful events. (They do so, claimed Schutz,
through the use of typifications.) Moreover, and as with Mead, Schutz saw
meaning as specific to contexts of action. In that sense he claimed that mean­
ing is 'indexical' and 'reflexive' . That is to say, meaning is both situated in and
(at the same time) constitutes the setting in which social action occurs.
Ethnomethodologists such as Garfinkel and Cicourel, of course, followed
through on these insights from the late 1 960s onward.
Studying the meaning of a text is not, of course, the same as studying the
meaning of action. In the case of action and interaction the 'author' and the act
are bound together in time and place, and so it is possible to talk in terms of a
situated analysis. With text, however, author and product are forever divorced.
Text is always autonomous in that respect, and the 'author' is necessarily absent.
And it is for such reasons that hermeneutics is faced with problems that are
somewhat different from those that confront the student of social interaction.
Indeed, by facing such problems, hermeneutics has been forced to introduce
various subtleties into the study and analysis of 'meaning'.

Copyrighted Material
112 USING DOCUMENTS IN SOCIAL R E S EARCH

In this last respect an important figure on our quest to get to grips with
meaning is Hans-Georg Gadamer. In fact, we might say that the problem of
understanding was in many ways turned on its head by the work of Gadamer
(1 975) . For; unlike Dilthey, Gadamer argued that understanding formed the
basis not only for enquiry in the social sciences, but of human life and enquiry
in general. Understanding for Gadamer, however, 'is not based on "getting
inside" another person' ( 1 975: 345) . Instead, he suggests, we would be far better
considering understanding as a form of translation. So, for example, when a
person translates a text from one language to another language they do not
seek to reproduce every wish and whim of the original author in a one-to-one
correspondence of original and copy. More likely, the translator decodes
the original from a declared culturally formed platform, and then recodes the
document in his or her own style, and with his or her own words. The trans­
lation is in that sense a fusion of two parties - the meaning embodied in the
original text is reproduced through a meaning interpreted by the translator.4 It
is, in the language of Gadamer, a 'hermeneutical conversation' ( 1 975: 349) .
What is more, it is clear that the culture in terms of which the interpreter
interprets is every bit as important in this process as is the culture in which the
original text was first produced. So interpretation demands a 'fusion of horizons' .
(In this regard, states Gadamer, the example of translating a text is an ideal one
because we can see that psychological processes are not in any way involved in
'understanding' .) It further follows that 'meaning' is not something to be dis­
covered, it is something to be produced. Indeed, there can never be a final,
once-and-for-all interpretation of anything. To understand is to understand dif­
ferently.s The latter point, of course, raises problems about the nature of truth
in social scientific enquiry. Oddly, given the title of Gadamer's book, how we
come to distinguish the true from the false is not an issue that he deals with.
A further attempt to locate understanding in terms of text is also available in
the work of Ricoeur ( 1 98 1 ) . Ricoeur is of interest for several reasons. Above all,
perhaps, he sees in text the ideal case in which meaning is divorced from con­
text. There is a 'distanciation' between text and author. That is to say, that once
text is sent out into the world its meaning is necessarily divorced from authorial
intention. So meaning cannot be equated with the indexical or the situated. Nor,
on the other hand, is it possible to fill the void of the absent author with the
meanings of the reader - as, say, Iser or Derrida seem to suggest. In that respect
Ricoeur is at one with Gadamer. He therefore recognizes that some kind of
fusion of horizons between the text and the reader must take place. The manner
in which he conceptualizes the fusion is, however, somewhat different. For what
Ricoeur suggests is that although the reader necessarily appropriates the text in
terms of his or her own lifeworld, the text also constrains and produces. That is
to say, the text - by virtue of its content - has a creative potential that acts upon
the reader, and, by its very nature, develops and broadens the reader in new ways
of thought and action. There is a sense, therefore, in which a text structures its
reader, and that, of course, is a point that has been illustrated in Chapter 5.

Copyrighted Material
 C O NTENT, MEANING AND REFERENCE :1:13

Now, it is not my aim to elaborate on these various claims, and still less to
adjudicate between them. Interested readers are better referred to the texts of
the cited authors (above) . I have sought only to pinpoint some of the central
issues. Three points, above all, need to be grasped. First, all of those who are
concerned with text take it for granted that text (and action) has meaning.
Secondly, they believe that meaning needs to be grasped or understood.
Thirdly, they argue that in order to 'understand' some degree of interpretation
is necessary. How that interpretation is to be carried out differs according to
each author. And, I might add, none of them specify how interpretation should
proceed in any concrete instance. In that respect they are more often con­
cerned with issues of ontology (of existence) than of epistemology (of know­
ing) . So I shall end this section by pointing towards a route that does carry
with it some practical implications for the analysis of text. It is a route that has
been signposted by the philosophers VW Quine (1 953) and Hilary Putnam
among others. Putnam (1 988) has argued that meaning talk invokes a form of
mentalism - or psychologism. So as with Mead and Gadamer he would argue
that 'Meanings aren't "in the head'" ( 1 988: 73), but he also considers the possi­
bility that we can do without meaning talk altogether. In a similar way Quine
argues that talk about meanings can be safely 'abandoned' (1 953: 22) . For, what
is more important than meaning is the process of 'reference' . Reference is
socially grounded. Indeed, what Putnam argues is that we ought to focus on
how words, sentences and ideas are used in social practice, always keeping in
mind that 'use is holistic'. That is to say, keeping in mind that there is always a
body of statements to consider in which the individual words and sentences
merely slumber. The social scientific focus needs, therefore, to be on the whole
as well as the parts - on the corporate body of statements rather than the single
claims. (Discourse is not a word that Putnam and Quine use in this context,
but that does not of course prevent us from referring to it at this exact point.)
In this view, then, what is important about the case of Anna 0 is not what she
meant or thought, nor even what Breuer or Freud meant or thought, but how
the parties involved variously referenced things (about the mind and body) in
the framework of Freudian discourse. One advantage of this manner of
approaching things is that, as we shall see, 'reference', unlike meaning, can be
made t�actable by the social scientific researcher.6 As ever, in order to see how
this might be so, we need to consider examples.

Referencing death

Let me begin with a relatively simple example and move outwards. In a study
of death in the city of Belfast undertaken during the 1 980s I looked at what I
called the social distribution of sentiments. The basis of the idea came from
some observations made by Durkheim ( 1 9 1 5) and his associate Robert Hertz
( 1 960) during the early twentieth century. Both writers claimed that what was

Copyrighted Material
114 USING D O C U M ENTS IN SOCIAL RESEARCH

significant about grief and mourning was not the inner psychological states of
the bereaved, but outward manifestations of sorrow. Sorrow was socially struc­
tured. For example, in many societies the deaths of infants hardly pass for com­
ment, whilst the death of a powerful man is often the occasion for considerable
social disruption. Even in contemporary western societies the deaths of the
stillborn, for example, are not considered to be an occasion for formal or
ritualized mourning and burial, and this is in marked contrast to our responses
to the death of, say, children. It follows therefore that we can gather an awful
lot of information about reactions to death by observing, in detail, what people
do at times of bereavement. Indeed, studying what people do is as instructive
(possibly more so) than speculation about what they 'feel' or 'think' or 'believe'.
In the city of Belfast it is common (in fact, it is more or less required) for
people to announce the death of a relative in the 'Deaths' announcements of
one of the major regional newspapers. This textual declaration informs the
world not only of a death, but also about the nature of burial or other arrange­
ments. More importantly, such notices offer an opportunity for people to make
comment on their feelings and thoughts and hopes for the deceased. Such
public declarations of sentiment are used in many different ways by different
kinds of contributor. For example, they are often used by friends and second­
degree relatives of the deceased's family to express solidarity with the bereaved.
At other times they are used to underline a claim to kinship. In some cases they
are used to underline the importance and significance of the person who has
died. And in yet other cases they are used to express religious or political affili­
ations in general. Consequently, such death notices offer a rich seam for study.
Yet in mining that seam, we have no need to call upon the concept of mean­
ing, or, indeed, of any other psychological concepts.
For example, one way in which we might study the use of such material is
to undertake an analysis of the number of such notices that each person
receives, as well as the origin (in terms of kin and friendship relations)' of the
notices, and the references that are contained within them. In short, to execute
a simple content analysis (Weber, 1 990) of the relevant newspaper columns.
That is just what I did, and the results are reported more fully in Prior (1 989) .
Here, I pick up on only a few of the emergent themes, most of which are
contained in Table 6. 1 .
We can see from a simple analysis of the results contained in that table
that the mean number of notices that each person gets alters according to
occupational background, age and marital status of the deceased. Manual
workers get more notices than non-manual workers, people under the age of
60 get far more notices than do older people, the married tend to get more
than the unmarried. There is also a variation in the number of citations by
gender of the deceased, but it is not 'significant' in a statistical sense.
There is no need to go into the statistical background of the examples (but
do note how such manipulations can have their uses in such exercises) . My
purpose here is simply to indicate the fact that 'reference' can in itself serve as

Copyrighted Material
 C O N T E NT, M EA N I N G A N D R EF E R E N C E 115

TABLE 6.1 The distribution of death notices in Belfast newspapers in a

1 0% sample of deaths, 1981
Number of Mean number of
cases In notices per person Standard deviation
Variable category In category from the mean
Occupation of deceased at death
·Non-manual 106 4 . 0* 3. 6
Manual 270 8 .6 9.1
Gender
Male 200 6.4 6. 8
Female 176 8.3 9.4
Age by category
01-59 65 12.4* 14.5
60+ 311 6.2 5.6
Marital status
Married 159 9.2* 10.0
Single/divorced/widowed 190 6.0 6.5
*t-test indicates a significant difference p > 0.001.

a key point for the social researcher. Hence, the simple act of inserting a notice
in a newspaper tells us things about the referencing of death in a western city
and how it is related to key social variables.We could, of course, go further and
look at referencing in more detail. Thus, we might ask, for example, what is
it that is contained in such notices? What, for example, are the religious and
political sentiments expressed in the notices and how do they vary according
to the social characteristics of the deceased? That would also serve as a valuable
research exercise - though it is not one that I wish to follow through at this
point. Were we to do that, however, we would then be drawn into a discussion
and analysis as to how the language of death notices is related to or expresses
a discourse (Potter, 1 996) on death. Note, however, that even at the stage of dis­
course the focus would be on how references are interlinked in the text rather
than with what they mean in this, that, or some other person's mind. But rather
than develop the death notice material any further it will serve us better to use
a new example.

Referencing metaphors

It has long been observed that metaphor is central to the representation of

illness. Indeed, in a rather famous literary analysis of this theme Susan Sontag
produced her fllness As Metaphor (1 978) . This was a study of the ways in which
metaphors were used to describe life-threatening illnesses such as TB and
cancer. For example, Sontag pointed out how war metaphors were prominent
in discussions of cancer as an illness. People fought cancer - to win. People
shore up their defences against the threat of cancer; they demand the latest

Copyrighted Material
116 USING DOC U M E NTS I N SOC IAL RES EARCH

weapons against the disease; they display immense courage and bravery in their
final struggle - and so forth. This use of war-related metaphor in discussions of
cancer is, of course, very widely used by politicians, scientists and, indeed,
authors. Proctor (1 995) provides an intelligent narrative of the ways in which
the metaphor was used in the USA during the 1 980s and 1 990s.
More recently, Seale (200 1 ) has taken up a similar theme and examined the
use of what he calls struggle language in media accounts of people with cancer.
Both his findings and his methods are of interest to us. Broadly speaking his
methods of study were as follows. Using an on-line database of English language
newspapers he looked at the ways in which people with cancer were repre­
sented in such documents during the first week of October 1 999.Almost 2,500
articles relating to cancer were identified and these were then read in order to
select the 358 that provided a story about a person with cancer. Content analy­
sis of the 358 stories was then used to map basic reference points such as the
types of cancer mentioned, the types of people represented (young, old, male,
female, black, white and so on), and the words that were used in the articles to
describe cancer. As you might guess, struggle language - as referenced by such
terms as fight, battle, survive, beat, struggle, victory, defeat - was uppermost in
the content. Such struggle language was not, however, necessarily related to
war images in a simple and straightforward way, but more frequently filtered
through a further set of metaphors that concerned sport and sporting activity.
Seale concludes by indicating how images of sport are closely tied into images
of fighting and war and how sporting activity is frequently viewed 'as standing
for a life and death struggle' (200 1 : 325).
A number of useful points arise from this analysis. The first is that, as with
the death notices example, counting what is referenced in text (straightforward
content analysis) can itself be of benefit in highlighting the concerns of a social
group. Secondly, content counts on their own are, however, rather limiting. In
order to exploit those counts one has to move into a realm beyond the mere
words on the page and into some form of discourse analysis. That is to say, one
has to begin to ask questions - and obtain answers - about how the various
terms and concepts that are counted are interlocked one into the other so as
to form a stance or position or 'weltanshauung' . Once again, however, one
need not call upon mentalist ideas in order to follow through with such an
analysis - as is evident from the next example.

Accounting for fatigue

Throughout this book I have suggested that to resort to inner mental proper­
ties of subjects is problematic for social science. Consequently a desire to study
the 'meaning' of a document is potentially misguided. Instead, we ought to
study what it is that is referenced in a document, and we have seen how that
might be done by use of the two examples above. However, we have also seen

Copyrighted Material
 CONTENT, M EA N I N G A N D R E F E R E N C E :1.17

how simple counting of content items has its limitations. For, we often wish
to go beyond noting that something is referenced and to ask questions about
how specific items are integrated into 'accounts' about this, that, or some
other matter.
The concept of an account is a useful one for our purposes (see Scott and
Lyman, 1 968) . Accounts are ubiquitous. They are usually offered in relation to
questions. For example, if you were to ask me why I became a sociologist, then
I would respond to you citing reasons, events, people, and possibly things that
were relevant to my becoming a sociologist. All of these items would be
wrapped up in an account - probably an account in narrative form. In most
cases accounts are tailored for a specific audience. So, my account as to why
I became a sociologist would be structured in such a way as to sound rational,
persuasive and coherent to the audience that I was addressing. I would include
within it only those items and relationships that I wished that audience to hear
about and that I considered to be culturally appropriate. I would probably
ignore things that I considered not to 'fit' with the requirements of the audi­
ence in question. For example, I may decide ·that a response to the question
that referred to a 'calling' would be regarded as somewhat bizarre in the
modern world. Consequently, I might prefer to structure my accoupt by refer­
ence to a series of influences upon me - books that I read, people that I knew,
problems that puzzled me. I might do this even if I felt that I was unsure why
I became a sociologist in the first instance. Mter all, saying 'I don't really know'
is not as impressive a response as one that contains a reasoned narrative that
refers to, say, a career plan. In addition, the central question might not be one
that I had ever considered, and so I might, perhaps, merely invent a response
for the sake of dealing with the question posed to me. This latter is always a
danger inherent in asking people for their motives, reasons and opinions, any
one of which may be invented for the purpose at hand.
In what follows I am going to look at accounts of illness. When people
fall ill , they often provide accounts about how the illness was caused, how it
developed, and how it might be best managed and treated. Psychologists some­
times refer to 'health beliefs' in such contexts. As you might guess, however,
reference to belief (as an inner mental state) is neither necessary nor particu­
larly useful - as the following details indicate.
As far as illness is concerned, we know that different people in different parts
of the world tend to construct accounts of illness that reflect their own histo­
ries and traditions and practices (see, for example, Lewis, 1 975; Prior et al.,
2000) . What people put into an account and how they connect the things so
referenced is a worthy topic of investigation in itself. All that is needed is that
we study how the account hangs together and what it contains.
Just to indicate how this might be done, I am going to look at some accounts
about fatigue. Feeling fatigued is a common problem (Meltzer et al., 1 995) .
Estimates of the proportion of people who feel totally exhausted at the end of
a working day sometimes reach as high as one-third of the adult population.

Copyrighted Material
118 USING DOCUMENTS I N SOCIAL RESEARCH

For most people the experience is temporary and short-lived. In a small

percentage of cases, however, the feelings of fatigue persist for long periods.
Sometimes there are clear reasons as to why this is so - perhaps the person is
suffering from some known physical malady or working overlong hours. I n
other cases, where clear explanations for the fatigue are absent, medical practi­
tioners and their patients are faced with a puzzle. How they seek to answer that
puzzle is the focus of this section.
Chronic fatigue (fatigue experienced for more than six months) is a rela­
tively rare occurrence. It is often referred to as a syndrome - that is, as chronic
fatigue syndrome (CFS) . In the USA it is known as CFIDS (chronic fatigue
and immune deficiency syndrome) . The causes of CFS are usually difficult to
pin down. Nor is it always clear how such fatigue might be overcome. As you
might expect, to ask 'what do you mean?' of someone who experiences CFS
is usually not very helpful. Indeed, CFS sufferers often argue that no one can
know what it 'means' to be fatigued unless they experience the condition. I n
fact, CFS sufferers normally place a heavy emphasis o n their experiential
knowledge of the condition as against the scientific stance that is adopted
towards CFS by physicians. In line with the tenets of this chapter, then, per­
haps we should ask different questions, such as, what do people who experi­
ence CFS reference in accounts of their suffering? What kinds of entities do
they recruit into such accounts? And how are such references and entities
interlinked?
One way of getting an account is to access patient support group literature
on the www. CFS sufferers are very active in patient support groups and con­
sequently they publish considerable amounts of literature on the nature of
the disease, its causes, its consequences, and its treatment and management. For the
purpose of this exercise I accessed just one patient support document on the
www. Table. 6.2 provides the results of a simple counting exercise executed
on that document. The table contains word counts for selected terms used in
the WWW leaflet.
As we can see, even a simple counting exercise - basic indexing - can help
with matters of reference/ though it is not of course surprising that the word
'fatigue' looms large in the table. In this leaflet it occurred 55 times in the
2,31 5-word text. Omitting prepositions, conjunctions and so forth (words such
as to, the, and, or, but) , there are in fact only five words that get 50+ mentions
in the entire text. (The two words omitted from Table 6.2 are 'subject' and
'patient' .) And it seems fair to assume that the word frequency order is in some
way related to the weight of concerns that sufferers have.
At the top tier of citation (50+ uses) there is comparatively little problem in
identifying the key concerns. As we move down the word frequency list, how­
ever, it becomes evident that counting on its own has severe limitations. For
example, the table also suggests that research is important to CFS sufferers as
are viruses and virology, fibromyalgia (a related condition to CFS) and disease.
Depression, psychology, psychiatry, immunology, neurology, genetics, and an

Copyrighted Material
 CONTENT, M EA N I N G AN D R E F E R E N C E 119

TABLE 6.2 Occurrence of selected

words in a 2,315-word patient support
group leaflet on chronic fatigue
syndrome
Fatigue 55
Chronic 51
Illness 50
Syndrome 46
Research 29
Virus/viral/virology 23
Disease 19
Fibromyalgia 18
Depression 14
Immune/immune-related/immunology 9
Genetic 4
Psychology/psychological 4
Neurology/neurological 4
Psychoneuroimmunology 2
Psychiatric/psychiatrists 2
Mental 1
Mind 1

unusual hybrid referred to as psychoneuroirnmunology, also get mentioned -

along with hundreds of other words. However, it is not clear how these refer­
enced items connect one to the other and how they are combined into a
coherent account of what CFS might be.
Indeed, to follow through on the relationships between words we need to
look at context, and using a simple concordance program this too is possible.
All that is required is that we call up the context in which a given word is used.
To illustrate some relevant issues I reproduce below the line contexts in which
'genetic' was referred to.

Extract 6.5. 1

Is CFS genetic?

The cause of the illness is not yet known. Current theories are looking at the possibilities of
neuroendocrine dysfunction, viruses, environmental toxins, genetic predisposition, or a com­
bination of these. For a time it was thought that Epstein-Barr Virus (EBV), the cause of
mononucleosis, might cause CFS but recent research has discounted this idea. The illness
seems to prompt a chronic immune reaction in the body, however it is not clear that this is in
response to any actual infection - this may only be a dysfunction of the immune system itself.

Extract 6. 5.2

Some current research continues to investigate possible viral causes including HHV-6,
other herpes viruses, enteroviruses, and retroviruses. Additionally, co-factors (such as
genetic predisposition, stress, environment, gender, age, and prior illness) appear to play an
important role in the development and course of the illness.

Copyrighted Material
120 USING DOCUMENTS IN SOCIAL RESEARCH

Extract 6. 5.3

Is CFS genetic?

Several studies suggest that there may be a genetic component to CFS. This is not
surprising since CFS seems to involve immune dysfunction to some degree, and immune­
related illnesses often have a genetic component. The evidence on this point is not clear.
And the fact that there seem to be cluster outbreaks of this illness seems to argue against
genetics as being the sole factor.

These three extracts are instructive for a number of reasons. First, they suggest
that CFS sufferers tend to focus attention on the physical causes of illness. As
one can see, the presence of viruses and immune responses loom large within
the selected text. Such concerns are partly reflected in the word frequency
table above. In the case of ' genetic and genetics', however, it hardly ranks any
higher than reference to psychiatry and psychology. So we might be drawn
into arguing that psychiatry and psychology are as significant to CFS sufferers
as are genetics and immune responses. In a sense they are. However, the way in
which psychology and psychological factors are integrated into the text sug­
gests that they are evaluated in a somewhat different way from the physical or
somatic causes of disease. Look at the following extracts.

Extract 6. 5 . 4

Emerging illnesses such a s CFS typically g o through a period o f many years before they
are accepted by the medical community, and during that interim time patients who have
these new, unproven illnesses are all too often dismissed as being 'psychiatric cases'. This
has been the experience with CFS as well.

Extract 6. 5 . 5

[CFS] i s not depression, nor does depression cause [CFS] .

These extracts suggest, then, that where psychiatric factors are referenced, they
are referenced in what we might call negative contexts. They suggest that CFS is
not a condition in which psychiatry has any major role to play, though they also
hint that someone might have suggested a link between psychological factors and
CFS - a link that is to be flatly rejected.
The relationship between psychological factors and the symptoms of CFS is a
contentious and complicated one. It often forms the nub of debates concerning
CFS. For example, in the UK, members of the medical profession produced a
report on the syndrome in 1 996 (Royal Colleges, 1996). The report covered many
aspects of CFS - including an overview of the role of viruses, immunology and
psychological states in CFS. Their conclusions suggested that CFS was not likely
to be caused by either viruses or immunological disorders. On the other hand
they did suggest that psychological factors were likely to be implicated in the
syndrome and that one of the most effective forms of management for CFS seemed
to be a psychological therapy known as cognitive behaviour ther:mv Thll�. the

Copyrighted Material
 CONTENT, M EAN I N G A N D R E F E R E N C E 121

report stated that 'Depression and/or anxiety represent the strongest risk factors
so far identified for CFS' (1 996: 1 5), but cautioned against any 'simple equation of
CFS with psychiatric disorder' ( 1996: 1 5) . As we have seen, the first claim is quite
at variance with what many sufferers argue to be the case.
Concentrating on how entities are linked and integrated into accounts is
always more productive than a simple focus on what is referenced. And this is so
with patient and professional accounts of CFS. In the paragraphs above I have
focused, ever so briefly, on how psychological conditions are integrated into
accounts. Broadly speaking, when CFS patients refer to psychological symptoms
they usually interpret them as the consequences of illness rather than as a cause,
and therefore as quite marginal to their real illness. Medical professionals on
the other hand tend to argue in the opposite direction, and to suggest that
psychological problems often lay at the base of a patient's difficulties and physi­
cal experiences. Naturally, an enquiry into how such disagreements are resolved
in practice requires an analysis of doctor-patient interactions. 0Jie certainly
cannot discover how things are resolved (in practice) by using documents alone.)
Nevertheless, a study of documentation is often useful as a precursor to the study
of interactive episodes. For documentation often lays out the basis of practical
reasoning and behaviour - as it does in the case of CFS. Thus, in a study of CFS
doctor-patient interactions undertaken recently I was interested in whether
debates in the literature were reflected in what CFS sufferers spoke about in a
clinic. The results of the initial research are described in Banks and Prior (2001).
Here I shall focus on only one extract from one patient. The patient is discussing
a visit to a medical specialist (lines 75) . Of particular interest is the way in which
the speaker references the link between his illness and his psychological problems
(lines 78-9) . As you can read for yourselves, the causal link between depression
and physical illness is attended to, but in such a way that the symptoms of
depression are seen as a result of some other illness rather than a cause of the
illness itself. In lines 75-6 the patient refers to a private consultation with a CFS
specialist. In line 77 a mention is made of ME which is an alternative expres­
sion for CFS often used by patients in the UK. (ME is an abbreviation of
myalgic encephalomyelitis - a physical disorder of the skeleto-muscular system.)
In lines 78-9 the patient, crucially, reverses the causal direction that is suggested
in the Report of the Royal Colleges (1 996) mentioned above.

75 (Patient 33): Anyway like he examined me and whatever, and um, we had
a good talk and er,
76 what happened then, he wrote back to my doctor like I said ,
he didn't tell anything, he
77 did say to me there are symptoms of M E, the rapid fatigue
you've got and whatever.
78 I told him, I said I know I was depressed at the time, but it's
not depression it' s the
79 fact that it's the rapid weakness is getting me down.
80 ( Doctor) : Yes.

Copyrighted Material
122 USING DOCUME NTS IN SOCIAL R ES EARCH

Conclusions

Documents have content and content requires analysis. Indeed, document

content is usually thought to contain insight into people's thoughts, ideas and
beliefs. These are often considered the meat and drink of the social scientist as
well as of the literary theorist.
One obvious question to ask with respect to the content of any document
is 'what does it mean?' And the analysis of the meaning of text is, as we have
seen, one with a long and tortuous history (one that involves the history of
hermeneutics) . We have also glimpsed at how the study of meaning can entice
us into a methodological minefield. For whose meaning are we to study and
whose meaning is at stake? Are we to study meaning as a quality of writing
authors or meaning as an interpretation of readers? Following Gadamer and
Ricoeur, of course, we would have to conclude that the study of meaning
requires a dialogue between interpreter and interpreted in the manner out­
lined. But, do we need to focus on meaning at all?
I have suggested (following Quine and Putnam) that we can safely abandon
questions about meaning and, instead, look at reference. Better to ask such
questions as, 'what is it that is referenced within documents?' than to ask, 'what
does this mean?' For, if nothing else, references can be indexed, and then
counted in the manner described above. However, we also need to be aware
that the problem with straightforward content analysis is that the same words
and terms can be used to reference different entities. Thus, the term 'ME' might
be used by physicians and patients equally but, in the light of what I have said
above, it is most likely that they would be referring to markedly different
things. Consequently, in order to appreciate the force of textual discourse one
has to do more than count. One has to attempt to get a picture of the ways in
which the network of references interlock. It is, perhaps, what we might call a
matter of intertextuality.8
So let us sum up the suggested strategy for dealing with content:

• Documents have content, and content requires analysis.

• Rather than focus on the 'meaning' of a word, sentence, paragraph or
document, it is far more fruitful to ask about what is referenced within the
document.
• In studying patterns of reference, simple content analysis has its place.
• Content analysis on its own, however, will be insufficient to highlight the full
pattern of referencing between objects cited in the text. Reference therefore
needs to be studied in context.
• As well as examining the words and terms that are recruited and connected
within a text, we need also to look at who enrols and who attacks the text
in everyday projects. Above all , we need to study how people use text in
action.

