Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Digitizing Identities
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This book explores contemporary transformations of identities in a digitiz-

ing society across a range of domains of modern life. As digital technology
and ICTs have come to pervade virtually all aspects of modern societies, the
routine registration of personal data has increased exponentially, thus allow-
ing a proliferation of new ways of establishing who we are. Rather than rep-
resenting straightforward progress, however, these new practices generate
important moral and socio-political concerns. While access to and control
over personal data is at the heart of many contemporary strategic innova-
tions and domains as diverse as migration management, law enforcement,
crime and health crisis prevention, “e-governance,” internal and external
security, new business models, and marketing tools, we also see new forms
of exclusion, exploitation, and disadvantage emerging.

Irma van der Ploeg is Senior Research Fellow at UNU-MERIT at Maas-

tricht University Economic and Social Research Institute on Innovation and
Technology.

Jason Pridmore is an Assistant Professor in the Department of Media and

Communication at Erasmus University in the Netherlands.
 Routledge Studies in Science, Technology and Society
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

1 Science and the Media 8 Science Images and Popular

Alternative Routes in Scientific Images of Science
Communication Edited by Bernd Hüppauf and
Massimiano Bucchi Peter Weingart

2 Animals, Disease and Human 9 Wind Power and Power Politics

Society International Perspectives
Human–Animal Relations and the Edited by Peter A. Strachan,
Rise of Veterinary Medicine David Lal and David Toke
Joanna Swabe

3 Transnational Environmental 10 Global Public Health Vigilance

Policy Creating a World on Alert
The Ozone Layer Lorna Weir and Eric Mykhalovskiy
Reiner Grundmann
11 Rethinking Disability
4 Biology and Political Science Bodies, Senses, and Things
Robert H. Blank and Samuel M. Michael Schillmeier
Hines, Jr.
12 Biometrics
5 Technoculture and Critical Bodies, Technologies, Biopolitics
Theory Joseph Pugliese
In the Service of the Machine?
Simon Cooper 13 Wired and Mobilizing
Social Movements, New
6 Biomedicine as Culture
Technology, and Electoral Politics
Instrumental Practices,
Victoria Carty
Technoscientific Knowledge, and
New Modes of Life
Edited by Regula Valérie Burri 14 The Politics of Bioethics
and Joseph Dumit Alan Petersen

7 Journalism, Science and Society 15 The Culture of Science

Science Communication between How the Public Relates to Science
News and Public Relations Across the Globe
Edited by Martin W. Bauer and Edited by Martin W. Bauer, Rajesh
Massimiano Bucchi Shukla and Nick Allum
 16 Internet and Surveillance 24 Commodified Bodies
The Challenges of Web 2.0 and Organ Transplantation and
Social Media the Organ Trade
Edited by Christian Fuchs, Kees Oliver Decker
Boersma, Anders Albrechtslund
and Marisol Sandoval 25 Information Communication
Technology and Social
17 The Good Life in a Transformation
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Technological Age A Social and Historical

Edited by Philip Brey, Adam Perspective
Briggle and Edward Spence Hugh F. Cline
18 The Social Life of Nanotechnology
Edited by Barbara Herr Harthorn 26 Visualization in the Age
and John W. Mohr of Computerization
Edited by Annamaria Carusi,
19 Video Surveillance and Social Aud Sissel Hoel, Timothy
Control in a Comparative Webmoor and Steve Woolgar
Perspective
Edited by Fredrika Björklund and 27 The Leisure Commons
Ola Svenonius A Spatial History of Web 2.0
Payal Arora
20 The Digital Evolution of an
American Identity 28 Transparency and Surveillance
C. Waite as Sociotechnical Accountability
A House of Mirrors
21 Nuclear Disaster at Fukushima Edited by Deborah G. Johnson
Daiichi and Priscilla M. Regan
Social, Political and Environmental
Issues 29 The Fukushima Effect
Edited by Richard Hindmarsh A New Geopolitical Terrain
22 Internet and Emotions Edited by Richard Hindmarsh
Edited by Tova Benski and and Rebecca Priestley
Eran Fisher
30 Digitizing Identities
23 Critique, Social Media and the Doing Identity in a Networked
Information Society World
Edited by Christian Fuchs and Edited by Irma van der Ploeg
Marisol Sandoval and Jason Pridmore
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 Digitizing Identities
Doing Identity in a Networked World
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Edited by Irma van der Ploeg

and Jason Pridmore
 First published 2016
by Routledge
711 Third Avenue, New York, NY 10017
and by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2016 Taylor & Francis
The right of the editors to be identified as the authors of the editorial
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

material, and of the authors for their individual chapters, has been asserted
in accordance with sections 77 and 78 of the Copyright, Designs and
Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or
utilized in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in
any information storage or retrieval system, without permission in writing
from the publishers.
Trademark Notice: Product or corporate names may be trademarks or
registered trademarks, and are used only for identification and explanation
without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Names: Ploeg, Irma van der editor. | Pridmore, Jason, editor.
Title: Digitizing identities : doing identity in a networked world / edited
by Irma van der Ploeg and Jason Pridmore.
Description: New York : Routledge, 2016. | Series: Routledge studies in
science, technology and society ; 30 | Includes bibliographical
references and index.
Identifiers: LCCN 2015035987
Subjects: LCSH: Online identities. | Identity (Psychology) | Information
technology—Social aspects. | Identification.
Classification: LCC HM851.D5485 2016 | DDC 303.48/33—dc23
LC record available at http://lccn.loc.gov/2015035987
ISBN: 978-1-138-79463-4 (hbk)
ISBN: 978-1-315-75640-0 (ebk)

Typeset in Sabon
by Apex CoVantage, LLC
 Contents
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

List of Figures ix
List of Table xi
Acknowledgements xiii

Introduction: Digitizing Identities 1

IRMA VAN DER PLOEG AND JASON PRIDMORE

PART I
Sharing and Connecting: Friends and Consumers

1 The Touristic Practice of Performing Identity Online 21

ANDERS ALBRECHTSLUND AND ANNE-METTE ALBRECHTSLUND

2 A Social API for That: Market Devices and the Stabilisation

of Digital Identities 37
JASON PRIDMORE

3 Caring for the Virtual Self on Social Media: Managing

Visibility on Facebook 60
DANIEL TROTTIER

4 Shaping Children’s Consumer Identity Within Contemporary

Dutch Market(ing) Practices 77
ISOLDE SPRENKELS AND IRMA VAN DER PLOEG

PART II
Growing Up: Children and Guardians

5 Risk Identities: Constructing Actionable Problems in

Dutch Youth 103
KAROLINA LA FORS-OWCZYNIK AND GOVERT VALKENBURG
 viii Contents
6 Swimming in the Fishbowl: Young People, Identity, and
Surveillance in Networked Spaces 125
VALERIE STEEVES

7 Makers of Media Wisdom: Translating and Guarding

Media Wisdom in the Netherlands 140
ISOLDE SPRENKELS AND SALLY WYATT
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

PART III
(Mis)Behaving: Suspects and Deviants

8 Data Mining ‘Problem Youth’: Looking Closer But Not

Seeing Better 163
FRANCISCA GROMMÉ

9 Sorting (Out) Youth: Transformations in Police Practices

of Classification and (Social Media) Monitoring of
‘Youth Groups’ 184
VLAD NICULESCU-DINCA, IRMA VAN DER PLOEG, AND
TSJALLING SWIERSTRA

10 Identifying the Perpetrator: An Ethnographic Study

of CCTV in Police Work in Denmark 206
PETER LAURITSEN

PART IV
On the Move: Migrants and Travellers

11 The Digital Evacuee: Mediation, ‘Mobility Justice,’ and

the Politics of Evacuation 221
PETER ADEY AND PHILIP KIRBY

12 The Datafication of Mobility and Migration Management:

The Mediating State and Its Consequences 242
DENNIS BROEDERS AND HUUB DIJSTELBLOEM

13 Migrants at/as Risk: Identity Verification and Risk-Assessment

Technologies in the Netherlands 261
KAROLINA LA FORS-OWCZYNIK AND IRMA VAN DER PLOEG

Contributors 283
Index 289
 Figures
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

2.1 Data permissions for users 44

2.2 Data permissions regarding friends of users 44
2.3 Extended data permissions requested from users 45
2.4 Data permissions for Twitter 46
2.5 User permission request on Facebook from Oprah.com 47
2.6 User permission request on Facebook from Skype 48
2.7 User permission request on Twitter from The Economist 49
9.1 Evolution of problematic youth groups in the Netherlands
by number and type 189
11.1 Evacuation study sheet 229
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Table