Copyrighted Material
 CONTENT, M EAN I N G A N D R EFER E N C E 123

In the following chapter we will look at some examples of how people use
words to picture the world and to place actors within that world. The examples
are drawn from two fields of human endeavour - science and anthropology.
Our task, as ever, will be to unravel what it is that people are doing with words
rather than with what they are intending - or 'meaning' .

RESEARCH EXERCISE

Exercise 6.1

As we have noted in Chapter 3, people with psychiatric disorders tend

to be portrayed somewhat negatively in the mass media. For example,
serious psychiatric d isorder is often associated with violence and
extreme violence is, in turn , often interpreted as 'madness'. Using
electronic search facilities examine how mental i llness and the
mentally i l l are referenced in newspapers. Make sure to select a
variety of such papers and confine the text search to specific dates -
say 1 January 2000 up to and including 31 July 2000. Compare the
reference terms used in the different papers. Pay particular attention
to any associations between psychiatric d isorder and violence, and
psychiatric disorder and genetics. Some helpful leads on the problems
can be found in Philo (1996).

Notes

1 Hermeneut - an interpreter, especially of scripture.

2 In terms of the linguistics of Saussure ( 1 983) the meaning of a term is given by its
position in a chain. Meaning can therefore change if and when position in a system
changes. The metaphor of a linguistic chain was extended beyond language and into
other systems of symbolic meaning - myths, clothing - by a later generation of
structuralists such as Barthes ( 1 985) and Levi-Strauss ( 1 969).
3 Dilthey's desire to differentiate the human sciences (or Geisteswissenschaften) from
the natural sciences (or Naturwissenschaften) stimulated debates that lasted through­
out the twentieth century. A good overview of these debates and their broad
methodological implications for social researchers is provided by Giddens (1 978).
4 For example, the English translator of Freud's work - Lytton Strachey - translated
simple German terms that were used by Freud, such as das Ich and das Es, into 'Ego'
and ' Id', thereby adding a guise of scientific and technical precision to plain, every­
day, words.
5 A fellow traveller in this regard is Charles Taylor (1 987).
6 In what might be regarded as a parallel manner, Silverman ( 1 993) has argued that
talk ought to be scrutinized so as to uncover strategies of practical reasoning rather
than meaning.

Copyrighted Material
:1.24 USING DOC U M E NTS I N SOCIAL RESEARCH

7 Many search engines currently in use on the WWW find relevant 'pages' using a
similar, word counting, technique.
8 ' Intertextuality' is a 'gadget' invented by the French literary critic/philosopher Julia
Kristeva (1 980) . The term refers, in part, to the notion that the meaning of a single
text is always bound up in its relations with others that are contemporary to it. The
concept also implies that texts are never singular or unique but comprise bits and
pieces of other texts written at other times and in other places.

Copyrighted Material
7
Do i ng Th i ngs with Words

Traveller's tales 125

The scientific report 127
From Kiriwana to Inishkillane 134
Conclusions 141
Notes 144

Traveller's tales

During the period 1 9 1 1-12 Apsley Cherry-Garrard travelled with Captain

Scott to the Antarctic. He was not included in Scott's team for the final leg of
the journey to the South Pole, and consequently he managed to survive the
rigours of Polar life whilst Captain Scott and his companions perished. On his
return to England, Cherry-Garrard reports that he was asked, by the Antarctic
Committee, to write the 'Official Narrative' of the 1 9 1 1 -1 2 Polar expedition.
The narrative was to take the for m of a scientific report on a scientific expe­
dition. Cherry-Garrard wrote the report, but he tells us in the introduction to
his 1 922 book, The Worst Journey in the World, that he felt that he was unable
to publish it under the auspices of the Official Committee, because, ' I could
not reconcile a sincere personal confession with decorous obliquity of an
Official Narrative' (Cherry-Garrard, 1 994: lii) . In short, Cherry-Garrard sensed
that there was an important difference between the nature of a scientific
narrative and a personal narrative. Yet, his book is full of scientific observations -
about the Emperor and other species of penguin, about seal life and the behavi­
our of whales and birds. He uses, at times, quite extensive footnotes in his
report, and makes reference to a number of scientific journals and books. Many
of his observations are cross-referenced with observations recorded in Scott's
diary and other sources - validated as it were. Nevertheless, The Worst Journey
in the World is not regarded in any way as a science text. Instead it is commonly
regarded as a 'travel' book - albeit one that is often described as the very best
twentieth-century example of that genre.1
The travel book is, of course, a distinctive form of literality that, under other
circumstances, might provide us with an instructive site of enquiry. (And so

Copyrighted Material
126 USING DOC U M ENTS I N S O C I A L R E S EARCH

too, the travel writer as a variant of the author-function.) I n this chapter,

however, our interests are to be restricted to two other sites of writing. The first
is the scientific paper, and the second is the ethnographic report. In both cases
we will be interested to focus on how people can 'do' science with words. I n
particular we will examine the manner i n which the relationship between the
author/discoverer and the (natural and the social) worlds that they investigate
are modulated through text. How, for example, the different types of , author'
report on the world; how they represent that world in their reports; and how
they inscribe themselves and their discoveries in the reporting process. As we
shall see, scientific (and anthropological) authors are normally keen to repre­
sent text as a neutral conduit for the transmission of observations and discov­
eries. The scientific text, in that sense, is seen as a means by which the scientist
can communicate with a readership in a manner that avoids contaminating the
data. Consequently, text is not seen as having any role in the constitution of the
various 'facts' that it reports upon. The extent to which that claim is support­
able will, however, become evident as this particular report unfolds.
Now, the analysis of documents of the kind referred to above is often under­
taken in a frame of what is called discourse analysis. Unfortunately, and as we
have noted before, 'discourse' and 'discourse analysis' are rather fuzzy terms. As
with all other words, their meaning can only be derived from their use. In
reviewing the multiple interpretations of such terms, van Dijk ( 1 977, I : 3)
argues that discourse analysis deals with 'talk and text in context' - that is, with
linguistic activity - but he also points out that the word 'discourse' can refer to
a much wider array of phenomena, such as styles of thought and analysis. This
wider vision of discourse - as, say, a set of closely integrated ideas and practices
about the nature of the world - was one favoured by Foucault ( 1 972) . The
latter was keen to indicate how the world comes to be known and understood
through discursive practices, and how a change in discursive regime can change
the world (and social relationships within it) itself. In short, how reality is con­
stituted through discourse.
This broader (Foucauldian) use of the term has its uses - especially in so far
as it encourages us to look at the manner in which discourse structures what
we might call the furniture of the world (that is, objects, 'facts' and 'things').2
We shall defer to it in our examination of the two realms of discursive practice
that have been selected. In relation to the realm of'science', however, we shall
primarily follow in the wake of a group of late twentieth-century writers -
such as Gilbert and Mulkay ( 1 984) , Knorr-Cetina ( 1 983) , Latour and Woolgar
( 1 979), Lynch ( 1 985) and Woolgar ( 1 9 8 1 ; 1 988) - all of whom look at the ways
in which science is created in and through discourse. That is, we shall adopt the
argument that 'science' is not a body of knowledge or practices independent of
scientists (such as, say, a true and correct explanation of the mechanics of
'nature'), but rather constituted in what scientists say and do. And we shall call
upon a similar argument in our investigation of anthropological reports. In the
latter case, however, we can supplement our claims with references to the work

Copyrighted Material
 DOING TH I N G S WITH WORDS 127

of those troubled by the possibility that anthropological writing might, after all,
not to be too distant from forms of travel writing. So 'ethnography', in parti­
cular, might be regarded as a form of fiction rather than a form of scientific fact
gathering (see Atkinson 1 990; 1 992b; Geertz, 1 988; van Maanen, 1 988) . Some
writers, of course, have made the claim that all forms of writing are forms of
fiction (see, for example, Derrida, 1 99 1 ) , but that - for various reasons - is an
awkward position to hold with any consistency. What we can say, however, is
that there are different forms of writing in the world and that they each claim
a different relationship to external reality. How that relationship is represented
in text is, in large part, the subject of this chapter.
Just to narrow down our concerns a little further it might be useful to add
that the specific focus of the chapter will be on the rhetoric of science and of
anthropology. That is, on the ways in which authors of anthropological and
scientific texts use words so as to persuade us of the veracity of their reports
and of their status as objective observers. We need also to. remain aware that
rhetoric is normally regarded as only one feature of discourse. Thus, semantics
(or the study of meaning) , narrative, argumentation, semiotics, pragmatics and
much more can also form the foundation for discourse analysis, and interested
readers will find some excellent pointers to those various modes of analysis in
van Dijk's (1 997) text. However, some writers (such as, for example, Bazerman,
1 988) argue that the study of rhetoric can encompass most of these latter
forms. In so far as that may be so, then our concern with scholarly rhetoric will
serve us well.

The scientific report

In 1 96 1 Ernest Nagel opened his book on The Structure oj Science with a

chapter in which he drew contrasts between common sense and scientific
thinking. Science, he claimed, organizes its knowledge and its findings in a
systematic manner; common sense does not. Science provides rigorous expla­
nations for events in the world; common sense does not. Science subjects its
arguments and findings to consistent criticism and analysis; common sense does
otherwise. In drawing such contrasts, Nagel developed a representation of
science that firmly belonged to what is generally referred to as the 'positivist'
camp (see also Chapter 8) . Such positivistic pronouncements (Giddens, 1 978)
on what science 'is' were frequently developed and published during the
nineteenth and twentieth centuries. In all such pronouncements 'science' was
seen as something set apart from other ways of thinking. I n particular it was
viewed as a means of obtaining objective knowledge about the external world,
knowledge that could be rigorously tested and retested until the truth of any
matter became plain to see. In short, science was viewed as a mirror to nature.
How science is to be demarcated from other forms of thought is a matter
for philosophy and philosophers, and there have been various detailed attempts

Copyrighted Material
.128 USING DOCUMENTS IN SOCIAL RESEARCH

to draw the required line (see Gillies, 1 993) . Such attempts are not of any concern
to us at the moment. What we seek to examine is exactly how scientists man­
age (in practice) to convince us of their special status as knowing subjects.
What, for example, are the textual and other strategies that they use to get their
ideas across - both to the likes of lay persons such as ourselves and other sci­
entists? Answers to that question require something other than philosophical
analysis. Indeed, it requires close examination of what scientists do, and that
becomes possible only when we adopt an anthropological or sociological
approach to the problem.
Anthropological studies of scientists are a comparatively recent phenome­
non. In the main such studies have focused on the laboratory as a site of
anthropological fieldwork - see, for example, Barley and Bechky ( 1 994) .
Charlesworth et al. (1 989) , Knorr-Cetina (1 983) , Latour and Woolgar (1 979)
and Lynch ( 1 985) . Most of this work has been executed in the realm of
the biological sciences. One of the key findings of such work is that text and
documents are fundamental to the very nature of laboratory work. (For where
are 'ideas', 'thoughts' , hypotheses, rules of procedure, results, findings and so
forth if they are not contained in documentation?) Indeed, Woolgar (1 988: 68)
sees documents as resting at the base of the scientific discovery process. This
certainly makes sense in so far as things (objects) and processes are only made
visible through drawings, photographs, pencil and ink traces, scribbles on
black- and whiteboards, memos, and scientific papers and books. Indeed, it may
be said that scientific objects are constituted in and through the use of such
documents; that texts produce scientific knowledge. Thus Latour and Woolgar,
in their study of the La Jolla Laboratory, argue, for example, that scientists are
'manic writers' ( 1 979: 48) , and are constantly generating scientific facts through
the use of written and other traces. Other traces - temperature charts, for
example - are routinely derived from what Latour ( 1 983) calls inscription
devices. That is to say, devices that turn material substances into documentary
form. So one of Latour's injunctions for any research project is to 'look at the
inscription devices' (1983: 1 6 1 ) :

No matter if people talk about quasars, gross national products, statistics or anthrax
epizootic microbes, DNA or subparticle physics; the only way they can talk and not be
undermined by counter-arguments as plausible as their own statements is if, and only if,
they can make the things they say they are talking about easily readable.

I n the positivist world of philosophers such as that of Nagel (mentioned

above) , such devices would have been viewed as mere technical instruments
that aid the scientist in his or her assessment of the true and real properties of
an object. That is to say, they would have been treated as a mirror on nature. In
the world of the 'constructivists' (Knorr-Cetina, 1 983) , however, inscription
devices are not merely adjuncts to scientific work, but central to the construction
of scientific objects themselves. What the object 'is' is in large part constituted

Copyrighted Material
 DOING THINGS WITH WORDS :129

by its trace. Indeed, for Latour (1 983) and Latour and Woolgar ( 1 979), the
laboratory as a whole is viewed as a productive force - a force that manufactures
the world so that 'scientific activity is not "about nature", it is a fierce fight to
construct reality' (1 979: 243) . A similar idea is expressed by Charlesworth et al.
( 1 989: 1 5 1 ) who regard laboratories as 'data-generation systems' - systems in
which documentation (photographs and charts as much as writing) plays a major
role. (See also Rapp (2000) on the laboratory manipulation of chromosomes.)
We have already examined the role of documents in scientific action in
Chapter 4. Here, we shall concentrate on how documents (in the form of
research grant applications, journal papers, textbooks and autobiographical
accounts of scientific discovery) also construct an image or representation of
science itself. In particular, on how documents generate an image of science as
a set of activities that interrogates 'nature' , and subsequently produces true and
useful knowledge about how the natural world works.
In order to achieve the latter, scientists may be said to adopt a special way of
'writing science' (Myers, 1 990) . For example, the use of first-person pronouns
(that is, T and 'me') is usually considered unsuitable in scientific prose. (Perhaps
this was one reason why Cherry-Garrard regarded his 'personal' account of the
Antarctic expedition as unscientific.) The use of the passive, rather than the
active, voice is also preferred. For example, a report in a medical journal might
state that 'patients were injected with x, y and z ' , rather than 'we injected the
patients with x, y and z ' , still less, 'I injected patients with x, y and z ' . (Lynch
(1 985) also provides an interesting example of what he calls the 'missing agent'
feature of scientific writing in relation to his study of rat brains) . This kind of
writing has the effect of distancing the scientist from the activities being
reported upon, so that they are more likely to be seen as exterior to the world
that they investigate and manipulate. This is sometimes all the more important
when animals are being discussed in experimental reports. For animals are very
often killed. Yet, one is unlikely to read in a scientific report a statement to the
effect that 'we killed 80 animals and dissected their brains' (see, for example, Birke
and Smith, 1 995). More likely, one would read something to the effect that '80
,
rat brains were harvested over a 48 hour period and assayed in the
,
laboratory using [XYZ techniques) - or, 'n animals were sacrificed under nem­
butal anethesia' (Lynch, 1 985) , and so forth. A sentence of the latter type has
many advantages. It does not mention killing, and no human agent is apparently
involved in the harvesting/sacrificing process. Harvesting, in any case, is a word
normally applied to the gathering up of grain, fruit and vegetables, and so the use
of the word serves, perhaps, to obscure the difference between vegetable and
other forms of life. Finally, only the rat brains were harvested - the fact that the
brains were attached to the rats is - shall we say - subtly elided.
Woolgar ( 1 98 1 ; 1 988), in particular, has devoted some time to analysing the
rhetorical techniques that scientists use in their reports to manufacture good
'science' . It would serve us well to review some of his claims here. For exam­
ple, he points out the importance of the scientific paper's overall setting. Thus,

Copyrighted Material
130 USING DOCU MENTS I N SOCIAL R ES EARCH

a good paper ought to be published in a 'good' journal. The authors should

have sound institutional affiliations ('good' universities or research institutes) ,
and, ideally, there should be a number of authors (a team, perhaps) . Clearly, a
lone, amateur, scholar publishing his or her work in a local or regional news­
paper is unlikely to be taken seriously. In fact, Latour ( 1 987) points out that the
adjective scientific is never attributed to isolated texts. On the contrary, it is an
adjective attributed to a document only when that document is produced
through a dense network of collaborators. The network will include 'researchers',
reviewers of scientific papers, editors, members of scientific committees and so
on - most of whom work in teams. This attention to setting is important, and
we shall see in Chapter 8 that whether a scientific finding is or is not published
and widely quoted seems, on the face of things, to be related to whether the
findings are positive rather than negative or neutral. In academic life, of course,
journals are sometimes placed in a hierarchy of influence and significance and
so scientists ought not simply to get their work published, but to get their work
published in 'first-class' refereed journals. So the New EnglandJournal ofMedicine
or The Lancet or Nature are seen and regarded as prestigious publications. A
published book chapter, on the other hand, would be seen as less prestigious.
Indeed, by drawing attention to the 'setting' of a text in this manner we can
begin to see how discourse is invariably structured by things outside of the
text. Thus, Charlesworth et al. ( 1 989: 1 7 1 ) argue that scientific papers are
designed first and foremost so as to be published rather than read. In other
words, the scientific paper is commonly regarded as an output device - a device
by means of which the productivity of authors and institutes can be judged.
Indeed, and as we have already seen, there are many things beyond the text3
(and language) per se that need to be addressed in any study of a scientific
paper. Discourse, in other words, always involves extra-linguistic processes.
Bazerman ( 1 988), for example, in his study of the scientific paper as a genre,
makes reference to the importance of various gatekeepers in maintaining firm
boundaries between what is and what is not to be regarded as scientific dis­
course. Such gatekeepers Uournal editors and reviewers) are as significant to the
structure of a rhetoric - to the fabrication of the scientific voice - as are the
writers themselves. For now, however, we need to look within rather than
without the text for an image of how scientific papers 'work'.
To persuade, scientific papers need references, lots of them. Latour ( 1 987)
argues that scientific authors use references in the same way as builders use
cement - to bind the structure together in such a way that each little bit of the
paper has to be attacked before the overall structure can collapse. Thus scientific
text, it is argued, is highly stratified with technical details and their associated
references. The text, as it were, recruits allies - in other texts - so as to over­
whelm with a sheer weight of numbers. In his examination of citation prac­
tices and their role in the development of scientific practices, Bazerman ( 1 988)
uses the concept of 'intertextuality' to refer to the wider web of textual rela­
tions that a scientific paper is required to display. Latour argues that the use of

Copyrighted Material
 DOING THINGS WITH WORDS 131

citations ensures that weak rhetoric becomes stronger as time passes, and as
more and more papers by the author (and that includes the laboratory as an
author) are published (Latour, 1 987: 1 03) . References in bibliographies and
footnotes are important.4 So too is an examination of the networks and
alliances in terms of which references are cited (Baldi, 1 998) . Yet, the rhetori­
cal force of any scientific paper is dependent on much more than references
alone. Thus, Woolgar, for example, indicates that the 'textual opening' of any
paper is significant in so far as it structures the reader in his or her approach to
the reading. That is to say, openings funnel the reader's attention onto specific
issues and inform the reader as to how the paper is going to develop and con­
clude. Textual entrees are usually achieved by the use of an 'Abstract' and 'key'
words - both functioning so as to mark out an audience to which the paper is
believed to be relevant. Paper titles can also be used so as to exclude certain
kinds of readers - and, by implication, to circumscribe an elite few who would
really 'know' what the paper is about. (One is reminded here of Goody's (1 968)
observations on the manner in which script can be used to structure an elite
of experts.) The use of 'headings' in the paper may also function to structure
the reader and the reading of the text. For example, in some scientific journals
there might be a formal way of proceeding through the paper - from an out­
line of the problem, through to the 'method' used to investigate it, an account
of the 'findings' and a 'discussion' of the implications of those findings.
Woolgar also makes reference to the use of 'pathing' and 'sequencing'
devices. These devices are used to indicate how the scientific results being dis­
cussed were arrived at. They often involve the use of a narrative of discovery -
pointing out precursors of the work, or how the work is the latest in a long
line of scientific developments, plugging gaps in the existing narrative. Pathing,
then, provides a trail to be read and locates the scientific results in a context of
cumulative scientific discovery. Sequencing provides a narrative in terms of
which conclusions have been arrived at - and the sequence is invariably pre­
sented as inevitable. Thus, in terms of the sequence, alternative readings of the
facts are closed down in the light of 'what happened next', and the focus is
increasingly placed on 'relevant' (rather than background and irrelevant) events.
Clearly, de Certeau's claim (1 984: 1 86) that, 'Narrations have the power of
transforming seeing into believing and of fabricating realities out of appear­
ances' has much resonance in scientific texts. Finally Woolgar refers to 'logic',
the process by which a whole series of events are necessarily connected so that
other ways of seeing the world are gradually occluded (Lynch (1 985) also
makes reference to similar techniques) .
More important from our standpoint, however, is the use of what Woolgar
refers to as 'externalizing devices' in scientific discourse. Externalizing devices
operate so as to distance the natural world from the social world of the scientists.
For example, the very use of the word 'findings' - as referred to above - is
suggestive of things alrea"dy in the world, but found and discovered by investi­
gators. This distancing effect can also be maintained by the use of carefully

Copyrighted Material
132 USING D O C U M ENTS I N S O C I A L R E S EARCH

structured sentences. Thus one might read a scientific report to the effect, say,
that 'we were able to measure for the first time X,Y and Z'. Such a statement
naturally implies that the object or objects being measured exist (in the exter­
nal world) long before the scientist discovers them, and that measurement of
such objects had proved problematic to other scientists. (Naturally, a construc­
tivist reading would highlight how the very act of , measuring', in part, serves
to constitute the object being studied.)
A focus on dimensions as a property of the object, rather than on measure­
ment as a product of the scientists, is usually most evident in cases of scientific
controversy. Thus Collins and Pinch ( 1 993) , for example, provide an excellent
case study of measurement problems in their examination of the cold fusion
debate that arose during the 1 990s, a debate in which various teams of scien­
tists argued about whether observable traces were a product of a new physical
process or an artefact of the laboratory set-up. Latour ( 1 983; 1 987) contends
that these kinds of problems are never resolved by appeal to superior forms
of evidence or 'fact', but simply through techno-political manoeuvring. In
Latour's world it is simply the weight of the laboratory that matters. Strong
laboratories struggle against weaker ones. They are able to carry out more and
more complex experiments, they are able to define the nature of reality, and
when they do so they are able - by sheer weight of personnel - to overwhelm
the opposition with more and more scientific papers.
This vision of science as a contest between heavyweights is an interesting
one. Yet, it does not help us to avoid the fact that heavyweights still need to
persuade in terms of the strategies referred to above. Indeed, both Latour
( 1 987) and Woolgar ( 1 988) have, at different times, drawn attention to the ways
in which the status of scientific facts is also underpinned by the use of what
they refer to as 'modalizers' . Modalizers comment on the factual status of
scientific claims, and can be systematically used so as to bolster one's own facts
and to cast suspicion on the facts of others. Consider the following examples.

Newton's second law of motion asserts that force is equal to mass

multiplied by acceleration (F = rna) .
Newton thought that his equation F = rna took the form of a
universal law.
Newton believed that F = rna held for all time and for all place. It
was left to Einstein to correct this misapprehension.

We can see here that the factual status of the claim F = rna is gradually eroded.
Interestingly it is eroded by associating the claim with human agency. That is
to say, the gradual implication of Newton's beliefs and thoughts underwrites
the eventual claim that the universality of the physical law was somehow in
error. Latour and Woolgar (1 979) argue that these kinds of modalizers are read­
ily found in scientific papers and that the factual status of claims can be alter­
nately bolstered and undermined by their use. The use of modalizers can

Copyrighted Material
 DOING THI NGS WITH WORDS 133

undoubtedly be observed in various forms of scientific discourse, though one

suspects that they are much more likely to be found in spoken rather than in
written discourse.
Some years ago Gilbert and Mulkay ( 1 984) analysed how scientists' dis­
course changed according to context. The two contexts that they focused upon
were the scientific paper and the semi-structured interview. Gilbert and
Mulkay noted a variety of differences in the linguistic registers of the two dis­
courses. For example, in the interviews scientists were more likely to inteIject
references to 'I', 'me', 'we' and 'they' and to account for scientific discoveries in
serendipitous ways. For example, in recounting the moment at which a given
discovery was made, a scientist might refer to 'X' having run into the canteen
with a 'what if?' statement - a statement that within 30 seconds made every­
thing 'fall into place' or hit them like a 'bolt from the blue' , whereas in the
scientific paper all discovery processes are presented as a result of measured and
deliberate manipulations. Sudden bolts from the blue in coffee shops are simply
never mentioned.
The use of what Gilbert and Mulkay called the 'empiricist repertoire'
enabled scientists to portray actions and beliefs as flowing naturally from their
systematic interrogation of nature. In our terms we might refer to it as 'mirror
talk' . In informal conversation, however, scientists tended to eschew this kind
of talk for talk that made frequent reference to networks of workers, personal
characteristics of specific scientists, to hunches and guesswork. Gilbert and
Mulkay refer to the latter as forming the 'contingent' repertoire. In the con­
tingent repertoire clashes between personalities, competition for resources,
political intrigue and the like are all mentioned as an intrinsic part of the
scientific process. And the two, contrasting, forms of discourse are in many ways
represented in talk about the success and failure of scientific work. So scientific
discovery that is subsequently validated is seen as a result of following rigorous
scientific method, whilst scientific discovery that is subsequently rejected or
placed in limbo is seen as a product of personal and contingent factors - as the
work of bad scientists.
The contingent, situated or occasioned character of scientific discovery is
also emphasized in the work of Knorr-Cetina ( 1 983) , though her work draws
as much attention to the role of talk in the laboratory as it does to the role of
text. Thus, she points out that 'the occasioned character of scientific work first
manifests itself in the role played by that which visibly stands around at the
research site, that is by facilities and measurement devices, materials which are
in stock, journals and books in situ in the library' (Knorr-Cetina, 1 983: 1 24) .
Just how such material is called upon in practice has been investigated, in part,
in Chapters 3 and 4 of this book.
The use of the strategies and devices and repertoires referred to in the fore­
going paragraphs is not, of course, peculiar to science and scientific writing. It
can be found in all other arenas of text. Thus, Potter (1 996), for example, draws
on examples drawn from news coverage to illustrate these same processes and

Copyrighted Material
134 USING DOCUMENTS IN SOCIAL R E S EARCH

procedures. More importantly, such procedures are also an integral part of

social scientific reporting itself. In order to see just how that is so, we need to
journey once more - not to the Antarctic, perhaps, but certainly far afield.