12.1 International tourist arrivals 1995–2009 (in millions)

245
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 Acknowledgements
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This book results from several years of pleasant and fruitful cooperation
among many people. At its origin lies the work we conducted within the
DigIDeas project team, consisting of Vlad Niculescu-Dinca, Karolina La
Fors-Owczynik, Isolde Sprenkels, Jason Pridmore, and Irma van der Ploeg.
In addition, two international workshops, one held in Amsterdam in
November 2011 and one in Maastricht in June 2013, provided us with the
invaluable input of many much-appreciated colleagues, some of whose work
is presented in this volume alongside our own.
All this was made possible by a generous Starting Grant provided by the
European Research Council, and by the COST-Action Network Living in
Surveillance Societies, which we gratefully acknowledge.
Irma van der Ploeg
Jason Pridmore
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 Introduction
Digitizing Identities
Irma van der Ploeg and Jason Pridmore
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

DIGITIZATION OF IDENTITY

With digitisation and automation processes pervading virtually all aspects

and domains of society over the past two decades, the routine registration
of personal data has increased exponentially. When in numerous contexts
technologically mediated and automated interaction started to replace phys-
ical and face-to-face encounters, depriving the interacting partners of tradi-
tional, trusted ways of establishing to each other who they are, a strong need
for new identification practices emerged.
Securing interactions, controlling access to buildings or information, pro-
viding personalised services, assessing rights and entitlements, and tracking
movements of people, goods, and data all led to an increasing reliance on
information technology (IT)–based technologies for automated recognition
of individuals. Thus a wide range of systems and methods has been devel-
oped and implemented during the previous decennia, constituting a special-
ized field within IT and information and computer science called ‘identity
management (IDM),’ Some of these methods are based on something a per-
son knows, like a PIN code or password; another set relies on something a
person has in possession, like a smartcard or (electronically enhanced) docu-
ment, and a third set is based on inalienable traits like physical features, such
as fingerprints, iris patterns, or a facial image. Such front-end mechanisms
are connected to elaborate back-end systems and databases, connecting the
identifiers with personal files, track records, or enabling access to services.
All these new mechanisms thus involve registration and storage of personal
data in the ‘identity management’ part of the IT system, rendering individu-
als ‘known’ to the system and recognizable as such in subsequent interac-
tions and exchanges.
Today our lives are full of identification and authentication systems
like these; whether we do our online banking, book a flight, access the
parking lot at work, cross an international border, or apply for a visa, we
have to prove who we are—or more accurately, we have to prove that we
are who we claim to be, and therefore entitled to engage in that particular
(trans-)action.
 2 Irma van der Ploeg and Jason Pridmore
But the digitization of identities involves more than using information
and communication technologies (ICTs) for establishing who someone is.
Through the mere growth of digital communications, interactions, and
transactions, vast amounts of personal data are registered, because the tech-
nology is designed to log information on every action, thus generating and
storing details of people’s behaviours. When we use identifiers like person-
name-records (PNRs), biometrics, registration numbers, IP addresses, or
electronic product codes (EPCs), these data can be linked and connected
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

to track individuals or produce personal profiles unprecedented in detail,

availability, and durability. Such more elaborate profiles also constitute
digital identities, in the sense that they tell a story about who one is: past
choices, movements, preferences, connections, and personal history can be
assembled from a range of different domains to form a more or less detailed
‘picture.’ It is in this sense of digital identity that, as one authority in
IDM research reminded us, ‘digital identities always grow, never shrink’
(Pfitzmann, 2007), meaning that what is registered about a person (the num-
ber of attributes), in however dispersed systems, only increases over time;
once registered, such attributes are virtually impossible to erase. Thus, one’s
digital identity, or digital persona (Clarke, 1994), can cast long shadows
over one’s life (Blanchette and Johnson, 2002). Digital identities in this sense
can be retrieved from numerous databases that may be connected and sub-
jected to algorithmic searches and analyses to find correlations and patterns,
from which to define groups and categories. Thus, as much as a unique,
individual profile, such digital identities also serve to define and allocate one
to a group or category that can be targeted within specific risk-management,
marketing, or prevention-policy strategies. Digital identities are created and
maintained for their predictive capabilities. Law enforcement agencies look-
ing for suspects, health organizations and researchers looking for risk indi-
cators, and businesses wanting to ‘know’ their customers are among the
most eager to avail themselves of these data.
Clearly, access to and control over personal data is at the heart of many
contemporary strategic innovations, ranging from migration and border
management, law enforcement, crime and health prevention, ‘e-governance,’
and internal and external security to new business models and marketing
tools. The awareness that these developments come with great potential risks
and challenges to fundamental rights like privacy and non-discrimination,
however, has increased as well, as evidenced in an increasing number of
reports and opinions published by international authoritative bodies like
UNESCO, the OECD, the Article 29 Data Protection Working Party of the
EC, and the European Group on Ethics in Science and New Technologies.
It is also central to the work of a growing number of civil society and activ-
ist organizations concerned with ‘digital rights,’ privacy, and surveillance
issues, such as Bits of Freedom, Privacy International, and the Electronic
Frontier Foundation to name but a few examples. Further, continuing efforts
towards legal regulation are central to how digital identities can and perhaps
 Introduction: Digitizing Identities 3
‘should’ be done, something culminating in revised legislation on the Euro-
pean Union (EU) level regarding personal data.
Feeding into these increasingly widespread concerns is the immense pop-
ularity of social media, where people voluntarily present detailed textual
and visual accounts of their lives to mass audiences. Through personal pro-
files, blogs, uploading of pictures, and sharing experiences in online com-
munities, individuals voluntarily ‘share’ extensive accounts of their personal
and professional lives with chosen circles of ‘friends,’ but as often with less-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

defined audiences and little awareness of the potential for secondary use,
surveillance, and other adverse consequences to them. This too qualifies as
construction of digital identities, in the sense that it often involves a careful
and conscious effort to experiment with and shape one’s image to oneself
and others. These technologically enabled new practices have received a lot
of scholarly attention as well, starting with the first seminal studies of still
text-based online social communities like ‘MUDs (multi-user dungeon)’ end
‘MOOs (MUD-object oriented)’ in the 1990s, up to the numerous analyses
of today’s nearly global and ubiquitous online presence of hundreds of mil-
lions of people through platforms such as Facebook, Twitter, Instagram,
Pinterest, LinkedIn, and more.
Posing the question in what ways new information technologies and digi-
tal media mediate and transform identities in the information society, this
book connects with the fact that ‘identity’ is also a key concept in contem-
porary social theory and in contemporary conceptualisations of the relation
between technology and society, ethics, and normativity. Thus, a field of
enquiry emerges at the crossroads of theoretical, technological, and societal
developments, representing opportunities for frontier research.

THEORIZING IDENTITY

Clearly, the phenomena we are describing here have taken on such enor-
mous proportions that they sometimes appear to be defining of our time
and age. Moreover, they seem to affect not only how, where, and when we
establish to others who we are but also how we think about ourselves and
others, how we know ourselves and others know us, how we form and sus-
tain relationships of various kinds, and how we act and enact our plans and
projects. In other words, this wide-ranging digitization touches upon our
very identities in numerous ways.
But which theoretical resources do we need to make sense of it all? Are
we really talking about one phenomenon here, or are there significant dif-
ferences that the much-, if not over-used term ‘identity’ fails to differentiate
sufficiently? What do we mean when we talk about identities and identifica-
tion anyway?
In addition to the public ethical, legal, and political concerns about digital
identification, academic interest from the social sciences and humanities has
 4 Irma van der Ploeg and Jason Pridmore
been triggered by the sudden prominence of ‘identity’ in our highly tech-
nologized society as well. Of course, within these scholarly traditions, ‘iden-
tity’ has been a recurring, much-theorized theme predating the ‘information
revolution’ and our current mobile communication-technology-saturated
everyday-life by decades. Psychologists, social theorists and philosophers
have been working on the question what constitutes identity, how it comes
about, and why it matters for a long time and in many different ways.
Much commented upon already is the wide range of angles, disciplines,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and traditions that over many years have been brought to bear on under-
standing identity. Unsurprisingly, this now vast literature yielded not one but
a range of more and less diverging concepts of ‘identity.’ During the late 20th
century, for instance, it became central to various strands of social theory,
including post-colonial, feminist, and queer theory, where it got deeply con-
nected with social politics. Next to these strands of social theory, it figures
as a key term in psychology in various ways, including psychoanalysis and
social and developmental psychology. Philosophers, in their attempts to
bring more analytical clarity to the concept, have, paradoxically, contrib-
uted to the proliferation of different understandings of ‘identity’ as well.
All this yields a vast range of different accounts of what ‘identity’ consists
of, how it comes about and is known, and where and to whom it matters.
Over the past decade, the coming together of the digitization of society and
the prominence of issues of identity has opened up a new space for research
on ‘digital identities’ in all the fields mentioned, giving rise to a surge of new,
often interdisciplinary scholarship that builds on these older traditions in
various ways.
We will refrain from attempting to provide a complete overview of the
many approaches to ‘identity’ one finds here but, instead, briefly explain
the ones that played a key role in the research that led to this book. We will
thus briefly explain what ‘identity’ looks like from the respective ‘schools of
thought’ associated with identity management discourse, sociology, surveil-
lance studies, and science and technology studies. But before that, and in
order to better be able to characterize these different takes on identity, we
will highlight three key distinctions we found helpful in understanding these
various approaches and the differences among them.