From Kiriwana to Inishkillane

Daniel Defoe's novel Robinson Crusoe (1 956 - first published in 1 7 1 9) is a book

that provides, among other things, a masterful representation of the 'anthropo­
logical' eye in the age of reason. The main story concerns a shipwrecked
Englishman who subsequently struggles to subsist on an isolated island. He
finds a 'native' companion, called Friday, and, with him, marches out against
cannibals and mutineers and quells both. Within the narrative we see emerge
some dominant themes of a newly minted capitalist ethic. They concern the
western 'civilisation' of the non-western world (indeed, the word civilisation
first appears around the time of Crusoe) ; the subordination of , nature' to the
will of humankind; and the systematic use of tools to produce and improve the
most rudimentary form of social organization. More to the point, perhaps, one
can view the novel as an allegory of the writing process itself (de Certeau,
1 984: 1 36) . That is, a process in which blank spaces are colonized, a system of
objects is produced by a dominant subject, and the natural and social world
transformed - even created - anew.
As with many things produced by humankind, the social world of others -
strangers, foreigners, outsiders, 'natives' - is and was realized in a special form
of literality. Between (roughly) 1 9 1 4 and (even more roughly) 1 980, that form
could be best found within academic anthropology, and, especially, that mode
of anthropology known as ethnography. (Though it is never quite clear
whether ' ethnography' refers to a textual account of anthropological fieldwork,
or the practical anthropological work that supposedly underpins it - see, for
example, Tedlock, 2000.) In any event, ethnography as a textual genre truly
appears only after the First World War, and it is thereafter set apart from the
numerous traveller's tales, fiction and journalism that surrounded it. Not until
the 1 980s would critics begin to question whether the line that had been
drawn between ethnography and works of fiction was truly impervious.
Ethnographic reports, then, were originally offered to readers as containing
objective scientific accounts of life in other cultures. They were documents
produced by trained anthropologists - people who had intentionally left the
comforts of London and Cambridge behind them, and planted themselves in
native villages across the globe, so as to experience, analyse and describe life
therein. (I shall leave aside for the moment the fact that ethnography - especially
as it was developed in North America - was later used as a method for study­
ing 'exotic' locations in the advanced industrial world.)
One of the earliest and most influential of all ethnographers was Bronislaw
Malinowksi ( 1 884-1 942), who studied the inhabitants of the Trobriand

Copyrighted Material
 D O I N G T H I N G S WITH WO R D S 135

Islands - in the western Pacific - between 1 9 1 4 and 1 920. In many ways it was
he who laid down the basic rules of ethnographic work, and his ' Introduction'
to Argonauts if the Western Pacific (1 922) provides one of the clearest accounts
as to how a twentieth-century ethnographer should proceed. In that I ntro­
duction, Malinowski tells us that ethnography is a science, and he contrasts
such a science with 'speculative theorising' which he refers to as ethnology.
No doubt, Malinowski was here thinking of such work as that produced by
E .B. Tylor in his Primitive Culture ( 1 87 1 ) . The latter, a two-volume work, focused
on mythology, philosophy, religion, art and customs of non-western peoples. It
did so in an overarching theoretical context of human evolution, and it relied
for its evidential base on information gathered from missionaries, sailors,
travellers and other assorted observers. None of his 'observations' were made
by Tylor in person. Instead, he merely compiled, ordered and theorized the data
that had been amassed by others. In place of this somewhat distanced form of
investigation, Malinowski sought to ground anthropology in first-hand obser­
vation - in the science of ethnography. Thus, ethnography, he says, presents us
with 'the empirical and descriptive results of the science of Man' ( 1 922: 9) . It
can only be produced under strict conditions. For example, in order to produce
lasting and reliable evidence about the lives of others, the scientific ethno­
grapher needs to pitch his 'camp right in [the native's] village' ( 1 922: 6) . He or
she should then aim to give a complete survey of the phenomena under study,
and not confine him- or herself to a study of the singular or the peculiar or to
the exotic. Lots of individual cases of events and happenings should be studied.
The ethnographer should 'exhaust as far as possible all the cases within reach'
( 1 922: 1 4) , and Malinowski referred to this precept as a 'method of statistic
documentation by concrete evidence' ( 1 922: 1 7) . Photographs should be taken
where possible. Drawings should be made if appropriate. (Malinowski does not
explicitly tell the reader to use drawings and photographs, but as he incorpo­
rates them into his work I am assuming that he is teaching us by example.)
Minute, detailed observations ought to be kept, and the ethnographer ought
also to compose an ethnographic diary. Characteristic narratives, statements,
typical utterances should also be collected. Results should be presented in
charts or synoptic tables. The final goal of the ethnographer should, however,
be to 'grasp the native's point of view, his relation to life, to realise his vision of
his world' ( 1 922: 25, emphasis in the original) .
Some years after writing that account, Malinowski wrote another account
of Trobriand life: Here are the opening sentences from the Preface to that
publication.

Once again I have to make my appearance as a chronicler and spokesman of the

Trobrianders, the Melanesian community so small and lowly as to appear almost negligi­
ble - a few thousand 'savages', practically naked, scattered over a small flat archipelago of
dead coral - and yet for many reasons so important to the student of primitive humanity . . .
In this book we are going to meet the essential Trobriander.Whatever he might appear
to others, to himself he is first and foremost a gardener. (Malinowski, 1 935: ix)

Copyrighted Material
136 USING DOCUMENTS I N SOCIAL R E S EARCH

Clearly, then, Malinowski's science was not as objective as he claimed (see

Stocking, 1 983) . At the crudest level one might say that it was produced in a
frame of white, European, male assumptions about advanced and primitive
societies; the civilized and the primitive world; the place of the female and the
essential nature of ' Man'. In retrospect (and the retrospect is important here) ,
we need only look at Malinowski's language to recognize the absurdity of his
claims to objectivity, detachment and scientific discourse. Indeed, many years
later, Malinowski's Diary - written during his years in the Islands - was
published (Malinowski, 1 967) . It indicates, among other things, that the ethno­
graphic findings were not generated purely and simply from observational
work. In fact, it appears as if many of his significant 'observations' of native life
(such as those concerning the Kula system) had in fact been stimulated by dis­
cussion with other westerners who were present in the Islands. (We might say
that such westerners were hidden auteurs of the ethnographic report.) And it
is certainly the case that the complete erasure of westerners from his published
accounts of 1 922 and 1 935 enabled Malinowski to present himself as a
detached, isolated, scientific observer of native life. The Diary also reveals that
he actually hated the 'lousy villages' (1 967: 1 29) of the Trobrianders and
thought comparatively little of the natives - often describing them in blatantly
racist terms. The Eurocentric and distinctly 'malestream' terminology called
upon by Malinowski was, however,just one of many ways in which the anthro­
pologist was able to construct an exotic 'other' on which the entire academic
discipline of social and cultural anthropology depended.
During the period in question ( 1 9 1 4-22) the 'other' that was observed was
invariably non-western and non-white. In the next and the following decades,
however, the mysterious and exotic 'other' was to be found in almost any setting.
The immigrant 'zone' of downtown Chicago (Zorbaugh, 1 929), Chicago's gang­
land (Thrasher, 1 927), the conservative setting of a mid-west town (Lynd and
Lynd, 1 929), and the street corner of a Boston neighbourhood (Whyte, 1 955),
were all suitable subjects for ethnographic study - though in the latter cases in
the frame of academic sociology rather than anthropology. But whatever the set­
ting, the key task of the ethnographer was to underscore that notion of ' other­
ness'. For whether it be of the inhabitants of the slum, or the gang, or the
box-car, or of darkest colonial Africa, the theme that the ethnographer wished to
emphasize above all was the contrast between a puzzling 'they' and a familiar 'us' .
Here, for example, is an extract from the Foreword to an anthropological study
of Middle America as represented by the residents of Muncie, Indiana.

Whatever may be the deficiencies of anthropology, it achieves a large measure of objec­

tiviry because anthropologists are by the nature of the case 'outsiders' . To study ourselves
as through the eye of an outsider is the basic difficulry in social science, and may be insur­
mountable, but the authors of this volume have made a serious attempt, by approaching
an American communiry as an anthropologist does a primitive tribe. (Lynd and Lynd,
1 929: Foreword)

Copyrighted Material
 DOING THINGS WITH WORDS 137

One task of the ethnographer, then, is to position him- or herself as a neutral

(but participant) outsider reporting upon the activities and beliefs of strange
and puzzling ' others' . It necessitates a precarious, perhaps impossible, point of
balance. For on the one hand the ethnographer has to present him- or herself
as a mere conduit for the processing of scientific information and, on the other,
as someone who is intimately and deeply involved with 'the native's point of
view'. Indeed, one might ask why, if all that is needed is the native's point of
view, do we not simply ask a 'native' to write up the study? Or, if the ethno­
grapher is merely acting as a conduit for scientific observations, then why doesn't
the ethnographer simply publish the field notes? The answers to such questions
are, of course, clear. The ethnographer is much more than a mere conduit or
a crystal through which the worlds of others are refracted. He or she is a teller
of tales (van Maanen, 1 988) - someone who has to create his or her commu­
nity and represent it in writing.s In that respect, the scientific volumes that
the anthropologist creates have been closely allied to the traveller's tales
recruited by Tylor, and even the literary imaginings of a Daniel Defoe (see
Geertz, 1 988) .
Now, it is undoubtedly the case that the very act of recording the lived
worlds of others in writing changes, in the most fundamental of ways,
the nature of human culture. Thus Goody (1 977) and Ong (1 982) , for example,
indicate how the translation of the symbols and thoughts of people from oral
cultures into writing cannot but fail to alter the very nature of the thought
system that is being discussed. Malinowski's synoptic charts and tables, for
example, put into written form what was only spoken of in Trobriand society
and it is quite clear that the systematization belongs to the Polish-born
observer rather than the western Pacific Islanders. Even Malinowski, of course,
recognized that ethnography involved construction. Thus, in his post-fieldwork
Coral Gardens, he pointed out that:

The main achievement in field-work consists, not in a passive registering of facts, but in
the constructive drafting of what might be called the charters of native institutions. The
observer should function not as a mere automaton; a sort of combined camera and
phonographic or shorthand recorder of native statements. While making his observations
the field-worker must constantly construct. (Malinowski, 1 935: 3 1 7)

Similarly, in the Argonauts, whilst discussing the features of the Kula (see
Chapter 9), Malinowski ( 1 922: 84) had pointed out how the Trobriand
Islanders had no outline of their social structure and that, consequently, 'the
Ethnographer has to construct the picture of the big institution', (emphasis in
original) .
All ethnographies are constructions. Indeed, ethnographic field notes may be
said to construct the very circumstances they claim to describe (Clifford and
Marcus, 1 986) . For, in many cases, what is represented in the text appears there
and there alone. Often, such acts of creation are blatant. Thus, Haddon's family

Copyrighted Material
138 USING DOCU MENTS I N SOCIAL RESEARCH

trees - mentioned in Chapter 1 - were not in any sense visible within Torres
Straits society. In the same manner, Lienhardt's (1961) description of Dinka
cosmology - executed through a constant comparison with 'western' ideas - is
a product of Lienhardt's thinking rather than of the Dinka. And the same might
be said of Evans-Pritchard's (1 937) view of the Zande. Put simply, then, ' [T] exts
do not simply and transparently report on an independent order of reality.
Rather, the texts themselves are implicated in the work of reality-construction'
(Atkinson, 1 990: 7) . How the text puts that reality together is something
that Atkinson (1 990) and van Maanen ( 1 988) have reported on in detail. Indeed
Atkinson, as with van Maanen, suggests that scientific ethnographies should
more correctly be seen as 'persuasive fictions' ( 1 990: 26) . The ethnographer
merely uses various rhetorical tricks to persuade us that the descriptions being
offered are 'real', 'true', 'accurate', 'telling it like it is'. For example, one such
trick Atkinson refers to as 'hypotyposis' ( 1 990: 7 1 ) - the use of highly graphic
passages of descriptive writing to portray some scene or other in a vivid
manner. Here, for example, is an illustrative extract from an anthropological
account of the west of Ireland, penned during the 1 960s. The ethnographer is
talking of his meeting with Joseph - a middle-aged man who had problems
with his 'nerves' .

I n 1 968 came another breakdown. I was present for the first three days of Joseph's dis­
tress. At seven o'clock one morning he began forcing himself to vomit. Straining every
muscle in his diaphragm he sank to his knees by the chair in which he had spent a sleep­
less night. He pulled an old newspaper from the range side, placed it to catch a few gobs
of mucus and bile which ran off his chin. The effort forced tears to his eyes. The crisis had
been precipitated by local opposition to Joseph's intention of selling part of his land to
strangers. (Brody, 1 974: 1 05)

We can see here how Brody positions himself as a direct, first-hand, observer
of Joseph's distress, not only by use of the phrase 'I was present', but also by
moving into a 'thick description' of Joseph's behaviour, and of the setting in
which he found himself (the range, the chair, the newspaper, the gobs of
mucus) . The eyewitness status of the account is, of course, designed to under­
line the authenticity of Brody's description of life in I nishkillane. He knows
of village life intimately, and this intimate, experiential knowledge ofJoseph's
distress is further linked to something bigger - knowledge concerning the
ownership of land in a small community. Indeed, the two issues go hand in
hand, and serve as evidence for a broad and general thesis that Brody is
attempting to establish. Namely, that far from being an integrated and cohe­
sive community, Inishkillane was a community riven by divisions and factions.
There were divisions between the enterprising and the traditional members
of the community, and between the married and the single and isolated.
Joseph, of course, belongs to the latter group. His only friend, apparently, is the
ethnographer.

Copyrighted Material
 DOING THINGS WITH WORDS 139

This interpolation of the eyewitness (or what Geertz (1 988) refers to as

'I-witness') account is a staple of any ethnographer. For, in manufacturing an
ethnographic study the first and foremost task of any author is to establish his
or her 'being there' - on the spot - in the native's camp. Various devices can be
used to underpin such 'being thereness' . Some involve the positioning of the
author - as in the Brody case, above. More usually, they focus on underlying
the 'otherness' of the subjects, such as may be achieved, for example, by the use
and display of photographs and sketches. (One of the earliest attempts to make
systematic use of photographs in ethnographic study was that of Bateson and
Mead - see, Sullivan ( 1 999) - though, as we have already noted, photographs
appear in much earlier works than theirs.) Indeed, ethnographic accounts of
both western and non-western societies are normally peppered with (black and
white) photographs to varying effect (see Becker, 1 986) . But photographs
without eyewitness accounts are usually insufficient to convey the sense of
first-hand knowledge. This is perhaps why, in many of the North American
(sociological) ethnographies, the first-handed nature of accounts is bolstered by
the use of documents written by the subjects of study themselves. Thus
Thrasher's The Gang (1 927) - a study of gang life in early twentieth-century
Chicago - often makes reference in its footnotes to such things as a 'gang boy's
own story', or a manuscript 'prepared by a former gang member' and so forth.
Whyte's Street Corner Society (1 955), on the other hand, uses verbatim conver­
sation of'Doc' and others as data, and this is these days the most common ploy
of the ethnographer. Producing such realist tales, as van Maanen ( 1 988) calls
them, requires the ethnographer to call upon his or her subjects as 'informants'
rather than mere tellers of tales. Indeed, we have already noted in Chapter 5
how social scientists attempted to get the 'native' to speak directly to an acade­
mic audience through the construction of an autobiographical narrative - a
narrative that supposedly avoided the confounding and contaminating effects
of the ethnographer.
The ethnographic text is, then, a form of literality that enrols various rhetori­
cal devices, all designed so as to persuade readers of the text that they are deal­
ing with scientific, objective or factual accounts of first-hand observers, rather
than created images and representations. The anthropologist is thereby posi­
tioned as a mirror on society in the same way that the lab scientist is positioned
to serve as a mirror on nature. Ethnography is presented, so to speak, as a form
of 'factology'. Yet the extent to which the supposedly objective findings of
ethnographers can be replicated by other observers has, on occasion, been a
subject of some dispute. Thus, in Chapter 1 , for example, we made reference to
the manner in which Haddon and his colleagues seem to have completely mis­
understood the nature of 'adoption' in the Torres Straits. It is a worrying find­
ing for realists and positivists, because it suggests that the in-depth analysis of
anthropologists may be at variance with the truth. In this respect, possibly the
most instructive debate relates to Freeman's (1 983) study of Samoa, though

Copyrighted Material
140 USING DOC U M E NTS I N SOC IAL R E SEARC H

similar concerns have arisen in relation to anthropological work In other

regions of the world.
Anthropological texts about the nature of Samoan society had been widely
published by Margaret Mead during the 1 920s and 1 930s (see Mead, 1 928) .
Her focus had been mainly on the processes associated with adolescent develop­
ment. The broad sweep of her argument was that adolescence in Samoa was
relatively free from parent-child conflict and the kinds of sexual repression that
she had observed in her own (North American) society. Freeman, however, on
the basis of his observations - executed during the 1 940s and 1 960s - disputed
this portrait. In place of it, he substituted a picture of unduly repressive parent­
child relationships and of a closely monitored adolescence. It is a conflict of
interpretation that is in large part paralleled in anthropological studies of other
areas of the world - such as those made in the west of Ireland. In the latter case,
the likes of Arensberg (1 937) had witnessed friendship, co-operation and
mutual respect as key features of Atlantic coast culture. Yet, Brody's ( 1 974)
interpretation of the same setting emphasized (as we have seen above) conflict,
antagonism and social isolation. Both of these examples, then, throw doubt on
the role of the ethnographer as one who can offer a transparent window on
the world. And it is interesting to note how such conflicting interpretations can
be reconciled in the context of anthropology as a whole. Thus the realist or
positivist route to reconciliation emphasizes the time differences involved in
the respective fieldwork (suggesting that had the various anthropologists been
present in the same society at the same time, then they would have 'seen' the
same things) , whilst the post-modern turn in anthropology uses alternative
interpretations as evidence for constructivist claims. Our interest in this chapter,
of course, is not in the relationship between an anthropological account and the
external, independent world that it claims to report upon, but in how that
anthropological account is assembled in text.
Naturally, in the pre- 1 980 world - a world in which ethnography was
widely regarded as a 'scientific' exercise - it was possible to find reference to a
number of 'how to do it' texts and statements (as, for example, in Whyte's
( 1 955) methodological appendix, or that of Radcliffe-Brown, 1 958) . Such
how-to-do-it texts provided rules that could be, and often were, enrolled by
later authors to justify the status of new work as rigorous and systematic, and
to show how such work belonged to that corpus of scientific observation that
was ethnography. At times, the how-to-do-it accounts plunged into detailed
confessional tales in which the author related his or her innermost feelings
about the field to the reader. Geertz (1 988) reviews a number of examples.
Such confessional forms (van Maanen, 1 988) invariably sought to blur the line
between objective accounts and subjective impressions and may well have con­
tributed to the demise of scientific ethnography as a whole. In any event, and
for whatever reasons, after the 1 980s, fieldworkers tended to turn away from
describing how ethnography was done 'in the field', and towards an examina­
tion of how it was manufactured on the page - that is, to the study of rhetoric.

Copyrighted Material
 D O I N G T H I N G S W I T H WO R D S 141

To recognize that an anthropological account is dressed in rhetoric does not

mean, however, that it is contaminated and ought not to be. Rhetoric is simply
unavoidable in an academic or any other form of text. The task is to discover
and note the rhetorical devices that are used to produce reports. (Such as the
use of distancing styles, the mechanisms that draw an abrupt division between
the author of the ethnography and the informant through the use of quota­
tions, the use of photographs and so forth.) Anthropological narratives, of
course, are themselves composed out of field notes. Yet there are relatively few
accounts as to how field notes and the narrative of an ethnography relate to
each other (Atkinson, 1 992a; Burgess, 1 982; 1 984) . Certainly, field notes often
contain their own narratives, and may well be written with an eye as to how
they will ultimately relate to the final (published) text (Atkinson, 1 992b) . In
most cases, however, such narratives will be routinely fractured, fragmented and
decontextualized in accordance with the methodological precepts of fieldwork
manuals. These days such fragmentation is fashioned as much by machines as
by humans - for example, through the use of such data processing programs as
NUDe 1ST or NVivo (see Richards, 1 999) , programs that tend to be built
around the all-encompassing rhetoric of ' grounded theory' (Glaser and Strauss,
1 967) . Grounded theory, of course, supposedly allows theoretical insight to
emerge out of the data themselves, the researcher merely rearranging and making
explicit that which is implicit in the data.Yet, the fracturing of observational or
conversational data into 'nodal' points - based on themes and concepts selected
by the researcher - and the rearrangement of data via the use of , memos' and
research reports fundamentally alter the data that have been assiduously collec­
ted. In that respect, the social scientific researcher invariably and inevitably
imposes a new order on the world as reported by his or her informants (for
some examples see Charmaz, 2000) . As Goody ( 1 977) suggests, the act of
writing is never a neutral process.

Conclusions

In their Official Discourse Burton and Carlen (1 979) focused on the discourse
contained in official (UK) government reports on various law and order
problems - many relating to street disturbances in Northern Ireland during the
late 1 960s and early 1 970s. They were interested, for example, in how such things
as 'common sense', 'natural reason' and the like were recruited by the writers of
such reports so as to explain and account for the events under scrutiny. Yet,
Burton and Carlen's analysis very naturally operates within another form of
discourse (no writer can escape the iron cage) . In this instance it was a social
scientific discourse of a special kind. In particular the authors composed their
book in a conceptual grid drawn from the work of the French structuralists
and especially those of a Marxist bent. And from this distance it is interesting
to ask why they felt it necessary to recruit, so fervently, the likes of Althusser

Copyrighted Material
1.42 USING DOCU MENTS I N SOCIAL R ESEARCH

and Lacan (see Dosse, 1 997a and 1 997b) into their work. One reason must have
been that French structuralists formed a powerful intellectual group in the
Europe of the 1 960s and 1 970s.A serious intellectual work in the social sciences
had therefore to recognize and acknowledge that power, and to engage with
the debates of its members. One wrote sociology by referring to the convo­
luted language of the Parisians in the same way that, during the late 1 950s and
1 960s, one wrote sociology by engaging with the convoluted language of
Talcott Parsons. But who now reads Parsons? (parsons had once posed the ques­
tion, 'who now reads Spencer?' - given that Spencer had been one of the most
influential and powerful of figures in nineteenth-century biology, geology and
sociology.)
The recruitment and abandonment of authors and networks of authors
forms an interesting field for investigation in itself. How authors circulate and
are picked up and set aside can tell us much about the natural history of a dis­
course. An investigation of citation networks (Baldi, 1 998) in scientific papers,
for example, could well tell us about the various 'core-groups' (Collins and
Pinch, 1 993) that are actively engaged in a scientific controversy. Such a task
lies well beyond the aims of the present chapter, however. For here I have
elected to concentrate on what is in the text rather than on how the text is
recruited by networks of authors. In particular, we have looked at the ways in
which authors structure specific forms of literality, and how they locate them­
selves within such forms. More specifically, it has been suggested that the fol­
lowing questions constitute useful entrees into the analysis of the resources that
social scientists use in the research process.
How is the author-function played out in the document? Exactly who author­
ized the document, and how were they positioned in the production process?
How are authors made visible or invisible? Are we dealing with the work of
absent and unmentioned actors - such as coding clerks, an 'office', or hidden
Europeans? Or are we dealing with one of d'Hemery's identifiable 'auteurs'? Is the
work offered to us as the product of a team or of a committee - as is the case with
DSM-IV (see DSM-IV 'Acknowledgements') - or as the work of a lone scholar?
How does the rhetor position him- or herself in the creative process and to
what desired effect? As a first-hand observer (an 'I-witness')? An impartial
observer? A collector of tales? Or as a mere conduit (a 'mirror') for the trans­
mission of scientific data and information? Indeed, at what points does the
rhetor turn him- or herself from observer into participant, and at what points
does he or she appear as an absent and invisible cipher? For it is clear that
ethnographers in particular have often sought to present themselves in text in
a dual role - as an intimate of'others' and, at the very same time, as distanced
and objective observers of exotic life. (Social scientific interviewers usually
position themselves similarly.)
What resources were exploited in the process of production? In particular,
what papers and books were recruited by the author(s) to underpin their
rhetoric? What kind of network does the citation and recruitment process

Copyrighted Material
 DOING THINGS WITH WORDS .143

inscribe? These, of course, are questions that relate to issues of intertextuality -

see Chapter 6, note 7 .
How d o the documents structure their readers? That i s t o say, what specific
audience is structured by the title of the document, its place and manner of
publication? Thus Cherry-Garrard, with whom we opened this chapter, struc­
tured his narrative to appeal to a 'general' audience and eschewed an appeal to
the scientific community of his day. The very title of his work, The Worst
Journey in the World, was markedly different from that which might have been
expected to symbolize a scientific report. In exactly the same way, a paper enti­
tled, say, 'The arthropods Mimetaster and Vachonisia from the Devonian
Hunsriick Shale', in Palaontologische Zeitschrift 1 978 (cited in Gould, 1 989) ,
structures its readers in a radically different mode from a book or paper enti­
tled 'Looking for fossils' . ('We can only wonder about what Malinowski
intended by entitling one of his books The sexual life of savages.)
Finally, one needs to ask how, through the use of textual devices, the docu­
ment structures the world. In this respect, we should also be aware that it is not
simply ethnographers who have shaped the worlds that they claim to have expe­
rienced in objective and distanced measure, but also interviewers in general. The
analysis of interview data as manufactured text, however, lays well beyond the
remit of this chapter. With that in mind I shall draw attention to one further
limitation of our analysis so far - namely, that we have focused almost entirely
on literary stratagems, and somewhat ignored the role of diagrams and illustra­
tions. Yet, we know that scientific and other texts often use illustrations and
other images (see Bastide, 1 990; Lynch, 1 985; 1 990) to convey a notion of frozen
factuality about the objects being discussed. Just how they achieve that goal is a
subject worthy of enquiry in itself - as we have seen in Chapter 4.

RESEARCH EXERCISES

Exercise 7.1

In the l ight of the discussion above, examine ways i n which the social
scientific interview has been represented in some key methods texts
since the 1960s. Pay particular attention to the manner in which the
interviewer is ' positioned' vis-a-vis the interviewee. A useful starting
point for such an exercise might be Fontana and Frey (2000).

Exercise 7.2

More specifically, and using the work of feminist writers and writers on
feminist methodology, examine how the interviewer has been reposi­
tioned with respect to the interviewee. Some useful starting points may
be found in DeVault (1987 ; 1990), Oakley (1981) and Reinharz (1992).

Copyrighted Material
144 USING DOC U M ENTS IN SOCIAL R E S EARCH

Notes

1 The notion of genre is itself worthy of investigation, see Dubrow (1 982) .

2 An interesting philosophical position on the manufacture of facts and things -
especially in relation to 'science' - is provided by Heidegger ( 1 967), in what must
be one of his more readable works.
3 For the likes of Derrida ( 1 976; 1 988). of course, this position is untenable since a
text references only itself and, by the nature of language, can have no referents to
some objective ('rea!') world beyond the text.
4 On the function of the footnote in scientific discourse, see Grafton (1 997).
5 We must not overlook the possibility that the ethnographer's informants have also
to construct his or her community and culture in order to represent them. So the
act of construction becomes a multi-voiced process - see, for example, Rabinow
( 1 977) . The role of informants in the construction process has, however, been little
researched.