THREE DISTINCTIONS

Overall, there are three key distinctions that account for many of the differ-
ences one finds among the various notions of ‘identity.’
The first is the distinction between ‘identity’ in a narrow sense and a
broader understanding of it; thus in many contexts, ‘identity’ simply refers to
someone’s name and a few other attributes such as age, gender, and nation-
ality. Knowing someone’s identity can then be achieved by simply taking a
look at that person’s passport, for example. This is the concept of identity
 Introduction: Digitizing Identities 5
one finds in law enforcement practices and national ID-card schemes, for
example. In contrast to this stand a notion of identity that encompasses a
much wider range of other elements, including, for example, one’s tastes
and preferences, one’s belief system, one’s lifestyle, profession or position in
society, one’s life story, up to one’s deepest dreams and desires. Getting to
know a person’s identity in this sense involves a lot more then mere checking
of identity papers; in fact, there is hardly a limit to what can be considered
relevant to this notion of identity.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This distinction resonates with a key philosophical distinction one finds,

for example in the work of Schechtman (Schechtman, 1990), who explains
how philosophers, when analysing ‘identity,’ have tended to be occupied
with the question of re-identification, which involves ‘spelling out the neces-
sary and sufficient conditions for saying that a person at t1 is the same as a
person at t2’ (Schechtman, 1990, p. 79), resulting in in criteria of personal
identity over time. This she opposes to identity ‘self-knowledge,’ which
involves reference to the beliefs, values and desires that are ‘expressive of
who one really is.’ Thus, whereas the first concept provides an answer to the
question ‘What makes a person the same as herself through time and space?’
the second answers ‘What makes a person unique and different from oth-
ers?’ A similar distinction can be found in the work of Ricoeur, who articu-
lates it as the distinction between ‘idem’ and ‘ipse’ identity (Ricoeur, 1990).
It can be argued that most of the time when ‘identity,’ ‘identification,’ and
in particular ‘verification’ are mentioned in relation to digital systems, it is
the first, narrow conception, the re-identification, or the ‘idem identity’ that
is concerned: usually the aim of identification systems is merely to establish
that someone is the same person that first applied for the service, created
the account, got issued the new passport, or got allocated some benefit.
Moreover, it is the second concept, the ‘ipse’ or ‘self-knowledge’ variety of
‘identity,’ that is probably the socio-politically and morally far more sig-
nificant one, because it pertains to this idea of who one ‘really’ is in terms
of highly personal characteristics, choices, and values. It may therefore be
important to look closely at which of the two is concerned in any specific
case to make sure what exactly is at stake. However, this does not mean that
they are completely unconnected. To take the example of someone’s identity
being verified in the context of social benefit payment: While on one level
this constitutes merely establishing that this person is the same one that was
enrolled in the programme earlier, at the same time, the very act of verify-
ing their identity does establish them as ‘social benefit recipient,’ which, on
another level, is a significant socio-political and morally charged identity
aspect in the wider sense.
The second distinction is related to this but is different, as it relates not
so much to ‘content’ of identity but rather to the perspective from which
it is conceived. This concerns the difference between identity as something
attributed to someone by others and identity as some kind of private self-
understanding. These two notions are sometimes played out against each
 6 Irma van der Ploeg and Jason Pridmore
other in the sense that the validity of the one may be denied by referring to
the other. This occurs in particular in relation to the act of identification:
Here the distinction between the first- and the third-person perspective may
cover a world of difference. For instance, a person’s public image may be
deeply contrasting with their own sense of their ‘true self.’
A lot of the early work on ‘digital identities’ in relation to individuals’
appropriation of personal computers (PCs), the Internet, and participation
in online communities, such as Turkle’s (Turkle, 1984, 1995), is character-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ized by an emphasis on opportunities to experiment, play around anony-