Copyrighted Material
8
Doc u m ents as Evidence .
Researc h i ng the I nert Text

Conjectures and refutations 145

Evidence, evidence, evidence 148
The selection of evidence 150
The scope and robustness of data 154
Data extraction: indexing and coding 157
Conclusions: conjectural history 161
Notes 164

Conjectures and refutations1

A belief in the existence of a single, common source of civilization has been a

persistent one in human history. It is held even today, at the commencement
of the twenty-first century, where it often takes on fantastical forms - some­
times involving references to the existence of lost civilizations such as that of
Atlantis. Atlantis appears in ancient Greek myth as a vast and important land
situated somewhere 'beyond the Pillars of Heracles' . Its people and rulers were
defeated by the Athenians. The land was drowned by an act of the gods. In
modern mythology it is often taken to symbolize a sophisticated and mysterious
civilization that served as a key source of human knowledge - knowledge
about the domestication of cattle and the cultivation of seed, the smelting of
metals, the skills of navigation and astronomy, and much else beside. Among the
'much else' we might include building techniques, techniques and effects that
often appear to be similar in different parts of the ancient world, such as is the
case, for example, with the construction of pyramids.
Pyramids can be found in ancient world sites of Asia, Africa and the
Americas. On the basis that building structures are similar in different regions
of the ancient world, a number of amateur historians and archaeologists have
sometimes erected entire theories. Thus, Graham Hancock for example, in his
Heaven 's Mirror: Quest for the Lost Civilisation ( 1 999), has argued that human
knowledge about building technique and the heavens was passed down from
the members of a once-mighty civilization that had reached its zenith around

Copyrighted Material
146 USING DOC U M ENTS IN SOCIAL R ESEARCH

1 0,500 years before the Christian era (BCE) . At one point he surmised that the
single, central civilization was buried under the Antarctic ice, and so he does not
actually refer to Atlantis per se. However, he does claim that the members of
the lost civilization were great astronomers. They held detailed knowledge of
the stars and the constellations, and they left a legacy of information that was
later spread throughout the ancient world. We can see evidence of their handy
work even today. For example, the Great Pyramids at Giza are exactly aligned
so as to mimic the position of the stars in Orion's belt, whilst the temples of
Angkor Wat, in Cambodia, trace out on the ground the pattern that the con­
stellation of Draco forms in the northern sky (or, more precisely, as Draco
would have appeared to a human eye 10,500 years BCE). The strange city of
Tiwanaku near Lake Titicaka in Bolivia has similar (though, as yet, unfathomed)
astronomical properties. The evidence is present for all to see. Unfortunately,
according to Hancock, academic archaeologists are blinded by prejudice and
outdated theory, and unable to recognize the larger patterns that these observa­
tions point towards.2
The Great Pyramids at Giza (there are three particularly close together) are
indeed arranged in a rough line, where a smaller pyramid is offset from the
other two by some 40 degrees. In that respect they do appear to mimic the
three stars that lie in Orion's belt (one star offset from the other two by about
50 degrees to the north) . What is more, we know that the fourth-dynasty
Egyptians, who built the pyramids, were particularly interested in the cosmos
and oriented their tombs to reflect some of its features - especially the cardi­
nal points of the compass. They certainly knew where the direction of north
was, and they would have been well able to mimic the stars in Orion's belt
exactly - had they so wished. It is something of a puzzle, therefore, as to why
the smaller of the Giza pyramids is offset to the south and not to the north. In
other words, the layout on the ground, and the stars in Orion's belt, only match
exactly if we turn Egypt upside down, or if we turn the sky upside down and
leave Egypt as it is - whichever is the easier.
So our first piece of evidence only fits with the theory if we ignore the
cardinal points. Unsurprisingly, it is not an issue that is addressed in Hancock's
book. But let us proceed further.We know that there are at least 1 6 stars in Orion
and some 80 pyramids in Egypt. Yet it is only a (very bad) fit of three stars and
three pyramids that are selected for inclusion in the evidential base. And the same
problem arises with Angkor Wat and Draco. Again the constellation of Draco is
simply not matched by the temple pattern, and Hancock uses only 1 0 out of
more than 60 temples to support his case. More importantly, perhaps, from our
understanding of Khmer society we remain unaware that Draco (the dragon) had
any place in Khmer culture whatsoever. For neither dragons nor the constella­
tion seem to have figured in Khmer religious or political life. So, in the language
of social science, the evidence just does not 'triangulate' (Denzin, 1 978) .
So here we have a book that is classified by the publisher as 'non-fiction'.
The author presents his case in an academic, authoritative rhetoric. He cites the

Copyrighted Material
 D O C U M ENTS AS E V I D E N C E 147

work of archaeological experts and authorities. He talks of conjecture, evidence

and proof. To all appearances we have a thesis, and reference to substantiating
evidence. In that respect the book conforms, superficially at least, to a traditional
(positivist) scientific model. Yet the bits of evidence that we have so far exam­
ined have already thrown up a number of problems. The evidence is highly
selective. We are not provided with good reasons why confirming instances (of
matches between stars and buildings) are chosen whilst others are ignored.
Furthermore, we would expect the evidence to be consistent over many
sources. For example, if the · constellation of Draco were so central to Khmer
builders then we would expect to find references to its significance in various
symbols systems and drawings from the relevant period - but we do not.
What is worrying about Hancock's claims, then, is that they seem to be
supported by the flimsiest kinds of evidence. And that raises issues about what
'good', reliable evidence might look like. Unfortunately, that is not an easy
question to answer. One of the reasons for difficulty lies in the fact that - as
we know from sociological studies of scientific controversy - what counts as
good, reliable and well-validated evidence is always open to social negotiation.
Thus, in their analysis of specific controversies (such as the cold fusion debate,
for example) Collins and Pinch ( 1 993) point out how what counts as evidence
to one party of scientists often serves as no more than mere conjecture to
another.
Our interest in debates about evidence, of course, relates only to the use of
documents as sources of evidence. The topic is an important one because, as I
have indicated previously, in matters of social research, documents usually
appear only in so far as they serve as receptacles of evidence for some claim or
other. Indeed, as I have been keen to emphasize throughout the book, social
researchers are far more interested in asking questions about what documents
contain than with what people do with documents and how they manipulate
them in organizational contexts. Consequently, as researchers of the inert text
it would undoubtedly be of considerable help to us if we could appeal to a set
of generally accepted rules about evidence to demonstrate that our scrutiny of
document content was done 'in the right way' and to the highest standards -
rules that would help establish that our ultimate claims are valid and reliable.
Unfortunately, no such body of rules exists. That is not to say that there are no
rules (see, for example, Platt, 1 98 1 a; 1 98 1 b; Seale, 1 999) , only that their status
is always contested.
In what follows I am going to make reference to a rhetoric that relates to
evidence. It is a rhetoric that concentrates on distinguishing 'good' from 'bad'
evidence and it is one that is widely called upon in contemporary medical
research. In many respects it is a rhetoric that fits rather badly with the
demands of qualitative research. Nevertheless its key concerns serve to high­
light some issues that are central to persuading readers that one's research
results are robust, reliable and valid - good enough, in fact, to stake one's health
and well-being on them.

Copyrighted Material
148 USING DOCU MENTS I N SOCIAL RESEARCH

Evidence, evidence, evidence

Social scientific debates concerning the use of suitable evidential foundations

for the acceptance of theories and hypotheses are at least as old as the sciences
themselves. John Stuart Mill's (1 843) attempt to 'systematize' the rules of
scientific evidence and scientific method is among the earliest of such strategies -
and it still has an impact (see Lieberson, 1 992) . Mill, of course, wrote in the
shadow of the Enlightenment, an age where trust and faith in reason and truth
(especially as exemplified in a codified scientific method) was commonplace. It
seems a long way from Mill's System if logic to a late modern age where as far
as methodology and much else are concerned, 'anything goes', and where
science is regarded as 'much closer to myth than scientific philosophy is pre­
pared to admit' (Feyerabend, 1 975: 295) . Indeed, we might characterize our
own world as a world in which all claims are regarded as provisional and ten­
tative, and even equivalent; story writing every bit as good as traditional
ethnography (Denzin, 1 988) .
Now, one of the most powerful and persuasive rhetorics of science ever
developed emerged out of what is often referred to as the 'positivist' move­
ment. What positivism is, how it is expressed and how it came to dominate
western scientific discourse is a complex field of study in its own right
(Giddens, 1 978; Gillies, 1 993) . At this point we need focus on only a few of
its central tenets - in particular those that concern the application of what is
sometimes called the hypothetico-deductive method of scientific discovery.
The latter was developed in large part by post-positivist philosophers such as
Karl Popper ( 1 959; 1 963) and Ernest Nagel ( 1 96 1 ) during the first half of the
twentieth century. The method has many intricate features, but included
among them are the suggestions that science proceeds not on the basis of
making lots of unconnected observations, but rather on the basis of advanc­
ing theoretically informed hypotheses - hypotheses that, ideally, incorporate
law-like statements. That is, statements of the type advanced by, say, Sir Isaac
Newton to the effect that 'force' equals 'mass' multiplied by 'acceleration'
(F = Ma) . Or the even more famous Einstein equation that energy equals mass
multiplied by a constant squared (e = me), where the constant c is the speed
of light. In the rhetoric of positivism, however, such laws and hypotheses
are always provisional. In other words, we should hold to them only as long
as there is evidence to support them, and all and any evidence should be
collected and analysed in an objective and dispassionate manner. ('Experimen­
tation' is, of course, the preferred strategy for collecting evidence.)
Disconfirming instances of evidence should carry special weight. In other
words, evidence that does not fit a hypothesis should carry more weight than
any number of pieces of evidence that do. Unfortunately, we know from
empirical studies of what scientists actually do that there is invariably a diver­
gence between the theory as outlined above and what happens in practice -
see, again, Collins and Pinch (1 993).3

Copyrighted Material
 D OC U M E N TS AS EVIDENCE 149

During the 1 950s-1970s there were many who argued that the hypothetico­
deductive method of enquiry should be applied every bit as much to social
scientific research as to research in the natural sciences (see Brodbeck, 1 968) .
It was not, however, an argument that persisted quite so strongly beyond the
1 970s. Yet social scientists have been and are always concerned with rules of
good practice and with amassing reliable and valid data sources. In fact, social
scientists often make use of the twin concepts of reliability and validity to
distinguish between good and bad forms of evidence. The concept of 'reliabil­
ity' in social research refers to the requirement that the 'findings' of any research
programme are independent of the particular circumstances in which the
research was carried out. In other words, any researcher in similar circum­
stances and adopting the same research strategy should get more or less the
same results as the original researcher. The concept of validity relates to the
issue as to whether the research findings are actually providing appropriate and
valid evidence for the claims being made by the researcher. With respect to the
Graham Hancock example above, of course, the suggestion is being made that
his findings are neither reliable nor valid.
Naturally, most commentators on research practice usually fragment the
demands that results should be' reliable and valid into a series of i,ssues about
data collection and analysis. And so it is with people who comment on
research with documents. Thus, Burgess ( 1 984) , for example, insists that
matters of authenticity (of the documents) and distortion should be taken
into account. Platt ( 1 98 1 a) states that the researcher should pay attention to
issues of authenticity, the availability of documents, sampling of the docu­
ments and what inferences can be made from the documentation that is
studied. Scott (1 990) lists authenticity, credibility, representativeness and
meaning as matters to be attended by the researcher. Such lists of criteria are
commonplace and Seale ( 1 999) provides an excellent overview of the various
criteria that have been applied in judgements about the virtues of qualita-
. tive research in general.
Rather than review lists of criteria, however, it will be much more useful for
us to refer to an existing template of what is regarded in a scientific commu­
nity as good practice. The template that I am about to refer to is conunonly
drawn upon in the framework of what is called evidence-based medicine
(Sackett, 2000) . The latter is part of a strategy that was adopted in western
medical practice during the closing decades of the twentieth century. Its aim is
to determine what is and what is not 'effective' - in the way of drugs, therapies
and medical interventions of all kinds. Crucially, from our point of view, it
achieves results through the review and analysis of document content.4 Though
the documents reviewed are invariably restricted to those that contain specific
kinds of research findings. Nevertheless, it is argued that by examining docu­
ment content in terms of a strictly defined set of procedures, researchers can
produce robust and reliable conclusions about the effectiveness of professional
practice. The procedures add up to what is sometimes called systematic review

Copyrighted Material
. 150 USING DOCUMENTS I N SOCIAL RESEARCH

methodology. It is a methodology that may be looked upon as a form of modern

medical technology, a form of technology that recruits into its frame information
processing techniques, statistics, mathematics and social science.
In what follows I am going to deal with what are often regarded as three
central issues of systematic review procedures. They concern ( 1 ) the selection
of evidence, (2) the scope and robustness of data, and (3) data extraction. As we
shall see, it is simply not possible to lift the systematic review template off the
shelf and apply it to qualitative work without important modifications - even
if we limit ourselves to just three areas of interest. Nevertheless, the general
stance of the review techniques has a virtue that extends far beyond the realm
of medicine.

The selection of evidence

In terms of the rhetoric of good research practice, one of the problems with
the Hancock example, provided above, is that the stars and buildings that he
matched together were selected on a highly idiosyncratic basis. That is to say,
not all stars and not all buildings were selected for the analysis. Sampling of
populations - whether of people, things or events - is, of course, a legitimate
procedure. But as we have stated above, where it does occur, then the selection
process needs to be justified. Ideally, and according to systematic review tech­
niques, the reasons for including and excluding cases ought to be defined in
advance of any study. It is simply not good enough to select cases that fit an
hypothesis, and to ignore those that do not. In quantitative research the central
requirements for the selection of items from any population are usually
twofold. First, once the criteria for including or excluding documents in a
study are set, the subsequent selection of documents should result in a repre­
sentative sample of the sum total. Second, the selection should be unbiased or
random. In qualitative research, of course, the selection of , cases' can be justi­
fied on numerous grounds, and it is not always possible to define in advance
what is to be studied (see, for example, Charmaz, 2000).
Representativeness and randomness usually go hand in hand in quantitative
research practice. Even in qualitative research - with or without documents -
randomness can have its virtues. Thus a study of letters, wills, contracts, certifi­
cates or whatever that is based on a random selection from a wider population
of such documents will be regarded as producing more reliable conclusions
than a study that is 'biased' and restricted only to those that fit the arguments
being advanced. Random selection may also have other benefits. For example,
in the absence of random principles, the temptation to focus on the strange,
the exotic and the unusual at the cost of the uneventful and boringly normal
might be difficult to resist. Perhaps that is why Malinowski ( 1 922) always took
care to warn his followers to aim for complete coverage of the phenomena

Copyrighted Material
 DOC U M E NTS AS E V I D E N C E .15.1

under study and not to focus on the abnormal and exotic. I found this to be a
valuable principle in my own work on death certification (Prior, 1 989) . It was
work in which I found complicated sequences of cause of death far more
interesting than straightforward sequences. Without adopting the principle of
random selection (I opted for a simple random sample of 10 per cent of all
cases in a twelve-month period), I might easily have weighted the sample
unfairly towards the inclusion of 'difficult' and atypical medical narratives
of death.
An essential first step to random case selection, then, must involve identifi­
cation of all of the members of a given population. This, in turn, should lead
to the construction of a sampling frame (de Vaus, 1 996) . When we are dealing
with people we can normally call upon lists or registers - such as a list of death
certificates - from which to sample. No register ever made is without its prob­
lems, but most of the widely used types of register (such as electoral rolls,
school registers and so forth) have a good coverage of their populations. Such
frames are easy to identify and sometimes easy to get hold of. With documents,
however, life is rather more difficult. For example, there are many e-mail letter
writers in the world, but we have no rational way of finding out who they are
or what they are writing about. Nor do we know how what is saved in the way
of documentation differs from what is lost. Hence, without a reasonably com­
plete list of e-mail letter writers or letters written, any study of letter writing
would risk the possibility of focusing on only a small and highly unusual selection
of cases.
In their fascinating work on 'correspondence' Chartier et al. ( 1 997) provide
some interesting insights into the above problem. In the course of their
various research programmes into styles of letter writing from the Middle
Ages to the nineteenth century, for example, they focused on 'secretaires' or
manuals devoted to instructing their readers as to how a letter ought to be
written. One of the great virtues of the 'secretaires' is that they were published,
and we know that published documents are always listed (somewhere) . Further­
more published documents have a considerable chance of survival across the
centuries. So by focusing on the manuals rather than individual letters, it is possible
(if not desirable) to obtain a complete listing of relevant items. I t might also be
the case that by focusing on the letter templates, rather than letters first hand,
the researchers could be reasonably certain that a complete range of letter
styles were available for study. In the nineteenth-century case it was possible to
check the influence of the 'secretaires' on actual letters written by studying a
pragmatic sample of over 600 letters written between 1 830 and 1 865 and
lodged in the Paris Postal Museum. In addition, Chartier et al. , also encoun­
tered data from a French postal survey of 1 847 that listed almost every letter
posted in France from almost every postal box during the month of November
of that year. The latter were used to provide a picture of communication in
France during the mid nineteenth century. (Note how, in this example, the

Copyrighted Material
152 USING DOCUM ENTS IN SOCIAL R E S EARCH

1 2 ,-----

10

BMJ La ncet JAMA NEJ M

10 1 967- 1 97 1 0 1 972-1976 • 1 977-1 981 0 1 982-1986 • 1 987-1 99 1

FIGURE 8 . 1 The growth of risk as a focus for medicine: percentage of articles with
'risk' in title or abstract (after Skolbekken, 1995)

circulation of letters (documents) serves to map out the skeletal nature of a

social network.)
In a similar manner, of course, one could use travel manuals, recipe books
and other types of document as templates for other everyday activities, though
whether people travelled or cooked in accordance with the manuals is forever
a matter of empirical enquiry. Social scientists who focus on aspects of con­
temporary life can, of course, always ask their subjects to produce documents
anew. This is often the case when people are asked to produce diaries and
other accounts specifically for the research process. It is a well-established strategy,
and Plummer (200 1 ) lists numerous examples of it.
Another possibility is for the social scientists to request documentation that
people might ordinarily produce or use in their everyday lives. Such 'naturally
occurring' forms of documentation were called upon in one of the most
famous of all twentieth-century American sociological studies The Polish
-

Peasant in Europe and America. In that study, the authors, Thomas and Zaniecki
(1 958) , obtained their letters (754 in total) by advertising for them in a Polish
emigre journal. They offered letter holders 1 0 to 20 cents for each letter.
Unfortunately, the sociologists were not entirely clear about the process of
acquisition. Their acquisition strategy became known only many years after
the publication of the original study (Madge, 1 963) . And although the collec­
tion contained letters from almost every strata of Polish society, it could by no
means be regarded as a systematic sample of available letters. That is to say, it
was not random.

Copyrighted Material
 DOC U M ENTS AS EVIDENCE 153

Information technology can of course assist in document sample strategies -

though the sheer weight of results can often overwhelm the novice researcher.
Thus, by using computerized search devices it is not too difficult to come up
with a volume of results for documents that are listed on some database or
other. Thus, a search for, say, a full list of scientific papers on 'suicide' can easily
be obtained from a database such as Medline5 or Sociological Abstracts 6 . The diffi­
culty, of course, is not so much to get results but to limit the results to a man­
ageable size. Figure 8 . 1 (adapted from Skolbekken, 1 995) illustrates the potential
use of such searches. In this case the searcher was seeking evidence for the claim
that, as far as medical publications were concerned, an interest in 'risk' grew
substantially and significantly during the last quarter of the twentieth century -
that, in fact, there had been a risk epidemic in the realm ofjournal publications.
(The four medical journals included in the figure are The British Medical Journal
(BM]) , The Lancet, the Journal if the American Medical Association (lAMA) and the
New England Journal if Medicine (NE]M) .) As is evident from the chart, a simple
search of appropriate databases using 'risk' as the only key word produces a quite
stunning and persuasive result on the general thesis.
Unfortunately, and as we shall note shortly, not all research results find their
way into mainstream journals. Some research documents can be particularly
difficult to trace and to list. For example, certain kinds of hard-to-trace items
are sometimes referred to as the 'grey literature'. Fortunately, there is a database
(known as SIGLE? for such literature, and it often throws up otherwise
neglected items of documentation. But why some literature is 'grey' and
remains 'grey' is an interesting issue in itself.
As I have already suggested, calls for random and representative samples of
data are to be heard mainly from the mouths of survey and quantitative
researchers. With qualitative research, however, it is often the case that there are
sound reasons for side-stepping such requirements. For example, there might
be good reasons for selecting only certain kinds of document or certain kinds
of event where documents come into play. Thus, in his study of laboratory life,
Lynch ( 1 985) focused specifically on 'agreement' and instances of interaction
where agreements and disagreements arose - some of which involved dis­
agreements about the use of documents. Platt ( 1 996; 1 98 1 a; 1 9 8 1 b) has argued
that, in the history of both psychology and sociology, the (highly selective) 'case
study' method has played a major role in social research. As far as research with
documents is concerned, the bulk of case studies involved the use of life
histories (see Chapter 5), though case studies have also been used to study
organizations and events as well as people. (A good modern example of a case
study is provided by Vaughan, 1 996). Naturally, case studies by their very nature
cannot be representative. However, they can be indispensable for developing
theoretical insight, and for examining the fine detail of social life. (For a justi­
fication of the method, see Yin 1 994.)
We should also recall that in qualitative work the use of sampling to refine
ideas rather than to satisfY the demands of calculation is a well-established

Copyrighted Material
1.54 USING DOC U M E NTS I N SOCIAL R ES EARCH

principle. Thus, qualitative researchers often develop sample strategies that are
deliberately aimed at gathering cases that are not in any sense representative of
the general population. I nstead the sampling is devised so as to gain a particu­
lar insight into social processes (see, for example, Charmaz, 2000) . Thus, in
work with documentation, the use of such principles as purposive, pragmatic
sampling and 'theoretical' sampling could easily be justified. The notion of
theoretical sampling was, of course, popularized by Glaser and Strauss ( 1 967: 45)
in their discussions of grounded theory. According to their principles researchers
should select cases for study on an individual basis - seeking at all times to
expand the scope of their data and variations in the social practices being
studied. Furthermore, a researcher ought always to be on the look-out for cases
that disconfirm any current hypotheses. Indeed Glaser and Strauss argue that
sampling ought to proceed up to that point where no new data - and no dis­
confirming instances - seem to be emerging. It is a stage that they refer to as
'theoretical saturation' (1 967: 6 1 ) . The practical application of the methodology
was best illustrated in studies relating to dying trajectories - see, for example,
Glaser and Strauss ( 1 965) .
What matters above all, of course, is that the researcher should specifY in
detail why it is that the cases selected for study have been so selected, and what
the limits of the selection process might be (see Seale, 1 999) . With the
Hancock example that opened this chapter, it is not at all clear why discon­
firming instances of the general hypothesis (about stars and building layout)
were ignored and one is left wondering whether it was merely because the data
failed to fit the author's generalizations.

The scope and robustness of data

The technology of the systematic review depends, in part, upon the classifica­
tion of evidence into a hierarchy of reliability and 'strength' (see, for example,
http:// www.york.ac.uklinst/crd/report4.htm). At the top of the hierarchy are
research studies that concern experimental designs, in particular, designs in
which the allocation of cases to experimental and control groups has been
randomized. Such studies are often known as randomized controlled trials
(RCTs) . The growth, development and influence of RCTs in medical practice
is a worthy topic of investigation in itself, though one that falls far outside the
remit of this book (see, for example, Matthews, 1 995) .
Following the RCT in the contemporary hierarchy of evidence are other
trials where random allocation of individuals to experimental and control
groups has not been possible for some (good) reason. Mter that come what are
called cohort studies. These commonly rest on the analysis of existing groups
in society (say, age and gender groups) that can be compared in some system­
atic manner one to the other. The aim is to determine what kinds of factors in
two or more groups might lead to differences in outcome with respect to a

Copyrighted Material
 DOCU MENTS AS EVIDENCE 155

health intervention. For example, one might compare rates of lung cancer
among groups of people who smoke cigarettes and those who do not. At the
bottom of the knowledge hierarchy come 'opinions of respected authorities',
descriptive studies and reports of expert committees. Qualitative research
would probably be slotted in at this point - were it to figure at all.
So, in terms of qualitative research, the hierarchy of evidence outlined above
has little relevance. I have drawn attention to it, however, so as to emphasize
the fact that judgements about the reliability and robustness of evidence are
socially (in this case professionally) based. Nevertheless, and irrespective as to
whether one is engaged in quantitative or qualitative research, such judgements
have always to be attended to.
In the realm of documents, and with respect to reliability, it is clear, for
example, that evidence can often be forged. In Chapter 1 , I mentioned the
issue of forgery with reference to Ann Frank 's Diary and The Diary of a
Nobody. 8 Scott (1 990) , in his analysis of research with records, pays consider­
able attention to the issue of forgery in available documentation. Forgery and
fakery can, of course, be a problem in all fields of human endeavour. And
in some contexts it could be just as interesting to study forged as to study
genuine documents. 'Scientific' forgery or forgery ofscientific findings is especially
instructive as it highlights the nature of problems relating to evidence.9 A more
pressing problem in qualitative research, however, is more likely to arise from
the fact that what is studied is carefully selected, rather than forged or faked.
And biased selection might not be intentional - as it seems to have been with
Hancock's selection of stars and pyramids. To illustrate how this might be so we
need to examine Figure 8.2.
Figure 8.2 is called a funnel plot. In this particular case it is derived from a
study of ' risk communication' in the context of patients and medical practi­
tioners (Edwards et al., 2000) . Medical practitioners often have to communi­
cate information about, say, the attendant risks and benefits of different kinds
of treatment. For example, hormone replacement therapy that is provided to
older women has benefits for such women as far as the onset of certain types
of arthritis are concerned, but also carries a measure of risk as far as other
matters are concerned (as with, say, risk of thrombosis) . But how can such
information be best communicated? For example, when practitioners commu­
nicate information about risk it is not always clear whether they should use
only words (such as high, moderate and low risk) or whether they should
use numbers (percentages) , whether they should use diagrams or whether they
use only language and so forth. The Edwards et al. , study was concerned to
review the literature on such matters and to see what conclusions might be
reached with respect to advice on risk communication.
As was hinted at earlier, systematic reviews normally concentrate on the
'effect sizes' of various interventions. For example, we might want to know
whether a discussion between a doctor and a patient that involved the use of
diagrams was rated as more effective than one that used just words and numbers.

Copyrighted Material
156 USING DOC U M E NTS IN SOCIAL RESEARCH

800

700 -

600 -
W
N
iii 500 -
w
-I
c.. 400 -
:E
<t
en 300 -
>
X
0
a:
200 -
c..

1 00 -

o -

- 1 00 I I
-1 o 2 3 4
EFFECT SIZE

FIGURE 8.2 Funnel plot demonstrating the presence of 'invisible' knowledge: plot
of effect sizes in literature dealing with risk communication (Edwards et al., 2000)

Studies that test and report on such effect sizes are normally published in
academic journals and it is these papers that are usually collected together in
a reVIew.
When effects, such as the above, are studied, they are usually studied for a
population of people. And the key figure that is gathered from such studies is
the average or mean difference in response between those people who expe­
rienced one kind of intervention (say, they were given drug, or intervention, ' A')
from another (who were, perhaps, given drug, or intervention, 'B') . Maybe the
people given intervention 'A' recovered earlier, or lived longer and so forth. A
focus on the average or mean time for recovery or the average number of years
lived after intervention should show up differences in effect. In the risk com­
munication study the 'effects' were often reports of improved understanding or
improved compliance with medical advice.
A funnel plot extracts the information about such 'effects' - from all known
studies - and plots them on a graph in the shape of an inverted funnel. In
Figure 8.2 these effect sizes have been rescaled so that most of the data points
fall between +1 and -1 on the horizontal scale. We have no need to delve into
the manner of the rescaling here. What is important to note is that most of the
data points lay on the positive side of zero, while hardly any data points lay on
the negative side of an imaginary vertical line running through zero. In short,
part of the funnel is missing. It is an interesting shape. And what it suggests is
that 'negative' findings are absent from the published literature.