mously, or carefully shape and manage one’s online personal ‘profile,’ so
that concerned mainly this first-person notion of identity as one’s own
self-understanding. In contrast, those working towards what later became
known as ‘surveillance studies’ concerned themselves above all with the
third-person perspective involved in the type of identification involved in
personal data gathering and monitoring of people by others (e.g. Gandy,
1993; Lyon, 2003).
But in most social-theoretical approaches that include some form of this
distinction, a dependency between the two is assumed, in the sense that an
individual’s private sense of self and identity is not thought to be indepen-
dent of ‘the social’ or vice versa. Usually, both are seen as somehow mutually
influencing each other, with the theory providing ideas about the mechanics
of this connection.
The third distinction we want to highlight here is that between concep-
tions of identity that see it as something that is essentially ‘there’ and that
can be known, expressed, registered, proven, faked, and so on versus a con-
ception that sees it as the outcome of processes, practices, social interac-
tions, on-going construction, enactment, or performance. This distinction
cuts through the other two, in that the essentialist version can come in the
narrow and the broad variety and in the attributed-by-others as well as the
first-person form. Similarly, the other version—let’s call that the constructed
one—can be perceived as personal endeavour, mainly done from a first per-
son perspective, but also as that which results from others’ actions, attribu-
tions, and classifications; it can be restricted to the narrow version or include
the wider variety. Thus, to take the most counter-intuitive case: ‘Identifying
myself’ in the narrow sense, as this person with a particular name, gender,
date of birth and nationality, may be seen as merely stating a basic, objec-
tive fact about myself. But it can also be shown to be a result of a series of
actions and a set of historically and culturally contingent practices involv-
ing naming, birth registration, kinship definitions, producing ID documents,
and so forth.
Most concepts of identity one finds in the literatures and discussions
on digital identity can be characterized by the way they are positioned
with regard to these three distinctions. The tricky thing, and the reason to
elaborate these distinctions here, is that more often than not, it remains
quite implicit what variety of ‘identity’ is concerned when digital identities
 Introduction: Digitizing Identities 7
are discussed. Consequently, a lot of confusion, paradoxical claims, mis-
communication, and controversy occur, causing some to argue that the
concept has become analytically useless (e.g. Brubaker and Cooper, 2000;
Gutwirth, 2009).
However one feels about that assessment, it can hardly be denied that in
one form or another, ‘identity’ remains very much central to a wide range
of socio-politically controversial and ethically sensitive discourses and prac-
tices today. In particular, if one wants to make sense of the way digital tech-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

nologies have been and are involved in these, ‘identity’ and ‘identification’
often are prominently there, whether one likes it or not.
In this volume, we have gathered a series of contributions that approach
the topic of digital identities in a number of different ways. Within this
diversity, and broadly speaking, four key discursive practices can be found
as informing the notion of identity in each of these; in what follows we will
briefly characterize these.