Copyrighted Material
 DOC U M ENTS AS EVI D E N C E 157

Could that be because researchers only find positive things? Well, perhaps
that is so, but it is unlikely. In fact we know from the manner in which things
are generally distributed in the world that the funnel plot should look sym­
metrical. In other words, in the case of Figure 8.2, there should be far more
negative data points than are available in the graph. Or to put it another way,
the left hand side of the graph should be fuller than it is. Indeed, it seems likely
that what has happened here is that negative and inconclusive research results
have not appeared in published outlets, whilst positive results have.
The reasons as to why some research results are published whilst others are
not raises important questions about pathways to publication and the manipu­
lation of knowledge in general. The factors structuring such pathways remain
open to speculation, but one set of influences that affects publication is that
papers that report on effects that are relatively unequivocal and likely to 'make
a difference' are generally favoured over findings that are ambiguous or contra­
dict prevailing assumptions. The funnel plot provides an image of this tendency
in action.
With qualitative research of course the luxury of numbers is not available,
but the significance of the negative remains important. So it is often argued
that qualitative researchers should always pay special attention to cases that
seemingly fail to fit the pattern that has been observed by the researcher (Seale,
1 999; Silverman, 1 993) . This emphasis on the disconfirming case is sometimes
known as 'deviant case analysis' and it has a special function in qualitative
research. I shall return to deviant case analysis shortly. Before I do that, I need
to turn to broader matters of indexing and coding data.

Data extraction : indexing and coding

Systematic review procedure requires the use of data extraction protocols or

forms. In the case of a systematic review the form would require that details be
collected of where a study is published, authors, sample sizes, 'effect sizes' , con­
clusions and so forth. Naturally in qualitative work one is unlikely to have
access to or to be interested in details of quantitative data sources. Nevertheless,
the notion of having a standardized form that is applied to all 'cases' is a useful
one, otherwise there lTught be a tendency to select only data that fit a precon­
ceived notion or theory and to ignore the negative cases. (For an example of a
standardized data extraction form see Yin, 1 994.)
Extracting data by means of a standardized procedure has its uses. However,
it is often easier said than done. Thus we know, for example, that different
individuals can often see different things in the same data set and read differ­
ent messages from one and the same document. In terms of extraction proce­
dure, therefore, systematic review procedure requires that it should be
performed independently by at least two people. In such cases, disagreements
between data coders are bound to arise and procedures need to be established

Copyrighted Material
158 USING OOC U M ENTS I N SOCIAL R ESEARCH

for dealing with the problem. When it comes to the coding of qualitative data
other kinds of issues can also emerge (see Ryan and Bernard, 2000). In order
to illustrate such problems I am going to refer to some interview data that were
collected on a research project concerned with traumatic brain injury (TBI).
TBI can arise as a result of vehicle accidents or accidents in the home or
violent assault. When it occurs the individuals affected usually experience
impairment of cognitive and physical functions. Memory, speech, bodily move­
ments and social functioning can all be severely affected. Medical professionals
assess the outcome ofTBI according to many dimensions. How injured people
assess such outcomes, however, is not at all clear. In what follows I am going to
refer to some work that I undertook with some colleagues on how injured
people assessed the impact of their injuries on their daily life.
The research team initially decided to gather the data by asking injured
people about (1) how they acquired their injuries and (2) how such injuries
had affected them in the intervening years between the traumatic event and
the interview. Interviews were translated into written transcripts. It was the
transcripts (as documents) that were subsequently analysed.
In a case such as this it is difficult in advance to design a template for data
extraction. This is partly because the injured as individuals talk about different
events and different experiences. At the same time it is easy to see how some­
one seeking to code interview data might be seen to have a free hand about
deciding what was most and least important in the transcripts. So how might
we avoid falling into such a trap? One method of overcoming this problem is
to use a concordance program to index each interview. A concordance
program - as was stated in Chapter 6 - provides a complete list of words used
in an interview together with a count of the number of times that a word is
used. It also provides a context for word use - that is, a sentence or phrase in
which a given word use is embedded.
In the case of interviews it is inevitable that some words will be those used
by an interviewer and some by a subject. So the first task of the researcher is
to exclude interviewer talk from the results. What remains is then the vocabu­
lary of the interviewee. At this early stage, of course, one must always keep
uppermost in mind that words taken out of context can provide a severely mis­
leading picture of what the talk is about. Nevertheless, there is no doubt that
the words that people use - in letters, written confessions, interviews or what­
ever - relate directly to the issues that concern them. In semi-structured inter­
views that are being used as an example here, it will become clear that
interview subjects talk about what is relevant to them in words of their own
choosing. Indexing that talk provides an essential first-base image of what is
and what is not considered relevant in the framework of the interview.
Table 8 . 1 contains data from 1 2 interviewees, all of whom had experienced
head injury as a result of a road traffic accident or a violent attack on them by
others. The 12 interviews being used here were merely the first of 60. The
table indexes a selected series of words that were used in the interview together

Copyrighted Material
 TABLE 8.1 Occurrence of key words in 12 interview transcripts
Words Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8 Case 9 Case 10 Case 11 Case 12
accident 14 3 6 7 o 3 4 4 o 15 2 2
aggression o 1 o o o o o o 1 o 3 6
anxiety 2 o o o o o o o o 1 o 0
before 11 12 6 6 o 6 18 3 15 o 3 16
change 9 4 2 o 1 4 3 2 9 o o 6
concentrate 5 1 2 1 o 6 2 7 1 7 2 4
confidence o o o o o o 7 1 1 1 o o
depressed 2 1 o o o 6 2 1 8 4 2 o o
o
different 3 2 o 3 o 2 7 o 8 5 3 3 (")
disable o o 2 4 1 o o o o o o o c

� forget
friend
7
1
9
7
1
1
3
1
o
1
2
2
o
10
6
o
o
1
2
2
2
2
2
10
s:
rn
z
� o 2 1 4 2 3 o o o o 3 1 -t
�
girl/boys
Ul
head 1 14 4 12 2 17 7 8 7 5 5 7
CD »
0.. headache o 1 o o o 5 1 o o 2 o 3 Ul
injury 1 4 o 2 o 7 5 2 18 1 7 13
� loss 2 1 1 2 1 2 4 4 4 4 o 3
rn
<
CD 3 1 3 2 o 1 o 1 1 o o 2
lucky o
� memory 13 19 11 18 1 11 7 1 10 10 o 9 rn
Z
moods o o o 1 o 1 o o o o o 1 (")
personality o o o o o o o o o 2 o 2 rn

physical o 1 2 o 1 1 4 2 1 1 1 4
since 7 4 o 2 o o 1 5 11 3 o 3
sleep 14 4 o 1 1 4 o 1 o 9 2 11
speech 4 11 1 1 o 3 1 o 4 o 2 o
talk 9 5 2 5 1 14 5 7 4 7 10 4
walk 4 10 5 11 o 2 6 4 1 12 5 3
work 53 25 4 6 o 4 5 11 23 18 17 8
worry 1 o 1 o o 3 o 1 5 5 1 5
....
til
(J)
160 USING DOC U MENTS IN SOCIAL RESEARCH

with a word count. It is evident from the table that 'accident', 'head' and 'injury'
figure prominently in the table. This is of course nothing less than what we
would expect given knowledge of the context of the research interview.
Beyond that, however, we can see that interview subjects also mentioned spe­
cific symptoms of disorder (concentration, depression, memory, sleep and so
forth) . They also failed to mention with any frequency other supposedly ' com­
mon' symptoms of TBI - such as aggression and personality changes. (This
could be because only familiar others would notice these changes and that the
injured themselves are oblivious to such features.) Interestingly, subjects also
referenced the importance of such things as walking and work (as employ­
ment) to their everyday lives.
None of this, however, can tell us how these words are actually tied together
and woven within the narrative of the interviews. More importantly, this kind
of analysis can only produce an analysis of words used. It cannot throw light
on the underlying concepts that relate to the use of such words. Let us con­
sider some examples.
TBI involves traumatic life events. A life as lived before inj ury cannot be
regained and this loss is often central to those who experience TBl. The loss
involves loss of physical control, loss of many ordinary everyday experiences
such as the ability to follow a television narrative for more than a few seconds,
and loss or impairment of cognitive function (such as memory) . Again, as one
might deduce from elements in Table 8. 1 , many of these issues emerge through
the index. For example, 'loss' is specifically mentioned. Of all the elements of
life that are lost, however, one of the most important is a set of social relation­
ships. In particular injured people can become dependent on others. Dominant
husbands might become as childlike dependants. Adult children might return
home in a state of dependence very much as was experienced several decades
previously. In fact, intimate family relationships of dependence can often be
quickly reversed. Equally, friendships that existed before injury rarely carry on
after injury. Girlfriends and old friends disappear. The injured become relatively
isolated and lose social contacts beyond their immediate carers.
N ow, we might refer to these kinds of issues as issues of ' dependency', or
'social disruption', or 'bereavement', or we might use some other concept. The
point is that such terms will not appear in the transcript. And it is at this point
that we begin to move over from indexing to coding. For what we do when
we code data is to arrange and organize the data according to social scientific
perspectives and interests. Coding terms are terms that have to be read into the
interview by the researcher. And as we move from indexing to coding we
begin as researchers to add things to the data. We interpret - but hopefully we
do so on good grounds.
The upshot of this discussion, then, is that coding data begins to move us
beyond the language and terminology of the 'text' and into a realm of social
scientific analysis. This, of course, is j ust what is needed, but it moves the
researcher into the realm of speculation. As a result it is all the more important

Copyrighted Material
 DOC U M ENTS AS EV I D E N C E .16.1

that cocling is undertaken in a rigorous manner. And in the light of our previous
discussion it is also essential that special attention is given to negative or deviant
cases. Let us consider one or two examples.
The research interviews referred to above were structured as narratives.
Narratives are a particularly useful form of cliscourse to social researchers (see
Mattingly and Garro, 2000; Plummer, 200 1 ) , and their analysis never fails to
yield insight into the manner in which people organize accounts of their lives.
In the case of the head injury interviews the dominant narrative seemed to be
one of abrupt change and a certain kind of yearning for the old way oflife. On
some occasions the narrative of loss was emphasized repeatedly. This is for
example partly evident in Case 9 (Table 8 . 1 ) , whose reference to 'before' to
'change' and 'loss' and 'work' added up to a doleful narrative of a fundamen­
tally altered world - and this was so for both the subject and his carer. Yet not
all respondents apparently emphasized loss and change.
For example, in Table 8 . 1 , we see few references from Case 5. However, this
is where contextual information comes to the fore. This person had lost the
facility for speaking. He communicated via a voice box linked to a computer.
Not surprisingly, he reported getting easily tired, and the interview was abbre­
viated in its questioning and in its answering. In that light one might consider
the words used by Case 5 as being of particular value. It is of special note there­
fore that of the few words that are used those relating to loss and change are
present. Case 1 1 , on the other hand, makes no reference to change and loss.
Does this hint at a narrative that was different from others? Checking the full
transcript reveals that Case 1 1 had structured almost the entire interview in
terms of change and difference, but he spoke of what he 'used to' do, say, act
like, take an interest in and so on. In short he merely used a slightly clifferent
lexicon from most other people. The basis of the narrative was not in its
fundamentals clifferent from those of other respondents, and it seems justifiable,
therefore, to argue that the narrative produced by Case 1 1 also expressed a
sense of irreversible change. The larger point, of course, is that by examining
apparently deviant cases such as the above, generalizations or hypotheses about
expected outcomes can be tested robustly and rigorously - and are therefore
more likely to be accepted as 'evidence'. As such, our comments here have impli­
cations for the issues raised in the previous section on the robustness of data.

Conclusions: conjectural history

Conjectural history - as with conjectural social science generally - was, and is,
capable of producing any number of narratives of human development, though
all such narratives are derived from 'theoretical' rather than empirical consid­
erations. Undoubtedly the most powerful narrative frame to emerge during the
infancy of the social sciences was that of evolutionary theory. During the nine­
teenth and early twentieth centuries, evolutionary thought pervaded not only

Copyrighted Material
162 USING DOC U M E N TS IN SOCIAL RESEARCH

biology, but archaeology, anthropology, psychology, sociology and much else

besides (Burrow, 1 966). It is not, of course, our fate to investigate the emer­
gence and development of such ideas, but it will prove useful for us to focus,
ever so briefly, on one particular thesis that was associated with evolutionary
social theory. It was a thesis nesded in the study of human origins, and particu­
larly in the work of speculative archaeology. Put simply, it was a thesis to the
effect that human civilization originated in one particular region (usually
located as the Mesopotamian plains) and then spread outward to the remain­
der of the world. During the earlier part of the twentieth century, this thesis
concerning the diffusion of civilization served as a kind of spine on to which
many detailed archaeological observations were hooked. Indeed, grand and not
so grand hypotheses were attached to the diffusionist skeleton. Thus, Gordon
Childe ( 1 936) , for example, developed his synthesis of early pre-history around
a diffusionist thesis that was embellished with ideas drawn from a Marxist mate­
rialist history. Childe's theses, as well as those of many others, could, however,
only flourish in a world where scholars could be cavalier with chronology - in
particular, in a world where the dating of objects was a somewhat speculative
affair. In short, in a world where, as Childe ( 1 936) put it, 'Dates in years before
3000 B.C. are just guesses.'
Some 1 0 years after Childe published his book, carbon-1 4 (C 1 4) dating
techniques were adopted in his science. Such techniques enabled, for the first
time, reasonably accurate dates to be established for the creation of almost any
human artefact - pottery, clothing, seed remnants, bones or whatever. And
once it became possible to establish a detailed chronology of activity, the evidence
for the diffusionist thesis mentioned above appeared increasingly fragile.
Indeed, litde by litde it became clear that the chronology of early human activity
was more consistent with a thesis that argued for a multi-centred origin to
human productive processes rather than a single, central origin. What is more,
the time span of human history was gready expanded. Indeed, the rise of C 1 4
dating techniques implied a consequent collapse of the rather simple diffu­
sionist theory oudined herein, and it held profound consequences for archae­
ologists generally. Their adoption certainly had a deep personal effect on
Childe, who was subsequendy reported to have considered much, if not most,
of his life's work to have been in error.
Whatever his feelings, Childe at least recognized that theoretical speculation
must always be restrained by and checked against what counts as 'good'
evidence in a professional domain. Thus, during the 1 940s and 1 950s the
physics of radioactive decay (C 1 4 dating) had to be taken into account, and
generalizations and hypotheses about chronology had to fit with findings
derived from the new technology. Childe very properly recognized this, as well
as other rules and conventions relating to evidence and proof.
Rules and conventions about good and bad practice clearly differ in one
realm to another. Thus 'evidence' as presented in courts of law clearly differs
from what is taken to be evidence in ordinary, everyday discourse. Proof in

Copyrighted Material
 DOCU M ENTS AS EVIDENCE 163

mathematics is a radically different thing from proof in forensic science. Robust

conclusions reached via quantitative investigation depend on quite different
judgements from those used in qualitative enquiry. In all cases, however, there
should be a transparency about what is to count as valid and reliable evidence.
That is to say, a clarity about what was done in the research process, and how
conclusions were arrived at. For, one can always argue about the validity of the
rules of the game, but if the rules remain opaque then it is very difficult to dis­
tinguish between the conscientious and rigorous player and the cheat. In this
chapter I hope that I have indicated how one can play to reliable standards.
In summary, the following key points emerge:

• Researching the inert text requires one to attend to issues of reliability and
validity.
• Issues of reliability and validity in turn require that we state at the outset of the
research project what, exactly, we are seeking to achieve, and what is to be
included in the field of study.
• Selection (and exclusion) of documentary materials should be in accordance
with the principles established in the preceding point.
• In those instances where documentary materials have to be sampled, a thor­
ough justification for the sampling procedures needs to be provided.
• Indexing and coding of data need to be executed in a rigorous and unbiased
manner.
• Whilst drawing conclusions from data, always pay special attention to data that
apparently fail to confirm one's claims and generalizations.

RESEARCH EXERCISES

Exercise 8.1

Some of the most puzzling events of the last quarter of the twentieth
century concern the appearance and spread of a new kind of disease
in cattle, humans and other mammals such as domestic cats. The
disease has various names - for example, it is known as both ' mad
cow disease' and variant CJD. The UK government instigated a full
investigation into the origins and consequences of the disease. The
report is available in ful l at (http:#www.bse.org.uk/index.htm) Included
in the report are volumes of 'evidence' .

1. Read the ' Executive Summary' of the report to get a feel for the
issues and problems that were examined by the investigative
committee.
2. Draw up a list of things that the i nquirers counted as 'evidence' .
(Make sure to include the ' Materials' volume in your overview.)

Copyrighted Material
164 USING DOCU MENTS IN SOCIAL R ESEARCH

3. Draw some conclusions about the way in which the written

evidence that was offered to the committee members differed from
oral evidence in its variety, structure and function.
4. Scrutinize Volume 2 of the report entitled 'Science' and note how
'evidence' is used in that volume. Pay particular attention to the
relationships between hypotheses and conclusions, and the
grounds on which hypotheses are rejected.
5. By using Volume 2 or any other volume of the report draw some
conclusions about the way in which text and 'evidence' are related
in the report as a whole.
6. Finally, offer some suggestions as to how the report functions in
relation to its terms of reference. (The latter are listed in the
' I ntroduction' to the report.)

Notes

The title of a book by Karl Popper (1 963), a book in which he laid down princi­
ples for the acceptance and rejection of scientific assertions.
2 These issues were taken up in a BBe Horizon programme (go to http://www.bbc.
co.uk/science/horizon, and search for 'Atlantis).
3 One reason for failure to accept disconfirm.ing experimental results is that it is quite
normal for different runs of the 'same' experiment to produce different and diver­
gent results. Systematic review methodology overcomes this problem by accepting
all results and analysing the bigger picture that is produced by the sum total.
4 See http: //www.york.ac.uklinst/crd/report4.htm and http://www.cebm.net/
5 http://www.nlm.nih.gov/
6 http://www.csa.com/detailsV3/socioabs.html
7 http: //www.cas.org/ONLINE/DBSS/sigless.html
8 In this respect the case of the so-called Hitler Diaries is rather interesting, see Harris
(1 996).
9 A useful example for study concerns the 'missing link' forgery - known as the
forgery of the Piltdown man. Extensive documentation of the case and surround­
ing issues can be found at http://www.clarku.edu/-piltdown/pp_map.html

Copyrighted Material
9
Producti on , Con s u m ption and Exch a nge

Lessons from southern Sudan 165

Circuits of production and consumption 166
The sociology of the Kula 169
Documents as technology 171
A final plea 172
Notes 173

Lessons from southern Sudan

Documents? There never were any. There are no written orders, no ordnance
maps, cryptographs, leaflets, proclamations, newspapers, letters. The custom of
writing memoirs and diaries does not exist (most frequently there is simply
no paper) . There is no tradition of writing histories. Most importantly - who
would do this? (KapuScinski, 200 1 : 4)

Kapuscinski is speaking of life in southern Sudan in the last decade of the

twentieth century; of life in a war zone; of life among peoples such as the
Dinka and the Nuer, peoples who had been closely studied during the middle
part of the last century by anthropologists such as Evans-Pritchard (1 940) and
Lienhardt ( 1 96 1 ) . He is also speaking of a world that is lost - precisely because
there is no documentation to hold it in place. Without documents there are no
traces. Things remain invisible and events remain unrecorded. The only
resource is word of mouth accounts. It was in such worlds that anthropology
was developed. So it is, perhaps, understandable that the research act of the
anthropologist turned on matters of observation, and the recording of speech
and behaviours. How could one use documents in the absence of paper?
Naturally, in the modern affluent world - the world in which we live - docu­
mentation is central. The written order forms a cornerstone of modern life, a
point emphasized by Max Weber in his account of that quintessential organi­
zational form of the twentieth century - bureaucracy. Yet, in modern social
science, the written trace continues to be neglected. There are few studies con­
cerned directly, and not just tangentially, with the use of documentation in
contemporary life and even fewer texts devoted to the problem of research on

Copyrighted Material
166 USING DOCU MENTS I N SOCIAL RESEARCH

forms of documentation. For analysis of speech, on the other hand, the range
of materials is expansive. This is so even though the analysis of speech has,
necessarily, to be mediated through text; that is, via a transcription (see, for
example, Silverman, 1 996: 254) . I n that light it would be an interesting research
exercise in itself to look at the emergence of the various social scientific tools
and techniques that have been used to translate talk into text.
There is, of course, no obvious way to account for the differing fortunes of
speech and writing in social scientific research. Perhaps the neglect of the role of
writing in contemporary life serves as something of a pointer towards a funda­
mental blind spot in western culture. Such an undervaluing of the written, as
opposed to the spoken word, has of course been remarked upon elsewhere.
Thus, the anthropologist Jack Goody ( 1 968; 1 977) has frequently referred to
writing as a rich, yet neglected, field for research studies. More fundamentally,
the philosopher Jacques Derrida (see Howells, 1 998) has underlined, on a
number of occasions, how the written word has always been considered sub­
sidiary and secondary to the spoken word in the philosophical corpus of the
west. Yet the subordinate role of writing to speech is far from deserved, and I
hope that I have demonstrated just how that is so throughout this book.
One of the key claims embedded in previous chapters is that documents, and
especially written documents, can be taken as a field of research in their own
right. In particular, the study of the processes of production and consumption
(or use) of written materials provides two sturdy pillars around which interest­
ing and essential research programmes can be built and developed. Naturally,
in the hurly-burly of ordinary everyday activity issues of production and con­
sumption become entwined, and it is not always easy, as we have seen, to dis­
tinguish clearly between the one process and the other. To end our analysis
therefore let us consider some final examples. This, not merely so as to under­
line the central issues, but also in order to push our analysis in one last - and
rather important - direction.

Circuits of production and consumption

Consider a map - a road map, perhaps. We commonly think of maps and charts
as providing true and accurate reflections of the world out there. If they did
not, then how else would we find our way across the countryside? Yet clearly,
a map can never depict with precision the terrain it claims to represent other­
wise it would be useless. It is an issue well taken up, for example, in a short
story by Jorge Luis Borges entitled 'Of exactitude in science' (Borges, 1 975) .
The story concerns a map of an entire empire 'that coincided with it point for
point', but being too cumbersome to use, it was abandoned - tattered frag­
ments of the map later being found in the western deserts where it was used
to shelter 'an occasional beast or beggar' . Not surprising you may think, for a
useful map obviously has to involve abbreviation and abstraction - as well as

Copyrighted Material
 P R O DU CTI O N , C O NS U M PTI O N AND EXC H A NG E 167

scaling. A map, moreover, must contain identifiable items and patterns - features
that we can recognize and group together.
Abbreviation, abstraction, scaling, grouping, pattern identification; these are
essential human activities, and they are necessarily implicated in the manufac­
ture of maps and numerous other documents. In Chapters 2 and 4, in particu­
lar, we examined such processes in detail and in a variety of contexts. We also
noted in Chapter 2 how documents are not simply produced and manufac­
tured and consumed, but often used to produce new facts and new things. In
that respect it was suggested that documents can often become enmeshed in a
circuit of production. And there are many documents such as the DSM and the
ICD (see Chapter 2) that may be looked upon in that light as machine tools.
That is, tools used to produce other manufactured goods. Such documents are,
by their very nature, implicated in endless circuits of production.
However, documents are not simply an adjunct to fact-producing activities -
mere tools to be used. Rather, fact production lies at the heart of their very
being. Indeed, most of the things that we commonly consider to be 'facts of
nature', or facts about society, can be said to exist only in documents. This,
despite claims to the effect that the materials contained within this or that
document merely reflect or represent some external reality beyond the. document.
As has been pointed out in earlier chapters, the social science researcher should
therefore never be tempted to think of the descriptions, images and explana­
tions that are contained in documents as mere representations of the world as
it 'really is' . Instead the researcher should consider the possibility that the world
is actually constituted in and through documentation. The materials referred to
in Chapters 3 and 4 provided some first-class examples of how this can be so,
and here I offer a further example.
Norton 's Star A tlas (Ridpath, 1 989) is commonly referred to as the amateur
astronomer's 'bible' . It is a book of star maps. As with all maps, we find within
evidence of abbreviation, scaling, abstraction and fact production in abundance.
Evidently, then, the starry firmament as presented in a star map is a humanly
created phenomenon. Naturally, the matter of which stars and nebulae are
composed exists independently of human intervention and creation, but the
moment we categorize the stars into constellations and galaxies we divorce
ourselves from a self-sustaining and independent world and begin to engage
with a universe constructed by feeble humanity. In fact the map (document)
mediates the relationship between human beings and the universe. Thus, it is,
after all, only earthpersons who 'see' Orion and trace its 'movements' through
the northern sky in winter, and it is only earthpersons who view the blue­
white stars that make up Cassiopeia's 'W' as somehow belonging together.
Indeed, neither Orion, nor Cassiopeia, nor Pisces, nor any other feature of the
constellations appears outside of human documents, and without the texts they
would fail to exist at all.
Stars and star maps are a long way from social science perhaps. Yet, as with
the natural world, many features of the social world are also created with the

Copyrighted Material
168 USING DOC U M E NTS IN SOCIAL RESEARCH

aid of documents. Thus, we saw in Chapter 1 how the anthropological documents

of the Torres Straits anthropologists served to create kinship. And we noted in
Chapter 4 how a family tree (pedigree) as drawn up in a genetics clinic can act
so as to manufacture both families and genetic disorders. Indeed, borrowing
our terminology from the sociology of science, we described the machine that
draws human pedigrees as an 'inscription device' (Latour and Woolgar, 1 979:
5 1 ) . In the work of the latter, such a device is an apparatus that turns matter
into documentation - as, say, with the graphical trace of temperature changes
in a living body. In this book we have broadened the meaning to refer to any
device - a questionnaire, a statistical device, a table or chart - that is capable of
translating social concepts (such as the family, or social class) into visible facts.
Documents, then, represent and make things visible. They also construct.
They construct and stabilize objects and things (Chapters 2, 4 and 7), identities
(Chapter 5) and processes (Chapter 3) . Yet in researching documentation we
must forever remain aware that representation, construction, stabilization are
carried out in concert - in concert with other agents. So it is the production
and consumption of documents in their social settings that are important - how
the document fits into the entire network of activities and agents of which it
forms a part. That is the key to the research process.
In this light it is instructive to consider a further essay on maps, one written
by George Psathas ( 1 979) some years ago. In that essay, Psathas looked at how
maps were used in context. His specific focus was on the kind of maps that
people draw and dispense for and to others so as to find the forthcoming party
at 'our house' or some such. His sociological interest was on the reasoning that
was implicated in the drawing of such maps. For example, he pointed out how
direction maps are always drawn with reference to a destination rather than,
say, to the topography of a given neighbourhood. Indeed, the use of such maps
clearly implicates readers as well as writers (or in this case amateur cartogra­
phers) . For readers of such maps are invariably inveigled into following the
sequences drawn on the map. They are obliged, as it were, to perform the route.
Thus, in reading and using the map, the map-reader moves him- or herself
from point A to point B in a manner dictated above all by the map-maker.
Such use provides a good example of a process referred to as action-at-a-dis­
tance. It also serves to demonstrate how documents in use can structure and
pattern their readers (a process expanded upon in Chapter 5) . This is precisely
why I have argued that documents can often be considered as agents in, rather
than mere adjuncts to, networks of action. And, as was stated in the introduc­
tory chapter, documents in action often take on qualities of the broom set in
motion by Goethe's Sorcerer's Apprentice, or the monster unleashed on the
world by Mary Shelley's Frankenstein. That is, qualities that are independent of
the creator. Perhaps it is not so surprising, then, as Wheeler (1 969) indicated,
that records and dossiers necessarily have a career independent of the human
subjects that they supposedly report upon. But rather than refer back to issues
that have already been dealt with, it will prove more fruitful for us to consider

Copyrighted Material
 PRODueTION, C O N S U M PTI O N A N D EXC HANG E 169

one final feature of the map-making example. Maps are constructed, and maps
are used or consumed in practical activities. I Maps are also exchanged. Yet, of
this latter process we have so far said little.

The sociology of the Kula

In modern cultural studies there exists a concept of the cultural circuit (du Gay
et al. , 1 997) . The circuit supposedly incorporates 'moments' of production and
consumption as well as of representation, regulation and identity. The image of
a circuit is, in part, chosen to illustrate the dynamic and interlinked nature of
the ways in which cultural products are appropriated and produced in the
modern world.2 Documents can be considered likewise. Yet, somewhat oddly,
there is one important theme that is entirely absent from the ' circuit' that is dis­
cussed and analysed by the cultural studies cohort. It is the theme of circula­
tion itself.
The circulation and exchange of goods and services has often formed a
major topic of investigation by anthropologists. It was in that context, for
example, that Malinowski, in his Argonauts of the Western Pacific, examined the
sociology of the Kula. As Malinowski pointed out, 'The Kula is a form of
exchange' ( 1 922: 8 1 ) . It was a form of exchange carried on by peoples among
a ring of islands in Melanesia. The islands may be imagined as belonging to a
circle. Malinowski noted that in the clockwise direction of the islands flowed
long necklaces of red shell (called soulva) , whilst in the anti-clockwise direction
flowed bracelets of white shell (called mwalt) . Each type of article was con­
stantly being exchanged for the other. (The necklaces and bracelets were never
kept as private property.) Further, the act of exchange was normally accompa­
nied by extensive ritual, and associated with the Kula were many other activi­
ties - such as conm1erce and boat building.
In writing about the Kula, Malinowski was developing insights about social
life in general. Thus, he noted that the Kula system was not primary to eco­
nomic life, but rather to social life. In so doing, he was emphasizing the fact
that the exchange process cemented social partnerships, rather than economic
relations. His insights were built upon throughout the twentieth century, and
were later developed into what has been called social exchange theory (Ekeh,
1 974). Our interest in Malinowski's analysis is that we can consider documents
and their exchange and circulation in much the san1e way as Malinowski con­
sidered necklaces and shells. For, documents can also be used to trace out
patterns of social exchange and thereby the social networks that lay behind them.
For example, in the modern world the exchange of greeting cards is a
conill1onplace event. Birthdays, weddings and even deaths form occasions where
people send cards to oneanother. People also buy and send cards to those who
are ill - 'Get well soon cards'. But to whom, exactly, do people send such cards?
And what do the patterns tell us about social life? In a rather fascinating study

Copyrighted Material
 �
HO USING DOCU MENTS IN SOCIAL R ES EARCH

58 -- 59 --- 7' l 32

I 56
42

'[7" / \ 35
70

�
-----/ \
12

36

39

J\o \
3

73 10
/' 49 4 63 - 65

FIGURE 9 . 1 Information contacts among professionals in South Wales, 1999

Weiner et al. (1999) looked at the distribution of such cards among hospital
patients suffering from two different types of disorder. The first were medical
disorders (for example, heart diseases and diabetes) and the second were serious
psychiatric disorders. What they noted was that although both sets of patients
were hospitalized (for more than three days) , the medical patients were far
more likely to get visits, gifts and 'Get well soon cards' than were the psychia­
tric patients. Such a pattern of gift exchange, suggested the authors, tells us
something fundamental about how we view psychiatric illness and psychiatric
patients in our society. (Psychiatric patients are commonly regarded as not
being 'really' ill and even when they are so regarded it is often thought to be
their own fault - so sympathy is in short supply.) Although it was not part of
the planned investigation, the hospital researchers could have gone further, and
used the cards to do something more. They could have used the exchange of
cards to trace patterns of a social network.
The exchange of communications - such as cards - between individuals not
only facilitates the flow of information, but also marks out channels and bound­
aries of social influence. For example, Figure 9 . 1 is a network graph. It represents
a trace of information exchange events among a small group of primary-care
practitioners in Wales. The lines represent connections between people who seek

Copyrighted Material
 PRODueTION, CON S U MPTI O N A N D EXC H A NG E 171

'advice and information' from one another. Thus, by viewing the graph it is
possible to see clearly the existence of groups of people who seemingly consult
each other a lot, possibly because they work closely together - in the same build­
ing. For instance, individuals referred to as numbers 25, 26, 27, 28 and 29 evidently
belong together, as do the GPs numbered 1 7-22. It is also possible to see that
some people are contacted considerably (numbers 29 and 34, for example) whilst
others look relatively isolated (numbers 6 1 , 32 and 42).
Such a pattern of contact and communication is traceable in other ways -
say by asking people about their e-mail contacts.3 And what the results of such
questioning show is not simply evidence about who contacts whom, but also
evidence of social power and social influence. In the study concerning the GP
contacts, discussed here, it was possible to link the patterns in the graph to
other information about the GPs, so as .to understand why the patterns were
as they were. Professional factors, ethnic factors and gender factors all seem­
ingly structured the patterns of contact and communication. And although the
graph in Figure 9 . 1 represents an image of verbal communication it could, in
theory, be extended to represent an image of communication via the use of
written documents.
In his conversations with Latour, Serres ( 1 995: 1 6 1 ) refers to objects that
trace or 'make visible the relations that constitute the group through which
[the object] passes, like the token in a children's game' . The term that Serres
uses to refer to such objects is 'quasi-object' - a term chosen to emphasize the
fact that the world cannot be simply divided into objects and humans, for
objects and people belong in an ensemble that is often difficult to unravel. For
example, in the GP network just referred to one of the questions being stud­
ied related to the dissemination of innovations, in particular the dissemination
of knowledge about new drugs (pharmaceuticals) such as Viagra. At a distance
it might seem that it is easy to distinguish between such an object and the
human influences upon it.Yet, in practice it is not. For what the drug (Viagra)
is, is fashioned by what is said about it, by patterns of use, and by patterns of
reception. The technology that is represented by the drug is thereby wrapped
up in a whole series of relationships and networks. To understand the technol­
ogy we need to follow the thing (document) through its social channels. And
all the while the circulation of communications about the document will serve
to mark out the boundaries of social groupings and social relationships - as in
Figure 9. 1 .

Documents as technology

Goody ( 1 977) and Ong (1 982) point out that writing is a form of technology,
and a very important one at that. They argue, for example, that the move from
oral to written culture changed the technological potential of human society.
Thoughts once written were able to be re-analysed and re-examined. Derrida

Copyrighted Material
172 USING DOC U M E NTS IN SOCIAL RESEARCH

refers to this property of writing as its iterability (see Howells, 1998). Iterability
encourages the development of critical and scientific thought. Writing also
bolsters the development of rational systems of accounting and monitoring,
and encourages close scrutiny of the ways in which we list and classifY things.
Looking at documents as technology encourages us to think about how the
technology is linked into productive relationships. For example, Goody ( 1 968)
provides insight into how, in Imperial China and other traditional societies, the
complexities of (Chinese) script were consciously and deliberately manipulated
by the literati so as to underpin their monopoly on state activity. The develop­
ment and widespread use of a phonetic alphabet (as used in the world of Latin
Christianity) would clearly have undermined the exclusivity of such power
groups. Not surprisingly the adoption of phonetic forms of writing were
always opposed within traditional China.
By understanding how technology is used, who recruits it and allies them­
selves with it, how it is adopted, and adapted, and how it circulates, can form a
series of key entry points into the investigation of social life. This is especially
so when alternative technologies are available in the same time and place. For
then it becomes possible to see how the supporters of the alternatives line up
one against the other, and how the different forms of technology change social
relations in which they are enmeshed. Though we must be forever mindful that
technology is not merely the wires and widgets within the 'machine box' - see
Mackay (1 997) . It is always hardware plus social relationships that count, and
not simply hardware alone. It is a lesson that is well underlined in the work of
Bijker et al. ( 1 989) who provide numerous examples of technology in action.
In the case of documentation, of course, wires and widgets are usually of
marginal significance, though instrumentation is not. Thus, we have seen in
Chapter 4, for example, how measuring instruments of various kinds consti­
tute a form of technology. Indeed, social researchers usually refer to their ques­
tionnaires and the like simply as 'instruments' . Quality of life instruments,
health outcome instruments, survey instruments and the like constitute a par­
ticularly interesting genre of modern technology. Systematic review techniques
(as referred to in Chapter 8) may be said to form another. In all cases, the tech­
nology is linked to particular kinds of professional practice and political and
managerial manoeuvrings. And it is such relationships between things and their
contexts that determine the shapes and appearances of the elements. Conse­
quently, what is needed for work with documentation is a focus on relations
rather than on things in isolation.

A final plea

Production, consumption and exchange are three processes that are central to
the study of documentation in use. It might be objected of course that the
most important feature of documentation - content - has been given too low

Copyrighted Material
 PRODueTION, CONSUMPTION AND EXC HAN G E 173

a profile throughout the entire book. I might defend such a lack of focus by
saying that there are numerous other texts that deal with aspects of content
analysis, or discourse analysis, and that there would be little point in duplicat­
ing such work here. However, I offer no such defence. There is no need. Instead
I offer an analogy. My analogy concerns an operatic libretto (the set of words
and phrases that is sung) . Taken on its own a libretto rarely adds up to much.
The text as narrative is often disjointed, repetitive and lacking in depth. I can­
not think of a single one that might hold a person's attention as a gripping tale.
Yet, a libretto is not intended to be analysed in isolation. It demands to be
analysed in action. How it is integrated into the dramatic action on stage, how
it relates to the melody and rhythm of the music, how it is called upon
(recruited) and manipulated by the singers, how it is peiforrned all of these are
-

of primary importance. Its substance as displayed on the inert page is of only

secondary concern.

RESEARCH EXERCISE

Exercise 9.1

Consider identifying a small sample of people (say, N 5) who use text

=

messaging services on mobile phones. Question them about how they

use such messages. Pay particular attention to the fol lowing issues:
(1) How the messages circulate and the extent to which they function
to mark out the boundaries of a social clique. (2) How the text
messages are integrated into the social activities of senders and
receivers. (3) How the messaging technology functions in the everyday
l ife of the user ( Mackay's ( 1997) work might be useful here). (4) How
the shape and form of the text can itself serve as an agent in the
structuring of social interaction .
Then consider expanding the sample so as to gain coverage of
additional kinds of user or to explore elementary hypotheses that may
have emerged from working with the initial sample.

Notes

1 For another example of maps in use see Brody (1981 ) .

2 In this vein, Fiske ( 1 989) has illustrated how acts of consumption can, in turn, be
viewed as 'work' or production.
3 Citation searches can also be used to map out a network of researchers or an emer­
gent network or boundaries of research networks. See Webster ( 1 991 ) .

Copyrighted Material
B i b l i ogra phy

Abelen, E., Redeker, G. and Thompson, S.A. ( 1 993) 'The rhetorical structure of
US-American and Dutch fund-raising letters' Text, 1 3:3: 323-50.
ABPI ( 1 996) Compendium ojpatient information leaflets, 1 996-97. London: Datapharm.
American Psychiatric Association (2000) The Diagnostic and Statistical Manual oj Mental
Disorders. DSM-IV- TR. Washington, DC: American Psychiatric Association.
Arensberg, CM. (1937) The Irish Countryman. An Anthropological Study. New York:
Macmillan.
Aries, p. (1981) The Hour oj Our Death . Tr. H. Weaver. New York: Knopf.
Aries, p. (1 985) Images of Man and Death. Tr. ]. Lloyd. Cambridge, MA: Harvard
University Press.
Aristotle ( 1 995) Poetics. Tr. S. Halliwell. Cambridge, MA: Harvard University Press.
Ash, T.G. ( 1 997) The File. A Personal History. London: HarperCollins.
Atkinson, ].M. ( 1 978) Discovering Suicide. London: Macmillan.
Atkinson, P. (1 990) The Ethnographic Imagination. Textual Constructions oj Reality.
London: Routledge.
Atkinson, P. ( 1 992a) 'The ethnography of a medical setting. Reading, writing and
rhetoric', Qualitative Health Research, 2 (4) : 45 1-74.
Atkinson, P. ( 1992b) Understanding Ethnographic Texts. London: Sage.
Atkinson, P. and Coffey, A. (1 997) 'Analysing documentary realities', in D. Silverman
(ed.), Qualitative Research. Theory, Method, Practice. London: Sage. pp. 44-62.
Atkinson, R. ( 1 998) The Life Story Interview. London: Sage.
Austin,].L. ( 1 962) How to do Things with Words. Ed.]. o. Urmson and M. Sbisa. Oxford:
Clarendon Press.
Baldi, S. ( 1 998) 'Normative versus social constructivist processes in the allocation of
citations. A network-analytic model'. A merican Sociological Review, 63: 829-46.
Banks , ]. and Prior, L. (200 1) 'Doing things with illness. The micro-politics of the CFS
clinic', Social Science and Medicine, 52 ( 1 ) : 1 1-23.
Barnouw, D. and Van Der Stroom, G. (1 989) The Diary oj Anne Frank. The Critical
Edition. London: Viking.
Barrett, R. ( 1 996) The Psychiatric Team and the Social Diifinition oj Schizophrenia. An
Anthropological Study of Person and ntness. Cambridge: Cambridge University Press.
Barthes, R. ( 1 977) Image-Music- Text. Tr. S. Heath. London: Fontana.
Barthes, R. (1985) The Fashion System .Tr. M.Ward and R. Howard. London:Jonathan Cape.
Barthes, R. ( 1 990) SIZ. Tr. R. Miller. Oxford: Blackwell.
Barley, S.R. and Bechky, B.A. ( 1 994) 'In the background of science. The work of
technicians in science labs'. Work and Occupations, 21 ( 1 ) : 85-126.
Barton, D. and Hall, N. (2000) Letter Writing as Social Practice. Amsterdam: John Benjarnins.
Bastide, F. (1 990) 'The iconography of scientific texts: principles of analysis', in M. Lynch
and S.Woolgar (eds), Representation in Scientific Practice. London: MIT Press. pp. 1 87-230.

Copyrighted Material
 BIBLIOGRAPHY 175

Bauer, M.W and Gaskell, G. (eds) (2000) Qualitative Researching. With Text, Image and
Sound. London: Sage.
Bauman, z. ( 1 991) Intimations if Postmodernity. London: Routledge.
Bauman, Z. ( 1992) Mortality, Immortality and Other Life Strategies. Cambridge: Polity
Press.
Baxter, R. (1 673) A Christian Directory. Ann Arbor, MI: University Microfilms, 1 970.
Bazerman, C. ( 1 988) Shaping Written Knowledge. The Genre and Activity of the
Experimental Article in Science. Madison, WI: University of Wisconsin Press.
Becker, H .S. ( 1 953-4) 'Becoming a marihuana user', American Journal if Sociology,
59: 253-4.
Becker, H.S. ( 1 986) Doing Things Together. Evanston, IL: Northwestern University Press.
Benson, D. and Hughes, ].A. ( 1 983) The Perspective if Ethnomethodology. London:
Longman.
Berg, M. ( 1 996) 'Practices of reading and writing. The constitutive role of the patient
record in medical work', Sociology if Health and fllness, 18 (4): 499-524.
Berg, M. ( 1 997) Rationalizing Medical Work. Decision-support techniques and medical prac­
tices. Cambridge, MA: MIT Press.
Beverley, ]. (2000) 'Testimonio, subalternity, and narrative authority' , in N.K. Denzin
and Y.S. Lincoln (eds), Handbook of Qualitative Research. 2nd edition. Thousand Oaks.
CA: Sage. pp. 555-65.
Bijker, WE., Pinch, T. and Hughes, T. P. (eds) ( 1 987) The Social Construction if
Technological Systems. Cambridge, MA: MIT Press.
Birke, L. and Smith,]. (1 995) 'Animals in experimental reports. The rhetoric of science',
Society and Animals, 3 (1): 23-4 1 .
Bloch, A . and Bloch, C . ( 1995) The Song of Songs. A New Translation. New York:
Random House.
Bloor, M. (1991) 'A minor office. The variable and socially constructed character of
death certification in a Scottish city', Journal if Health and Social Behaviour, 32 (3):
273-87.
Borges, ].L. (1 975) A Universal History of Infamy. Tr. Norman Thomas di Giovanni.
Harmondsworth: Penguin.
Boswell,]. ( 1 976) Boswell's Life ifJohnson. London: Dent.
Bourdieu, P. ( 1 977) Outline of a Theory if Practice. Tr. R. Nice. Cambridge: Cambridge
University Press.
Bowker, G.c. (1 994) Science on the Run: Information Management and Industrial Geophysics
at Schlumberger, 1 920-40. Cambridge, MA: MIT Press.
Bowker, G.c. and Star, S.L. (1 999) Sorting Things Out. Classification and Its Consequences.
Cambridge, MA: MIT Press.
Bowling, A. (1 997) Measuring Health. A Review if Quality of Life Measurement Scales. 2nd
edition. Buckingham: Open University Press.
Breuer, ]. and Freud, S. ( 1 974) Studies on Hysteria. Pelican Freud Library Vol. 3.
Tr. ]. Strachey and A. Strachey. Harmondsworth: Penguin.
Brodbeck, M. (ed.) ( 1 968) Readings in the Philosophy of the Social Sciences. New York:
Macmillan.
Brody, H. ( 1 974) Inishkillane. Change and Decline in the West if Ireland. New York:
Schocken Books.
Brody, H. (1981) Maps and Dreams. London: Jill Norman and Hobhouse.

Copyrighted Material
1.76 BIBLIOGRAPHY

Brown, G. (1981) 'Etiological studies and the definition of a case', in ].K. Wing,
P. Bebbington and L.N. Robins (eds) , What is a case? The problem of diftnition in
psychiatric community surveys. London: Grant Mcintyre. pp. 62-9.
Bruner, ]. (1 987) 'Life as narrative', Social Research, 54 (1): 1 1-32.
Bruner, ]. ( 1 993) 'The autobiographical process', in R. Folkenflik (ed.), The Culture of
Autobiography: Constructions of Self Representation. Stanford, CA: Stanford University
Press. pp. 38-56.
Bulmer, M. (1980) 'Why don't sociologists make more use of official statistics?',
Sociology, 14 (4) : 505-23.
Burgess, R.G. (ed.) (1 982) Field Research: A Sourcebook and Field Manual. London: Allen
and Unwin.
Burgess, R.G. (1984) In the Field. An Introduction to Field Research. London: Allen and
Unwin.
Burgher, M.S. (1 997) A tlas of Mortality in Europe: Subnational Patterns, 1 9801 1981 and
1 9901 1991. WHO European Centre for Environment and Health. WHO Regional
publications. European series no. 75. Copenhagen: WHO Regional Office for
Europe.
Burgos-Debray, E. (ed.) (1 984) I, Rigoberta MenchU. An Indian Woman in Guatemala.
New York: Verso.
Burrow, ].W (1 966) Evolution and Society. Cambridge: Cambridge University Press.
Burton, F. and Carlen, P. (1 979) Official Discourse. On Discotlrse Analysis, Government
Publications, Ideology and the State. London: Routledge.
Callon, M. (1 989) 'Society in the making. The study of technology as a tool for socio­
logical analysis', in W. E. Bijker, T Pinch and TP. Hughes (eds) , The Social Construction
of Technological Systems. Cambridge, MA: MIT Press. pp. 83-1 03.
Casagrande,].B. (ed.) (1960) In the Company of Man. New York: Harper Row.
Castells, M . (1 997) The Power of Identity. London: Blackwell.
Chapman, C.J. (1 997) Cyrillic for Pedigree Drawing. Oxford: Cherwell Scientific
Publishing.
Charlesworth, M., Farrall, L., Stokes, T and Turnbull, D. (1 989) Life Among the Scientists.
An Anthropological Study of an Australian Scientific Community. Melbourne: Oxford
University Press.
Charmaz, K. (2000) 'Grounded theory: obj ectivist and constructivist methods' , in
N.K. Denzin and YS. Lincoln (eds) , Handbook of Qualitative Research. 2nd edition.
Thousand Oaks, CA: Sage. pp. 509-36.
Chartier, R., Boureau, A. and Dauphin, C. (1 997) Correspondence. Models of Letter Writing
from the Middle Ages to the Nineteenth Century. Tr. C. Woodall. Princeton, NJ:
Princeton University Press.
Chauvel, L. (1 997) 'L'uniforrnisation du taux de suicide masculin selon !'age: effet de
generation ou recomposirion du cycle de vie?', Revuefranfaise de sociologie, XXXV III
(4): 681-734.
Cherry-Garrard, A. (1 994) The WorstJourney in the World. London: Picador. (1st edition,
1 922).
Childe,VG. (1 936) Man Makes Himself. London: Watts.
Cicourel, A.V (1 964) Method and Measurement in Sociology. New York: The Free
Press.
Cicourel, A.V (1 976) The Social Organisation ofJuvenile Justice. London: Heinemann.

Copyrighted Material
 BIBLIOGRAPHY 177

Claus, E.B., Risch, N. and Thompson, WD. ( 1 99 1 ) 'Genetic analysis of breast cancer in
the cancer and steroid hormone study', American Journal of Human Genetics, 48:
232-42.
Clifford, j. and Marcus, G.E. (eds) ( 1 986) Writing Culture: The Poetics and Politics of
Ethnography. Berkeley, CA: University of California Press.
Cobb, R. ( 1 978) Death in Paris. 1 795- 1 80 1. Oxford: Oxford University Press.
Coleman, C. and Moynihan, j. (1 996) Understanding Crime Data. Haunted by the Dark
Figure. Buckingham: Open University Press.
Collins, H .M. and Pinch, T. (1 993) The Golem. Cambridge: Cambridge University Press.
Crapanzano, V. (1 980) Tuhami: Portrait of a Moroccan. Chicago: University of Chicago
Press.
Daniell, D. ( 1 989) Tyndale's New Testament. New Haven, CT: Yale University Press.
Darnton, R. ( 1 984) The Great Cat Massacre and Other Episodes in French Cultural History.
New York: Basic Books.
Darnton, R. (2000) 'An early information society. News and the media in eighteenth­
century Paris', American Historical Review, 1 05 ( 1 ) : 1-35.
de Certeau, M . ( 1 984) The Practice of Everyday Life. Tr. S. Rendall. London: University
of California Press.
de Duve, T. (1 993) The Definitively Unfinished Marcel Duchamp. Cambridge, MA: MIT
Press.
Defoe, D. (1 956) The Adventures of Robinson Crusoe. London: Shorter Classics.
Denzin, N.K. (1 970) The Research Act in Sociology. 1 st edition. London: Butterworth.
Denzin, N.K. ( 1 978) The Research Act. A Theoretical Introduction to Sociological Methods.
2nd. edition. New York: McGraw-Hill.
Denzin, N.K. ( 1 988) 'Qualitative analysis for social scientists', Contemporary Sociology,
1 7 (3) : 430-2.
Denzin, N.K. (1 989) Interpretive Biography, London: Sage.
Derrida, j. (1 976) Of Grammatology. Tr. G.c. Spivak. Baltimore, MD: Johns Hopkins
University Press.
Derrida,j. (1 977) 'Signature, event, context', Glyph, 1 : 1 72-97.
Derrida,j. ( 1 988) Limited Inc. Evanston, IL: Northwestern University Press.
Derrida,j. (1991) A Derrida Reader. Between the Blinds. Ed. P. Kamuf. London: Harvester
Wheatsheaf.
Desrosieres, A . ( 1 994) 'Official statistics and business: history, classifications, uses', in
L. Bud-Frierman (ed.), Information Acumen. The Understanding and Use of Knowledge
in Modern Business. London: Routledge. pp. 1 68-86.
DeVault, M.L. (1 987) 'Women's talk. Feminist strategies for analyzing research inter­
views' , Women and Language, 1 0: 33-6.
DeVault, M.L. ( 1 990) 'Talking and listening from women's standpoint. Feminist strate­
gies for interviewing and analysis', Social Problems, 37: 96-1 1 6.
De Vaus, D.A. (1 996) Surveys in Social Research. 4th edition. London: UCL Press.
Dosse, F. ( 1 997a) History of Structuralism. Vol. 1. The Rising Sign. 1 945- 1 966.
Tr. D. Glassman. Minneapolis: University of Minnesota Press.
Dosse, F. (1 997b) History of Structuralism. Vol. 2. The Sign Sets. 1 967-Present.
Tr. D. Classman Minneapolis: University of Minnesota Press.
Douglas,j.D. ( 1 967) The Social Meanings of Suicide. Princeton, NJ: Princeton University
Press.

Copyrighted Material
178 BIBLIOGRAPHY

Douglas, M. ( 1966) Purity and Danger. An Analysis of the Concepts of Pollution and Taboo.
London: Routledge & Kegan Paul.
Dubrow, H. ( 1 982) Genre. London: Methuen.
du Gay, P., Hall, S., Janes, L., Mackay, H. and Negus, K. ( 1 997) Doing Cultural Studies.
The story of the Sony Walkman. London: Sage.
Durkheim, E. ( 1 9 1 5) The Elementary Forms of the Religious Life. Tr. ].W Swain. London:
Allen and Unwin.
Durkheim, E. ( 1 9 5 1 ) Suicide. A Study in Sociology. Tr. ].A. Spaulding and G. Simpson.
London: Routledge & Kegan Paul.
Edwards, A. , Hood, K., Mathews, E., Russell, D., Russell, 1., Barker,]., Bloor, M., Burnard,
P., Covey,]., Pill, R., Wilkinson, C. and Stott, N. (2000) 'The effectiveness of one-to-one
risk-communication interventions in health care', Medical Decision Making, 20: 290-7.
Ekeh, P. ( 1 974) Social Exchange Theory. London. Heinemann.
Ellis, C. and Bochner, P. (2000) 'Autoethnography, personal narrative, reflexivity', in
N.K. Denzin and YS. Lincoln (eds), Handbook of Qualitative Research. 2nd edition.
Thousand Oaks, CA: Sage. pp. 733-68.
·
Emmison, M. and Smith, P. (2000) Researching the Visual. London: Sage.
Evans-Pritchard, E.E. (1 937) Witchcraft, Oracles and Magic among the Azande. London:
Oxford University Press.
Evans-Pritchard, E.E. ( 1940) The Nuer. A description of the modes of the livelihood and politi­
cal institutions of a Nilotic people. Oxford: Clarendon Press.
Evenhuis, H.M., Kengen, M. and Eurlings, H. ( 1 992) 'Evaluation of a screening instru­
ment for dementia in ageing mentally retarded persons', Journal of Intellectual
Disability Research, 36: 337-47.
Fay, L.E. (2000) Shostakovich. A Life. Oxford: Oxford University Press.
Feyerabend, P. (1 975) Against Method. London: New Left Books.
Finnegan, R. (1 996) 'Using documents', in R. Sapsford andV.Jupp (eds) , Data Collection
and Analysis. London: Sage. pp. 1 38-5 1 .
Finley, M.1. (1 977) The World of Odysseus. 2nd edition. London: Chatto and Windus.
Fiske, ]. (1 989) Understanding Popular Culture. Boston, MA: Unwin Hyman.
Fitzgerald, F. Scott ( 1 925) The Great Gatsby. New York: New Directions Books.
Folstein, M.F., Folstein, S.E. and McHugh, P. R. ( 1 975) 'Mini Mental State. A practical
method for grading the cognitive state of patients for the clinician', Journal of
Psychiatric Research, 1 2: 189-98.
Fontana, A : and Frey, ].H. (2000) 'The interview. From structured questions to negoti­
ated text', in N.K. Denzin and YS. Lincoln (eds) , Handbook of Qualitative Research.
2nd edition. Thousand Oaks, CA: Sage. pp. 645-72.
Foucault, M. (1 970) The Order of Things. London: Tavistock.
Foucault, M. ( 1 972) The Archaeology of Knowledge. Tr. A. Sheridan. New York: Pantheon.
Foucault, M. (1 973) The Birth of the Clinic. Tr. A.M. Sheridan. London: Tavistock.
Foucault, M. ( 1977) Language, Counter-Memory, Practice. Oxford: Blackwell.
Foucault, M. (ed.) ( 1 978) I, Pierre Riviere. London: Penguin.
Foucault, M. (1 979) 'What is an author?', Screen, 20: 1 3-35.
Foucault, M. ( 1 988) Technologies of the Self. A Seminar with Michel Foucault. Ed.
L.H. Martin, H . Gutman and P. H . Hutton. London: Tavistock.
Foucault, M. ( 1 99 1 ) The Foucault Effect. Studies in Governmentality. Ed. G. Burchill,
C. Gordon and P. Miller. London: Harvester Wheatsheaf.

Copyrighted Material
 BIBLIOGRAPHY 1.79

Freeman, C.]. , Brown, A., Dunleavy, D. and Graham, F. (1 996) The Report of the Inquiry
into the care and treatment of Shaun Anthony Armstrong. Middlesborough: Tees District
Health Authority.
Freeman, D. ( 1 983) Margaret Mead and Samoa. Cambridge, MA: Harvard University
Press.
Fujimura, J.H. ( 1 996) Crafting Science. A Sociohistory of the Questfor the Genetics of Cancer.
Cambridge, MA: Harvard University Press.
Gabe,J. (ed.) (1 995) Medicine, Health and Risk. Sociological Approaches. Oxford: Blackwell.
Gadamer, H .-G. ( 1 975) Truth and Method. Tr. W Glen-Doepel, London: Sheed and
Ward.
Garfinkel, H. ( 1 967) Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice Hall.
Gay, P ( 1 988) Freud. A Lifefor our Time. New York: WW Norton.
Gedye, A. ( 1 995) Dementia Scale for Down Syndrome. Manual. Gedye Research and
Consulting,Vancouver, Be.
Geertz, e. ( 1988) Works and Lives. The Anthropologist as Author. Cambridge: Polity
Press.
Geertz, e. ( 1 993) The Interpretation of Cultures. London: Fontana.
Gerth, H. and Mills, e.W (1 953) Character and Social Structure. New York: Harcourt,
Brace.
Giddens, A. ( 1 978) 'Positivism and its critics', in T. Bottomore and R. Nisbet (eds), A
History of Sociological Analysis. London: Heinemann. pp. 237-86.
Giddens, A. (1 984) The Constitution of Society. Cambridge: Polity Press.
Giddens, A. (1991) Modernity and Self-Identity. Self and Society in the Late Modern Age.
Stanford, CA: Stanford University Press.
Gilbert, G.N. and Mulkay, M. (1984) Opening Pandora's Box. A Sociological Analysis of
Scientists' Discourse. Cambridge: Cambridge University Press.
Gill, A.M. and Whedbee, K. ( 1 997) 'Rhetoric', in T.A. van Dijk (ed.), Discourse as
Structure and Process. Vol. 1. London: Sage. pp. 1 57-84.
Gill, R. (2000) 'Discourse analysis', in M.W Bauer and G. Gaskell (eds) , Qualitative
Researching with Text, Image and Sound. London: Sage.
Gillies, D. (1 993) Philosophy of Science in the Twentieth Century. Oxford: Blackwell.
Glaser, B.G. and Strauss, A.L. ( 1 965) Awareness of Dying. Chicago: Aldine.
Glaser, B. and Strauss, A. (1 967) The Discovery of Grounded Theory. Strategies for
Qualitative Research. Chicago: Aldine.
Goffinan, E. ( 1 959) The Presentation of the Self in Everyday Life. New York: Doubleday
Anchor Books.
Gogol, N.V. ( 1 996) Dead Souls. Tr. B. Guilbert. London:Yale University Press.
Goode, WJ. and Hatt, PK. ( 1952) Methods in Social Research. New York: McGraw-Hill.
Goody, J. ( 1 968) Literacy in Traditional Societies. Cambridge: Cambridge University
Press.
Goody, J. (1 977) The Domestication of the Savage Mind. Cambridge: Cambridge
University Press.
Gould, S.]. ( 1 989) Wondeiful Life. The Burgess Shale and the Nature of History. New York:
WW Norton.
Grafton, A. ( 1997) The Footnote. A Curious History. London: Faber.
Greimas, A.]. ( 1 987) On Meaning. Selected Writings in Semiotic Theory. Tr. P.]. Perron and
F.H. Collins. London: Francis Pinter.

Copyrighted Material
180 BIBLIOGRAPHY

Grossmith, G. and Grossmith, W ( 1 999) The Diary of a Nobody. London: Penguin.

Gubrium, J.F., Holstein, J.A. and Buckholdt, D.R. ( 1 994) Constructing the Life Course.
New York: General Hall.
Hacking, 1. (1 990) The Taming of Chance. Cambridge: Cambridge University Press.
Hacking, I. ( 1 995) Rewriting the Soul. Multiple Personality and the Sciences of Memory.
Princeton, NJ: Princeton University Press.
Haddon, A.C. ( 1 904) Reports of the Cambridge Anthropological Expedition to Torres Straits.
Vol. VI. Society, Magic and Religion of the Western Islanders. Cambridge: Cambridge
University Press.
Hak, T (1 992) 'Psychiatric records as transformations of other texts', in G. Watson and
R.M. Seiler (eds), Text in Context. Contributions to Ethnomethodology. London: Sage.
pp. 1 38-55.
Hall, S. ( 1 972) Encoding and Decoding in the Television Discourse. Media series SP No. 7.
University of Birmingham, Centre for Contemporary Cultural Studies.
Hancock, G. ( 1 999) Heaven's Mirror. Quest for the Lost Civilisation. Harmondsworth:
Penguin.
Handlin, 0. ( 1 953) The Uprooted. From the Old World to the New. London: Watts.
Harding, N. and Palfrey, C. ( 1 997) The Social Construction of Dementia. Confused
Professionals? London: Jessica Kingsley.
Hardman, C.E. (2000) Other Worlds. Notions of Self and Emotion Among the Lohorung Rai.
Oxford: Berg.
Harper, R.H.R., O'Hara, K.PA., Sellen, AJ and Duthie, DJR. ( 1 997) 'Towards a
paperless hospital?', British Journal of Anaesthesia, 78: 762-7.
Harris, R. ( 1 996) Selling Hitler. The Story of the Hitler Diaries. London: Arnold.
Healy, D. ( 1 997) The Antidepressant Era. Cambridge, MA: Harvard University Press.
Heidegger, M. (1 967) What Is a Thing? Tr. WB. Barton and V. Deutsch. Chicago: Henry
Regnery.
Hertz, R. ( 1 960) Death and the Right Hand. Tr. R. and C. Needham. London: Cohen &
West.
Hindess, B. (1 973) The Use of Official Statistics in Sociology. London: Macmillan.
Houtkoup-Steenstra, H. (2000) Interaction and the Standardized Survey Interview.
Cambridge: Cambridge University Press.
Howells, C. ( 1 998) Derrida. Deconstruction: From Phenomenology to Ethics. Cambridge:
Polity Press.
Hughes, D. and Griffiths, L. ( 1 999) 'On penalties and the Patient's Charter: centralism v
de-centralised governance in the NHS', Sociology of Health and fllness, 21 ( 1 ) : 7 1 -94.
Hughes, TP ( 1 989) 'The evolution of large technological systems', in WE. Bijker,
T Pinch and TP Hughes (eds), The Social Construction of Technological Systems.
Cambridge, MA: MIT Press. pp. 5 1 -82.
Iser,W ( 1 989) Prospecting. From Reader Response to Literary Anthropology. Baltimore, MD:
Johns Hopkins University Press.
Jackson, HJ (200 1 ) Marginalia. Readers Writing in Books. New Haven, CT: Yale Univer­
sity Press.
Jones, E. ( 1 964) The Life and Work of Sigmund Freud. Ed. L. Trilling and S. Marcus.
Harmondsworth: Penguin.
Jonsson, L. and Linell, P (1991) 'Story generations: From dialogical interviews to
written reports in police interrogations'. Text, 1 1 (3): 1 9-40.

Copyrighted Material
 BIBLIOGRAPHY 181

Jordan, M. (1 995) The Encyclopaedia of Fungi ofBritain and Europe. Newton Abbot: David
and Charles.
Kapuscinski, R. (2001 ) 'Death in Sudan'. Tr. K Glowczewska. New York Review of
Books, 48 (7): 4-ti .
Keating, P. and Cambrosio, A. (2000) "'Real compared to what?" Diagnosing leukemias
and lymphomas', in M. Lock, A. Young and A. Cambrosio (eds) , Living and Working
with the New Medical Technologies. Intersections of Inquiry. Cambridge: Cambridge
University Press. pp. 1 03-34.
Keller, E . F. (2000) The Century of the Gene. Cambridge, MA: Harvard University Press.
Klockars, C.B. ( 1 975) The Professional Fence. London: Tavistock.
Kluckhohn, C. ( 1 945) 'The personal document in anthropological science', in
L. Gottschalk, C. Kluckhohn and R. Angell (eds), The Use of Personal Documents in
History, Anthropology, and Sociology. New York: Social Science Research Council
Bulletin, 53: 78- 1 73.
Knorr-Cetina, KD. (1 983) 'The ethnographic study of scientific work. Towards a con­
structivist interpretation of science', in K D. Knorr-Cetina and M. Mulkay (eds) ,
Science Observed. London: Sage. pp. 1 1 5-40.
Knox, B. ( 1 990) ' Introduction', in Homer. The Illiad. Tr. Robert Fagles. New York:
Viking Penguin.
Knox, B. ( 1 996) 'Introduction', in Homer. The Odyssey. Tr. Robert Fagles. London:
Viking.
Kress, G., Leite-Garcia, R. and van Leeuwen, T, (1 997) 'Discourse semiotics', in
T.A. van Dijk (ed.), Discourse as Structure and Process. Vol. 1 . London: Sage. pp. 257-9 1 .
Kristeva,]. ( 1 980) Desire in Language. A Semiotic Approach to Literature and Art. New York:
Columbia University Press.
Kwasnik, B.H. ( 1 99 1 ) 'The importance of factors that are not document attributes in
the organization of personal documents'. Journal of Documentation, 47 (4): 389-98.
Langness, L.L. (1 965) The Life History in Anthropological Science. New York: Holt,
Rinehart and Winston.
Latour, B. ( 1 983) 'Give me a laboratory and I will raise the world', in K.D. Knorr­
Cetina and M. MuIkay (eds), Science Observed. London: Sage. pp. 1 4 1-70 .
.
Latour, B. ( 1 987) Science in Action. How to Follow Engineers and Scientists Through Society.
Buckingham: Open University Press.
Latour, B. ( 1 988) The Pasteurization of France. Tr. A. Sheridan. London: Harvard
University Press.
Latour, B. and Woolgar, S. ( 1 979) Laboratory Life. The Social Construction of Scientific Facts.
London: Sage.
Law, ]. ( 1 994) Organising Modernity. Oxford: Blackwell.
Law, ]. and Hassard, ]. (eds) ( 1 999) Actor-Network Theory and After. Oxford: Blackwell.
Lepper, G. (2000) Categories in Text and Talk. London: Sage.
Levi-Strauss, C. ( 1 969) Totemism. Tr. R. Needham. Harmondsworth: Penguin.
Levitas, R. and Guy, M. (eds) ( 1 996) Interpreting Official Statistics. London: Routledge.
Lewis, G. ( 1 975) Knowledge of fllness in a Sepik Society. A study of the Gnau, New Guinea.
London: Athlone Press.
Lidchi, H. ( 1 997) 'The poetics and politics of exhibiting other cultures', in S. Hall
(ed.), Representation. Cultural Representations and Signifying Practices. London: Sage.
pp. 1 5 1-222.

Copyrighted Material
182 BIBLIOGRAPHY

Lieberson, S. ( 1 992) 'Small N's and big conclusions. An examination of the reasoning
in comparative studies based on a small number of cases', in c.c. Ragin and
H.S. Becker (eels), What is a Case? Exploring the Foundations of Social Inquiry.
Cambridge: Cambridge University Press. pp. 1 0-1 1 8 .
Lienhardt, G . ( 1 96 1 ) Divinity and Experience. The Religion of the Dinka. Oxford: Oxford
University Press.
Lincoln,YS. and Gubba, E.G. (1 985) Naturalistic Inquiry. Beverly Hills, CA: Sage.
Lynch, M. ( 1 985) Art and Artifact in LAboratory Science. A Study of Shop Work and Shop
Talk in a Research LAboratory. London: Routledge & Kegan Paul.
Lynch, M. ( 1 990) 'The externalized retina: Section and mathematization in the visual
documentation of objects in the life sciences', in M. Lynch and S. Woolgar (eds) ,
Representation in Scientific Practice. London: MIT Press. pp. 1 53-86.
Lynd, R.S. and Lynd, H .M. (1 929) Middletown. New York: Harcourt, Brace.
Machler, S. (200 1 ) The Wilkomirski Affair. A Study in Biographical Truth. Tr. ].E. Wooels .
New York: Schocken Books.
Mackay, H. ( 1 997) 'Consuming communication technologies at home', in H. Mackay
(ed.), Consumption and Everyday Life. London: Sage.
Madge,]. ( 1 963) The Origins of Scientific Sociology. London: Tavistock.
Malinowski, B. ( 1 922) Argonauts of the Western Pacific. London: Routledge.
Malinowski, B. ( 1 935) Coral Gardens and their Magic. London: George Allen and Unwin.
Malinowski, B. ( 1 967) A Diary in the Strict Sense of the Term. Tr. N. Guterman. London:
Routledge & Kegan Paul.
Matthews, ].R. ( 1 995) Quantification and the Questfor Medical Certainty. Princeton, NJ:
Princeton University Press.
Mattingly, C. and Garro, L.c. (eds) (2000) Narrative and the Cultural Construction of fllness
and Healing. Berkeley, CA: University of California Press.
Mead, G.H. ( 1 934) Mind, Self and Society from the Standpoint of a Social Behaviorist. Ed.
C.W Morris. Chicago: University of Chicago Press.
Mead, M. ( 1 928) Coming of Age in Samoa. A Psychological Study of Primitive Youth for
Western Civilization. New York: William Morrow.
Meltzer, H . , Gill, B., Petticrew, M. and Hinds, K. (1 995) The Prevalence of Psychiatric
Morbidity among Adults Living in Private Households. OPCS Surveys of Psychiatric
Morbidity in Great Britain. Report 1 . London: HMSo.
Miles, C. and Irvine,]. ( 1979) 'The critique of official statistics', in]. Irvine, C. Miles and
I. Evans (eds), Demystifying Official Statistics. London: Pluto.
Mill, j.S. ( 1 843) A System of Logic, Ratiocinative and Inductive. Being a Connected View of
the Principles of Evidence, and the Methods of Scientific Investigations. London: Parker.
Mill, ].S. ( 1 989) Autobiography. London: Penguin.
Miller, N. and Morgan, D. ( 1 993) 'Called to account: The CV as an autobiographical
practice', Sociology, 27 ( 1 ) : 133-43.
Ministry of Health ( 1 954) Reports on Public Health and Medical Subjects. No. 95. Mortality
and Morbidity during the London Fog of December 1 952. London: HMSo.
Mol, A. ( 1 999) 'Ontological politics. A word and some questions', in ]. Law and
]. Hassard (eds), Actor Network Theory and After. Oxford: Blackwell. pp. 75-89.
Mol, A. (2000) 'Pathology and the clinic. An ethnographic presentation of two athero­
scleroses', in M. Lock, A.Young and A. Cambrosio (eels), Living and Working with the
New Medical Technologies. Intersections of Inquiry. Cambridge: Cambridge University
Press. pp. 82-1 02.

Copyrighted Material
 B I BL I O G RAPHY 183

Monaghan, L. (2001 ) Bodybuilding Drugs and Risk. London: Routledge.

Murphy, EA., Fauquet, C.M., Bishop, 0., Ghabrial, S., Jarvis, A., Martelli, G., Mayo, M.
and Summers, M. (eds) ( 1 995) Virus Taxonomy. Classification and Nomenclature of
Viruses. Sixth Report of the International Committee on Taxonomy of Viruses. Virology
Division. International Union of Microbiological Societies. Vienna: Springer Verlag.
Myers, G. (1 990) Writing Biology. Texts in the Construction of Scientific Knowledge.
London: University of Wisconsin Press.
Nagel, E. (1961) The Structure of Science. London: Routledge & Kegan Paul.
NHS Executive (1 994) Guidance on the Discharge of Mentally Disordered People and their
Continuing Care in the Community. HSG(94)27/LASL(94)4. London: Department of
Health.
Nukuga, Y. (2002) 'Between tradition and innovation in new genetics: The continuity
of medical pedigrees and the development of combination work in the case of
Huntington's disease' . New Genetics and Society, 2 1 (1): 39-64.
Nukuga, Y. and Cambrosio, A. ( 1997) 'Medical pedigrees and the visual production of
family disease in Canadian and Japanese genetic counselling practice', Sociology of
Health and fliness, 1 9 (5) : 29-55 .
Oakley, A. (1981) ' Interviewing women. A contradiction i n terms', i n H . Roberts (ed.),
Doing Feminist Research. London: Routledge & Kegan Paul.
Ong, W]. ( 1 982) Orality and Literacy. The Technologizing of the Word. London: Methuen.
Outhwaite, W ( 1 975) Understanding Social Life. The Method Called Verstehen. London:
Allen and Unwin.
Pepys, S. (1 970) The Diary of Samuel Pepys. 1 1 Vols. Ed. R. Latham and W Mathews.
Berkeley, CA: University of California Press.
Phillips, T. (ed.) (1 995) Africa. The Art of a Continent. London: Prestel Verlag.
Philo, G. (ed.) (1 996) Media and Mental Distress. London: Longman.
Pilnick, A. (2002) Genetics and Society. An Introduction. Buckingham: Open University
Press.
Pinch, T.]. and Bijker, WE. ( 1 989) 'The social construction of facts and artifacts: Or how
the sociology of science and the sociology of technology might benefit each other'
in WE. Bijker, T.P. Hughes and T. Pinch, (eds), The Social Construction of Technological
Systems. Cambridge, MA: MIT Press. pp. 1 7-50.
Platt, ]. (1981a) 'Evidence and proof in documentary research. Some shared problems
of documentary research'. Sociological Review, 29 ( 1 ) : 31-52.
Platt, ]. ( 1 98 1 b) 'Evidence and proof in documentary research. Some shared problems
of documentary research'. Sociological Review, 29 ( 1 ) : 53-66.
Platt, ]. ( 1 996) A History of Sociological Research Methods in America. 1 920-1 960.
Cambridge: Cambridge University Press.
Plummer, K. (2001 ) Documents of Life. 2. An Invitation to Critical Humanism. London: Sage.
Poole, R. ( 1 969) 'Introduction', in C. Levi-Strauss. Totemism. Tr. R. Needham.
Harmondsworth: Penguin.
Popper, K.R. (1 959) The Logic of Scientific Discovery. London: Hutchinson.
Popper, K.R. ( 1 963) Conjectures and Refutations. Growth of Scientific Knowledge. London:
Routledge.
Porter, T.M. ( 1994) 'Information, power and the view from nowhere', in L. Bud-Frierman
(ed.), Information Acumen. The Understanding and Use of Knowledge in Modern Business.
London: Routledge. pp. 2 17-30.
Potter,]. (1996) Representing Reality. Discourse, Rhetoric and Social Construction. London: Sage.

Copyrighted Material
184 BIBLIOGRAPHY

Prior, L. (1 988) 'The architecture of the hospital. A study in spatial organisation and
medical knowledge', BritishJournal oj Sociology. 39 (March) : 86-1 1 5 .
Prior, L. (1 989) The Social Organization oj Death. Basingstoke: Macmillan.
Prior, L. ( 1 992) 'The local space of medical knowledge. Disease, illness and hospital
architecture', in J. Lachmund and G. Stollberg (eds), The Social Construction oj fllness.
Stuttgart: Franz Steiner Verlag. pp. 67-84.
Prior, L. ( 1 993) The Social Organization oj Mental fllness. London and Newbury Park,
CA: Sage.
Prior, L. ( 1 997) 'Following in Foucault's footsteps. Text and context in qualitative
research', in D. Silverman (ed.), Qualitative Analysis. Issues of Theory and Method. Sage.
pp. 63-79.
Prior, L., Pang, L.c. and See, R H . (2000) 'Beliefs and accounts of illness. Views from
two Cantonese speaking communities in England', Sociology of Health and fllness,
22 (6) : 8 1 5-39.
Prior, L., Wood, F. , Gray, J., Pill, R. and Hughes, D. (2002) 'Making risk visible: The role
of images in the assessment of (cancer) genetic risk. Health, Risk and Society, 4 (3).
Proctor, R.N. ( 1 995) Cancer Wars. New York: Basic Books.
Psathas, G. ( 1 979) 'Organizational features of direction maps', in G. Psathas (ed.),
Everyday Language. Studies in Ethnomethodology. New York: Irvington. pp. 203-25.
Putnanl, H . (1 988) Representation and Reality. Cambridge, MA: MIT Press.
Quine, V.W (1 953) From a Logical Point oj View. 9 Logico-Philosophical Essays.
Cambridge, MA: Harvard University Press.
Rabinow, P. (1 977) Riflections on Fieldwork in Morocco. London: University of California
Press.
Radcliffe-Brown, A.R. (1 922) The Andaman Islanders. A Study in Social Anthropology.
Cambridge: Cambridge University Press.
Radcliffe-Brown, A.R. ( 1 958) 'The method of ethnology and social anthropology', in
M.N. Srinivas (ed.), Method in Social Anthropology. Chicago: Chicago University Press.
Ragin, c.c. ( 1 992) ' Introduction', in c.c. Ragin and H.S. Becker (eds) , JilIhat is A
Case? Cambridge: Cambridge University Press. pp. 1-17.
Rapp, R. (2000) Testing women, Testing the Fetus. The Social Impact oj Amniocentesis in
America. New York: Routledge.
Reiner, R. ( 1 996) 'The case of the missing crimes', in R. Levitas and M. Guy (eds) ,
Interpreting Official Statistics. London: Routledge.
Reinharz, S. (1 992) 'Feminist interviewing', in Feminist Methods in Social Research.
Oxford: Oxford University Press.
Reith, M. (1 998) Community Care Tragedies. London:Venture Press.
Richards, L. ( 1 999) Using NVivo in Qualitative Research. Thousand Oaks, CA: Sage.
Ricoeur, P. ( 1 977) 'The model of the text. Meaningful action considered as a text', in
F. Dallmayr and T. McCarthey (eds) , Understanding and Social Inquiry. Notre Dame,
IN: Notre Dame University Press.
Ricoeur, P. ( 1 98 1 ) Hermeneutics and the Human Sciences. Ed. tr. and intro.J.B. Thompson.
Cambridge: Cambridge University Press.
Ridpath, I. (ed.) ( 1 989) Norton's 2000. 0. Star Atlas and Riference Handbook. 1 8th edition.
Harlow: Longman.
Ritchie, J.H . , Dick, D. and Lingham, R. ( 1 994) The Report of the Inquiry into the Care
and Treatment of Christopher Clunis. London: HMSo.

Copyrighted Material
 BIBLIOGRAPHY 185

Ritvo, H. ( 1 997) The Platypus and the Mermaid, and Other Figments of the Classifying
Imagination. Cambridge, MA: MIT Press.
Rousseau , J.-J. (2000) Confessions. Tr. A. Scholar. Oxford: Oxford University Press.
Royal College of Physicians, Royal College of Psychiatrists and Royal College of
General Practitioners (1 996) Chronic Fatigue Syndrome: Report of a Joint Working
Group of the Royal Colleges of Physicians, Psychiatn·sts and General Practitioners. London:
Royal College of Physicians.
Ryan, G.W and Bernard, H.R. (2000) 'Data management and analysis methods', in
N.K. Denzin and YS. Lincoln (eds) , Handbook of Qualitative Research. 2nd edition.
Thousand Oaks, CA: Sage. pp. 769-802.
Ryle, G. ( 1 990) The Concept of Mind. Harmondsworth: Penguin.
Sackett, D.L. (2000) Evidence-based Medicine. How to Practice and Teach EBM. Edinburgh:
Churchill Livingstone.
Sacks, H. ( 1 984) 'On doing "being ordinary''', in J.M. Atkinson and J. H �ritage (eds),
Structures of Social Action. Studies in Conversational Analysis. Cambridge: Cambridge
University Press.
Sacks, 0. (1 986) The Man who Mistook his Wife for a Hat. London: Picador.
Sacks, H. ( 1 992) Lectures of Conversation. 2 Vols. Oxford: Blackwell.
Sarangi, S. (1 998) 'Rethinking recontextualization in professional discourse studies.
An epilogue', Text, 1 8 : (2) : 301-1 8 .
Saussure, F. d e ( 1 983) Course in General Linguistics. Ed. C . Bally and A. Sechehave in
collaboration with A. Riedlinger. Tr. R. Harris. London: Duckworth.
Schreiber, F.R. ( 1974) Sybil. London: Allen Lane.
Schutz, A. ( 1 962) Collected Papers. Vol. 1. The Problem of Social Reality. The Hague:
Martinus Nijhoff.
Scott, J. ( 1 990) A Matter of Record. Cambridge: Polity Press.
Scott, M.B. and Lyman, S.M. ( 1 968) 'Accounts', American Sociological Review, 33: 46-62.
Seale, C. ( 1 998) Constructing Death. The Sociology of Dying and Bereavement. Cambridge:
Cambridge University Press.
Seale, C. ( 1 999) The Quality of Qualitative Research. London: Sage.
Seale, C. (2001) 'Sporting cancer. Struggle language in news reports of people with
cancer', Sociology of Health and fllness, 23 (3) : 308-29.
Serres, M. ( 1 995) Conversations on Science, Culture and Time. With Bruno Latour.
Tr. R. Lapidus. Ann Arbor, MI: The University of Michigan Press.
Sharpe, K. (2000) Reading Revolutions. The Politics of Reading in Early Modem England.
New Haven, CT:Yale University Press.
Shaw, C.R. (1931) The Natural History of a Delinquent Career. Chicago. University of
Chicago Press.
Shelley, M.W ( 1 996) Frankenstein or the Modem Prometheus. London: Pickering and
Chatto.
Silverman, D. (1 993) Interpreting Qualitative Data. Methods for Analysing Talk, Text and
Interaction. London: Sage.
Silverman, D. (ed.) ( 1996) Qualitative Research. Theory, Method and Practice. London: Sage.
Silverman, D. (1 998) Harvey Sacks. Social Science and Conversation Analysis. Cambridge:
Polity Press.
Simmons, W (ed.) ( 1 942) Sun Chiif. The Autobiography of a Hopi Indian. New Haven,
CT: Yale University Institute of Human Relations.

Copyrighted Material
186 BIBLIOGRAPHY

Skolbekken, J.-A. (1 995) 'The risk epidemic in medical journals', Social Science and
Medicine, 40: 291-305.
Smith, D. ( 1 978) 'K is mentally ill. The anatomy of a factual account', Sociology, 1 2 :
.
23-53.
Smith, D.E. (1 984) 'Textually mediated social organization', InternationalJournal of Social
Science, 36 ( 1 ) : 59-75.
Smith, D.E. ( 1 990) Texts, Facts and Femininity. Exploring the Relations of Ruling. London:
Routledge.
Smith, R. (1 997a) 'Authorship. Time for a paradigm shift', British Medical Journal.
3 1 4: 992.
Smith, R. (1 997b) 'Authorship is dying. Long live contributorship', British Medical
Journal, 3 1 5 : 696.
Smith, S. and Watson, J. (eds) (1 992) De/colonizing the Subject. The Politics of Gender in
Women's Autobiography. Minneapolis: University of Minnesota Press.
Sontag, S. (1 978) fllness As Metaphor. New York: Farrar, Strauss and Giroux.
Spiegel, H. (1 997) 'Sybil - the making of a disease', New York Review of Books, XLIV
(7): 60-4.
Spiegelman, D., Colditz, G.A., Hunter, D. and Hertzmark, E. (1 994) 'Validation of the
Gail et al. , model for predicting individual breast cancer risk' ,Journal of the National
Cancer Institute, 86: 600-8.
Stanley, L. and Morgan, D. (eds) ( 1 993) 'Biography and autobiography in sociology',
Sociology, 27 (1): 1-197.
Stannard, D.E. ( 1 977) The Puritan Way of Death. New York: Oxford University Press.
Star, S.L. ( 1 989) 'The structure of ill-structured solutions: Boundary objects and
heterogeneous distributed problem solving', in M. Hunhs and L. Casser (eds) ,
Distributed Artificial Intelligence 2. London: Pitman. pp. 37-54.
Star, S.L. and Griesemer, J.R. ( 1 989) 'Institutional ecology "translations" and boundary
objects. Amateurs and professionals in Berkeley's Museum Vertebrate Zoology,
1 907-39', Social Studies of Science, 1 9: 387-420.
Stocking, G.W ( 1 983) 'The ethnographer's magic', in G.W Stocking (ed.), Observers
Observed. Madison,WI: University of Wisconsin Press. pp. 70-1 1 8 .
Sullivan, G . ( 1 999) Margaret Mead, Gregory Bateson, and Highland Bali. Fieldwork
Photographs of Bayung Gede, 1 93 6- 1 939. Chicago: University of Chicago Press.
Sutherland, E.H. (1 937) The Professional Thiefby a Professional Thief. Chicago: University
of Chicago Press.
Tambiah, S.]. (1 968) 'Literacy in a Buddhist village in North-east Thailand', in J. Goody
(ed.), Literacy in Traditional Societies. Cambridge: Cambridge University Press.
pp. 86-131 .
Taylor, C. ( 1 987) 'Interpretation and the sciences of man', in P. Rabinow and
M.W Sullivan (eds) , Interpretive Social Science. A Second Look. London: University of
California Press. pp. 33-8 1 .
Taylor, S. ( 1 982) Durkheim and the Study of Suicide. London: Macmillan.
Tedlock, B. (2000) 'Ethnography and ethnographic representation', in N.K. Denzin and
Y.S. Lincoln (eds), Handbook of Qualitative Research. 2nd edition. Thousand Oaks. CA:
Sage. pp. 455-86.
Thigpen, C.H. and Cleckley, H. (1 957) The Three Faces of Eve. New York: McGraw-Hill.
Thomas, WI. and Znaniecki, F. (1 958) The Polish Peasant in Europe and America. NewYork:
Dover.

Copyrighted Material
 BIBLIOGRAPHY 187

Thrasher, EM. (1 927) The Gang. A Study of 1 , 3 1 3 Gangs in Chicago. Chicago:

University of Chicago Press.
Tischkowitz, M., Wheeler, D., France, E . , Chapman, c., Lucassen, A., Sampson, J.,
Harper, P. , Krawczak, M. and Gray, J. (2000) 'A comparison of methods currently
used in clinical practice to estimate familial breast cancer risks', Annals of Oncology,
1 1 (4): 45 1-4.
Tolstoy, Leo ( 1 889) War and Peace. Tr. N.H. Dole. London: Walter Scott.
Torfing,J. ( 1 999) New Theories of Discourse. Laclau, Mouffe and Zizek. Oxford: Blackwell.
Tukey, J.W ( 1 977) Exploratory Data Analysis. New York: Addison-Wesley.
Tylor, E.B. ( 1 87 1 ) Primitive Culture. 2 Vols. London: John Murray.
Unwin, N., Carr, S. and Leeson, J. (1 997) An Introductory Study Guide to Public Health
and Epidemiology. Buckingham: Open University Press.
van Dijk, T.A. (ed.) (1 997) Discourse as Structure and Process. 2 Vols. London: Sage.
van Maanen, J. ( 1 988) Tales oj the Field. Chicago: University of Chicago Press.
Vaughan, D. (1 983) Controlling Unlawful Organizational Behavior. Social Structure and
Corporate Misconduct. Chicago: University of Chicago Press.
Vaughan, D. (1 996) The Challenger Launch Decision. Risky Technology, Culture, and
Deviance at NASA. London: University of Chicago Press.
Vaughan, D. ( 1 999) The Role of the organization in the production of techno­
scientific knowledge', Social Studies of Science, 29 (6): 9 1 3-43.
von Goethe, J. W (1 986) Faust. Part One. Tr. P. Wayne. Harmondsworth: Penguin.
Watson, G. and Seiler, R.M. (eds) (1 992) Text in Context. Contributions to
Ethnomethodology. London: Sage.
Watson, L.c. and Watson-Franke, M.B. (1 985) Interpreting Life Histories. An Anthro­
pological Inquiry. New Brunswick, NJ: Transaction Books.
Weber, M. ( 1 930) The Protestant Ethic and the Spirit of Capitalism. Tr. T. Parsons. London:
Unwin.
Weber, M. ( 1 979) Economy and Society. 2 Vols. Ed. G. Roth and C. Wittich. Berkeley,
CA: University of California Press.
Weber, R.P. (1 990) Basic Content Analysis. Newbury Park, CA: Sage.
Webster, A. (1991) Science, Technology and Society. Basingstoke: Macmillan.
Weiner, A., Wessely, S. and Lewis, G. (1 999) ' ''You don't give me flowers anymore". An
analysis of gift-giving to medical and psychiatric inpatients', Social Psychiatry and
Psychiatric Epidemiology. 34: 1 36-40.
Wheeler, S. (ed.) ( 1 969) On Record. Files and Dossiers in American Life . New York: Russell
Sage Foundation.
Whyte, WE (1 955) Street Corner Society. Chicago: Chicago University Press.
Wilkomirski, B. (1 996) Fragments. Memories ofa Childhood, 1 93 9- 1 948.Tr. C.B.Janeway.
London: Picador.
Wittgenstein, L. (1 958) Preliminary Studies for the 'Philosophical Investigations'. Ed.
R. Rees. Oxford: Blackwell.
Woolgar, S. (1981) 'Discovery. Logic and sequence in a scientific text', in K.D. Knorr
and R. Krohn (eds), The Social Process oj Scientific Investigation. Dordrecht: D. Reidel.
pp. 239-68.
Woolgar, S. ( 1 988) Science. The Very Idea. London: Tavistock.
World Health Organization ( 1 992) International Statistical Classification of Diseases and
Related Health Problems. t Oth. Revision. 3 Vols. London: HMSO.
World Health Organization ( 1 998) World Health Statistics Annual. 1 996. WHO: Geneva.

Copyrighted Material
188 BIBLIOGRAPHY

Yin, R.K. (1 994) Case Study Research. Design and Methods. 2nd edition. London: Sage.
Young, A. ( 1995) The Harmony of fllusions. Inventing Post-traumatic Stress Disorder.
Princeton, NJ: Princeton University Press.
Zerubavel, E. ( 1 979) Patterns of Time in Hospital Life. A Sociological Perspective. London:
University of Chicago Press.
Zhukov, G.K. (1971) The Memoirs of Marshal Zhukov. London: Cape.
Zimmerman, D.H. (1 969) 'Record-keeping and the intake process in a public welfare
agency', in S.Wheeler (ed.), On Record. Files and Dossiers in American Life. New York:
Russell Sage Foundation. pp. 3 1 9-45.
Zimmerman, D.H . and Pollner, M. (1971) 'The everyday world as a phenomenon', in
J.D. Douglas (ed.), Understanding Everyday Life. London: Routledge & Kegan Paul.
pp. 80-103.
Zorbaugh, H.W (1 929) The Gold Coast and the Slum. A Sociological Study of Chicago's
Near North Side. Chicago: University of Chicago Press.

Copyrighted Material
I ndex

accounts, 90-1, 105, 117 author, cont.

actants, 12, 28, 93, 105 recruitment of, 142
action-at-a-distance, 20, 57-8, as source of meaning, 109, 111, 112
67, 168 as subject, 11, 12 , 90
actor-network, 91 authorship, 18, 97, 131
ad-hocing, 41-2 autobiography, 17, 95, 101, 139
Africa, 1-2 , 165 autoethnography, 105
agency, 3, 129, 132 authenticity of documents, 169,
air pollution, 63-6 see also forgery
algorithms, 45
alphabet, as an ordering device, 47, 71 Barthes, R . , 18, 110, 123
Alzheimer's disease, 74-5, 77, see also being-at-risk, 37, 79-86
dementia behaviour,
analysis, measurement of , 76-7
Freudian, 107-8 reporting of, 55
of data, see data beliefs, as social properties, 91,
anonymity, 31, 65 117, 122
anthropology, 19-20, 97, 134-7, 140, bias in selection of documents, 150-7
see also ethnography Bible, 13-14, 22-3
Atkinson , P., 127, 138, 141 biographical work, 93-8, 103,
archaeology, 145, 162-3 see also autobiography
architectural plans as documents, 8-10 biographer, 96-8
Aries, P. , 7 blackboards as documents, 67
Aristotle , 11, 72 black box, 90
arrangement Bowker, G.C., 57, 60, 71, 86
in documents, 48, 72-3 boundaries and documents, 94, 171
of things, 25, 72, 88 boundary objects , 79, 87
art, 1-2 brains, images of, 73-6
assessment and documentation, Brody, H . 128
54-6, 58, 81-5 bureaucracy and writing, 60
astronomical maps, 167-8
audiotape as document, 105 Calion, M . , 91-2
auteur, 30-1, 48 cancer, 66, 79-81, 84-5, 115-16
author, cards, social distribution of, 169-70
as collective, 12 , 97, 131, 142 care, 63
as contributor, 11, 139, 142 cases,
death of, 18 defined , 45
de-centred , 90, 142 identified, 62, 93, 100
as function, 11, 12, 48, 90, 104, case study as method, 97, 153
126, 142 causes,
and readers, 16, 18, 89, 143 as social constructions, 33, 39, 65
hidden, 136, 142 vocabulary of, 33, 65, 90

Copyrighted Material
190 INDEX

cemeteries as records, 6-7 coroners, 39-41

CFS, see Chronic Fatigue Syndrome correspondence, 21-2, 151,
certification, 33 see a/so letters
characters, see writing and script counter-agents, see documents
charts, as agents
clinical , 83 counting, 31, 78, 114-16, 118-19
hospital , 57 crime, 31
psychiatric, 54 curriculum vitae as autobiography,
childhood, 8 95, 101, 103-4
chromosomes, 80, 88 cut off points and measurement,
Chronic Fatigue Syndrome, 118-21 45-6, 64, 77
Cicourel, A., 38, 44 Cyri llic, 81-3
circuits, Cytometer, 78
of communication, 100
cultura l , 169 Darnton, R . , 30, 47, 69, 71, 89
of production, 167 data,
circulation coding of, 31-2, 33-4, 41, 45, 76,
of documents, 54, 56, 62, 100, 157-61
169-70 collection of, 33-4, 44
of things, 169 enumeration strategies for, 21,
citations, 131, 142, 173 32-5, 42-5, 113-15, 116,
CIS-R, see Clinical Interview Schedule 118-19, 157-60
civilization, as concept, 134, 145 quantitative, 32-7, 43-6, 113-16
classification, selection of, 150-4
principles of, 25, 30-1, 72 statistical manipulation of, 36, 64
and social practice, 71, 72, 73, 78 death,
types of 33-4, 44, 56, 57, 72, causes of, 31-2, 65
86-7 , 88 organization of, 5-6
as world-view, 48, 70, 71 rates, 31
clients as defined by documents, 54-6 references to, 113-15
Clinical I nterview Schedule, 44-6, 76 de-centring
clinics, of author, 90
and records in, 50-1 of subject, 90
and talk, 54, 81-4, 86, 112 De Certeau, M . , 16, 47, 95, 109-10,
coding, see data 131, 134
Col lins, H . M . , 132, 142 decoding, 105
committees as authors, 65 dementia, identified by use of
computerized documents, see documents, 77
electronic documents Denzin , N., 146, 148
confessions, 93, 95, 105 depression identified by use of
consumption of documents, 10, 16, documents, 45-6
19, 28, 166-7, see a/so in accounts of illness, 102-21
document function, and users Derrida, J . , 4, 105, 110, 166, 171
of documents Diagnostic and Statistical Manual of
content of documents, 20-6, 115, Mental Disorders (DSM), 33, 44,
118-21, 172 46, 99, 142, 167
content analysis, 21, 114, 116, 122 diagrams, as documents, 83-6
contracts as documents, 67 diaries, 14-15 , 17, 152, 165
contributors as authors, 11 dictionaries, 24, 28, 71
conversation , see talk Diderot, 47-8, 71

Copyrighted Material
 INDEX 191

discourse, the concept of, 13, 25-6, documents, cant.

28, 29, 113, 115-16, 126-7 , social roles, 52-3, 56, 101, and
133, 141-2 subjects, 91, 93, 100-1, 103
disease, 46, 74-5 structure readers, 73, 131, 142, 168
DNA, 84-6 dossiers, 92, 105
documents, drawings, 8-10, 128, 135, 168
and data analysis, 149-61 DSM, see Diagnostic and Statistical
and identity, see identity Manual of Mental Disorders
and text, 5, 7 , 26, 28, 128, Durkheim, E., 25, 35-7, 43
see a/so text
as actors, 3, 14, 20, 66, 67 e-mail, 151, 171
as agents and counter-agents, editing as production , 16, 130
3, 13-14, 66, 91, 168 editors as authors, 130
as collective products, 11-13, electronic documents, 88, 103
26, 97, 142 encoding, 105
as field for research, 3, 5, 26, 149, encyclopaedias, 47-8, 70-1, 72, 101
152-3, 168 Enlightenment, The, 48, 148
as objects of consumption, see environment as object of study, 63-7
consumption epistemology, 113
as objects of production , see ethnography,
production nature of, 126-7, 136-7
as receptacles of content, 4, 14, origins, 134
20-2, 26, 42, 91, 122, role of informant, 19, 97-8, 144
147 , 172-3 role of invisible others, 136
as records, see records as science, 139-40
as resources, 3, 16, 42, as social construction, 137,
100, 165 142, 144
as sources of evidence, 14, 61-3, as research strategy
92, 147 , 149, 163-4 writing of, 137
as situated products, 4, 10, 15, see a/so field notes
19, 26, 51, 87 ethnomethodology, 38-9, 111
as technology, 4, 47, 71, 73, 81, Evans-Pritchard, E . E . , 138, 165
84-6, 96, 128, 150, 71-2 evidence
as tools, 167 nature of, 146-8, 162
as trace, 165 selection of, 150-4, 163-4
circulation and exchange of, 54, exchange of documents, see
56, 62 , 100, 142, 169-71 circulation
defined, 1-2 expertise and documents, 56, 85,
function versus content, 4, 21, 26, 102, 131
52, 66, 67-8, 91, 94, 94, 104 experts, 46, 65-6, 102-3, 105, 131
generative, 34, 43, 46-7
in organizational activity, 56, facts, manufacture of, 35-6, 38-9,
60-3, 81 42, 47, 63-6, 73-9, 82, 126,
make things visible, 73-86, 100, 129, 131, 167-8
128, 143, 168 fatigue, 115-18, see a/so Chronic
mediate social interactions, 51-2, Fatigue Syndrome
57, 59, 78-9, 81, 171 feminist methodology, 143
and other relationships, 126, 167 fiction, 101, 127, 134, 138,
produce things, 77, 82-3, 128, 143 139, 146
selection of, 150-4 field notes, 137, 134, 138

Copyrighted Material
192 I N D EX

fieldwork, see ethnography homicide, 61, 92-5

files, and psychiatric reports, 61-3
and clinics, 50, 56 Homer, 11
and Max Weber, 28 hospita l , 8-9, 54-7
and police, 3 hypertext, 88
see a/so records, reports hypothetico-deductive method, 148-9
findings, see facts hysteria, 107-8
fog, explanations of, 64-6
footnotes, 125, 131 I , 95-6, 98, 129, 133, 138-9
forgery of documents, 15, 155, 164 lCD, see International Statistical
Foucault, M . , 25-6, 33, 47, 65, 92, Classification of Diseases
95, 97, 110, 126 identity,
France, 30, 36, 92 and documentation , 20, 55, 57, 58,
Frank, Anne, 14-15 62-3, 91-2, 94, 96
Frankenstein, Dr, 14, 28, 168 as relational property, 98, 105
frequency counts in data analysis, creation of, 97-100, 102-3, 104
31, 119, 157-61 multiple, 98
Freud, S . , 107 theft of, 96, 105
Fujimura, J . , 47, 74 see a/so I , subjects and subjectivity
fungi, classification of, 72-3 i l lustrations, see images
images,
Gadamer, H .-G . , 112 in medical work, 75-7, 78, 83,
Garfinkel, H . , 38, 50-2 84-6
Geertz, C . , 127, 137 , 139, 140 in science 74, 84-6
genealogies, composition of, indexing documents 118,
19-20, 82, 83 compared to coding, 158-61
genetics, inscription devices, 67, 7 7 , 81-6,
and CFS, 119-20 128, 168
and mutations, 80, 84-5 International Statistical Classification
and pedigrees, 82-3 of Diseases (ICD), 33, 47, 167
and polymorphism, 85 interpretation ,
and risk, 80-3 of documents, 23, 82-6
gravestones as documents, 6-7 of meaning, 108
gender, 96 intertextuality, 122, 124, 130
generative documents, see documents interviews, 44-5, 97-8, 143, 158-9
genre, 125, 130, 134, 144 iterabil ity, 172
Goffman, E . , 28, 63, 101
governmentality and documents, 87 journals, 130
graphs, 35
grounded theory, 141, 154 kinship, 19-20
Goody, J . , 14, 131, 137, 166, 171-2 see a/so, genetic pedigrees
Guru systems, 89 Knorr-Cetina, K., 128, 133
knowledge, forms of, 127,
health, measurement of via 135, 138
documents, 32, 44-5, 76-7 Kula, The, 136, 137, 169-70
Heidegger, M . , 144
hereditary influences, see genetic laboratory, 84-6, 128, 129, 131, 132
pedigrees Lancet, The, 130
hermeneutics, 108, 112, 123 Latour, B . , 57, 67, 74, 7 7 , 83, 90,
HIV/AIDS, 102 91, 128, 130-1, 132, 171

Copyrighted Material
 INDEX 193

Leaflets, structure their readers 101, music as text, 17-18, 173

see also patient i nformation mutation, see genetics
leaflets Myalgic Encephalomyelitis (ME), see
letter, Chronic Fatigue Syndrome
as data, 21-2, 150-2
as form of writing, 151 Nagel, E . , 148
circulation of 151-2 narratives,
readers of, 21-2 as means of data collection, 158-61
Levi-Strauss, C . , 25, 72, 108, 123 ethnographic, 135-40
libretti , to be studied in action, 173 of death, 39, 65, 151
Lienhardt, G . , 138, 165 of disease, 102
life story, 97, see also autobiography, of events, 61, 64, 93, 95-6
biographical work of nature, 79, 161-2
literality, forms of, 30, 47, 95, 125, of science, 79, 125, 131
134, 139 negative cases in data analysis,
London, 63-4 148, 156-7
luminosity, 86 networks of communication and action,
Lynch, M . , 73-4, 128, 129, 131, 100, 130-1, 150-1, 171
143, 153 neuroimaging, 75-6
newspapers as a source of data,
Manuals, see ICD and DSM 114-16, 133-4
Malinowski, B., 134-7, 169 non-textual components of documents,
maps as documents in use, 166-8 5-10, 85-8
meaning, nosologies in action, 57
and reference, 113 types of, 33, 44, see also
as object of study, 23-4, 108, classification
109, 113 nursing records, see records
is manufactured, 112 NVivo, 141
is situated , 38, 109, 111
see also understanding Odysseus, 11
measurement, 44-5, 76, 132 Ong, W., 137, 171
membership categorization device opera as simile of text in action, 173
(MCD), 41, 56 organization as revealed by
Mental Health Inquiry Reports (MHI), documents, 56, 60-3, 81
61-3 outcome measures, 76, 130, 172
mental illness, 43-7 , 54-6, 59,
61-3, 93, 120 paintings as data, 7
metaphor, analysis of, 115 passports and identity, 20, 96
as ordering device, 71 Patient Information Leaflets (PIL), 101-2
methods of data analysis see data patient records, 54-6
M H I , see Mental Health Inquiry and identity, 102
Reports pedigrees, see genetics
'
mobil ization of documents, 13, 66, pencil-and-paper-diagrams, 74, 83, 86
86, 91 performance, 18-19, 28, 52, 63, 67,
modernism, 109, 111 73, 76, 92, 100-1, 103, 104,
motives, 39, 49 168, 173
multiple personality disorder (MPD), personal documents, 17, 21-2, 96
98-100 personality,
murder, see homicide disorders of, 56, 58, 62
museums, arrangements in, 88 multiple, 98-100

Copyrighted Material
194 I N DEX

pharmacists, 101-2 recontextualization, 62, 94, 105

photography, 135, 139 recruitment of documents, 58, 66, 67,
pictures, see paintings 100, see also mobilization
PIL, see Patient Information Leaflets referencing, 24, 113, 122, 130
Platt, J . , 97, 147 , 153 reflexivity, 51
police files, 30-1, 38 registers, see records
and statements, 94 reports,
pollution, careers of, 168
explanations for, 64-5 ethnographic, 126, 134-7
politics of, 63-5. governmental, 63-.-6, 163-4
Popper, K., 148, 164 health, 33, 44, 64-5
positivism, 127, 128, 148 international, 32
postmodernism, 109, see also mental health inquiry, 61-3
modernism police, 30, 38
power, expressed through scientific, 125-6
documentation , 47-9, 53, 67, 87 statistical, 31-46
print as writing, 28 rhetoric,
production of documents, 10-14, of ethnography, 137-41
30-48, 82-7, 129-32, 142, of medicine, 147
166-7 of science, 79, 80, 129-31
psychiatric measures, 43-6, 77 Ricoeur, P. , 28
and notes, 54, 62 risk, 152, 153, 155, see also
psychology and illness, 120 being-at-risk
psychologism, 109-13 Riviere, P. , 92-5
publ ications 130, 157, see also rule-governed procedures, 12-13,
citations 33-4, 41, 81-2
Putnam, H . , 113, 122 rule-books, see generative documents

quasi-object, 171 sampling of documents, 150-4

questionnaires, 44, 82 Saussure, F. , 12, 24, 123
questions; shape answers, 44-5, schizophrenia, 54-6, 100, 105
91, 99, 117 Schutz, A., 2-3, 111
science,
randomness and selection of data, and documentation, 73, 84-6, 128
149-51 , 153, 154 as a system of thought, 123,
readers, 127, 134
and meaning, 109-10 as mirror of nature, 126-9, 147
as performers, 18-19, 100 as social construction , 128-9,
shape writers, 16-18, 130 130, 144
reading, 16, 73, 89, 105, 130 rhetoric of, 79, 80, 129-31
reasons as social phenomena, 49 writing of, 129-31
records, 27, 29, 94 see also, positivism
cancer, 82 script,
clinical, 50-4 structures social action 53, 172
of identity, 94-5, 103 types of, 14, 52 , 53
nursing, 55-6, 62, 86-7 Seale, C., 116, 147, 157
organizational failure and, 62 selection of data, see randomness
psychiatric, 57 semiotics, 28, 123, 124
social work, 54, 62 sexual abuse, 98-9
in southern Sudan, 165 Shostakovich, D . , 17-18

Copyrighted Material
 I N DEX 195

signature, significance of, 12, 93, 105 translations,

Si lverman, D o , 123 of ideas into images, 73, 76, 168
Smith, Do, 28, 60, 67, 97 of talk into text, 54, 82-4
smog, see fog of text into talk, 53, 86
social roles and documents, 52-3, of text into text, 112
56, 101 transformations of data, 36, 156
sociological research and travellers' tales, 125, 135
documents, 3 Trobriand Islands, 134
Sorcerer's Apprentice, 3, 20, 29, 168
speech, and writing, 3, 28, 56, understanding as an aim of social
see also, talk research, 108, 110-13
Star, S o L , 57, 60, 71, 79, 86
. universities and documentation, 60
star maps, 167-8 use, as a focus for research, 4, 173
statistics, as social p�oducts, 30-49 users of documents, 4, 19, 42, 51,
structuralism, as school of thought, 58-9, 61, 66, 104, 147
25, 108, 123
subjectivity, 92, 95, 97-8 validity and reliabil ity in matters of
subjects, 90, 93, 94, 98, 100 research, 149
suicide, 34-40 van Maanen, J o, 127, 137, 139, 140
survey methods, 150-1 Vaughan, D o , 61, 69, 153
Sybil, 98-100 Verstehen , see understanding
systematic review as methodology, viruses, 88, 118-19
150, 155-6, 172 vitae, see curriculum vitae
vocabularies,
talk, 81-2, 86, 54-6, 59, 81-2, 86, of causation, 33, 65, 90
96, 98, 133, 158 of motive, 49, 90
tapes, see audiotape von Goethe, J oWo, 10, 27
taxonomies, see classification,
nosologies in action Weber, M o , 4, 23-4, 28, 110,
technology and documentation, 4, 47, 111, 165
81-6, 149-50, 171-2 Wittgenstein, L., 108
temperature charts, 57, 104, 128 Woolgar, S o , 74, 126, 128, 129,
text, 131-2
and documents, 5, 7, 26, 28, 128 World Health Organization (WHO), 32
and real ity construction, 138 World Wide Web (WWW ) , 2, 118
as distinct from talk, 28 writers, 16, 130, see also authors
as script, 14, 52-3 writing, 26, 57, 94, 95, 97-8,
in use, 122, 173 129, 131, 134, 166, 172
status of, 130 and print, 28
see also documents, and writing see also script
totemism, 25
transcripts, 158, 165 X-rays, 75, 104-

Copyrighted Material
Copyrighted Material