IDENTITY MANAGEMENT

The first concerns identity as it is encountered in digital practices of identi-

fication. In particular, the growth of ‘e-commerce’ and ‘e-government’ ren-
dered the anonymity of interacting partners as a problem that needed to be
solved to enable the establishment of trust relations required for economic
exchange over the net and governmental services to assess entitlements. An
entire new industry surrounding digital identity, identification, and ‘identity
management’ started booming by developing all kinds of means to ‘identify’
people when interacting with and through an information system.
This industry generated a discursive and technological practice that
dealt with ‘identity,’ in the first instance specifically in the narrow sense:
names, connected to home or IP addresses, Social Security numbers and
bank accounts, were what mattered here. Moreover, verifiable identities
were what mattered, thus leading to the very opposite of the celebrated ano-
nymity and freedom to playfully experiment briefly enjoyed by MUD and
MOO enthusiasts during the 1990s. Moreover, as ‘certified identity,’ this
typically concerns identity and identification from a third-person perspective
and of the pregiven, essentialist variety. Elaborate infrastructures involv-
ing usernames, account numbers, PIN codes, tokens, smartcards, security
mechanisms, and biometrics have been developed over the past decades to
ascertain whether someone really is the person they claim to be and, there-
fore entitled to access bank accounts or government or social services or buy
and sell online.
It is in these identity-management practices that the notion of identity men-
tioned briefly at the beginning of this introduction and exemplified by the
quote from Pfitzmann originates. Here, ‘an identity’ has come to denote a
specific dataset pertaining to an individual user of a particular IT system, like
 8 Irma van der Ploeg and Jason Pridmore
an account number connected to a name, possibly including address, date of
birth, a biometric, a picture, or an ID number, and so on. This kind of ‘iden-
tity’ looks like a thing, a dataset, a commodity, a file, an ‘account’ in an IT sys-
tem; it can be exchanged, moved about, sold, stolen, be provided or revoked
by an ‘identity provider,’ and must be ‘managed’ by the owner of the system.
Clearly, in many respects, this notion is very much at odds with the way
philosophers, social scientists, and other theorists have, through long and
respectable intellectual traditions, come to talk and think about identity.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Nonetheless, through the prominence of ‘identity management,’ whenever

people interact with systems, authorities, or services through digital systems,
this concept of identity has tended to spread elsewhere. Thus, for exam-
ple, governments have come to perceive themselves as ‘identity providers’
because of their traditional role in issuing passports and other state-certified
identity papers (Barnard-Wills, 2012).
But it is also in more traditional practices like policing and law enforce-
ment or migration and border management that these identity-management
systems and identification practices are prominent. Identifying a suspect or
checking passports at border-crossing points is also premised on this notion
of identity, as it increasingly involves the use of IT–based systems developed
within this IDM frame.
But the very spread of this particular, if not peculiar, concept of ‘identity’
is a reason for caution for academics. In the context of a collection of stud-
ies of ‘digitisation of identity,’ this concept of identity is the object of study
rather than a theoretical, interpretative resource as such. For that, the next
three identity discourses are the ones most prominent in this collection.

SOCIOLOGICAL THEORY

Sociological theory has long been concerned with the production of identity,
albeit alternatively discussed as the production of ‘selves’ or ‘subjectivities.’
This is, as indicated in our second distinction, largely done in relation to how
personal identity is understood in relation to the attribution of identity by
others. For instance, Charles Horton Cooley (1902) articulated the ‘looking-
glass self’ as an internalization of others’ perspectives—or at least the per-
sonal interpretation of what those perspectives might be. This was further
reflected in later symbolic interactionist thought, most notably of George
Herbert Mead (1932) and Herbert Blumer (1969), who saw identity—again,
in their case, a notion of self-understanding—as largely a reflective process.
Most importantly, their work began to describe how identity could be seen
as part of a process of meaning making that occurs between individual per-
sons and others in the course of day-to-day social life. The reflective pro-
cesses that this is seen to inherently entail was further developed by Anthony
Giddens in his description of ‘late modern’ identities (1991). For Giddens,
identity is produced through a process of continual self-monitoring in light
 Introduction: Digitizing Identities 9
of and in contrast with social structures and practices. In his perspective,
identity is a continual process of reflexivity—though we may retain cer-
tain identifying attributes, sociological theory largely reflects the narrative
production of identity through a broad definition connected to life stories,
beliefs, and personal histories.
This highlights the contrasts between self-knowledge and the external
attribution of identity by others; however, each of these perspectives is predi-
cated on the idea that identity is a product of interactions between persons
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and social groups. Erving Goffman’s work drew inspiration from symbolic
interactionism, but his dramaturgical focus on social contexts was one of the
most prominent early works to discuss how identities are performed (1959).
By using notions such as ‘face’ and an understanding of selves produced
through ‘scenes,’ Goffman’s flexible conceptualization of identity gained sig-
nificant interest as more essentialized notions of identity became less appeal-
ing in light of more complicated social and political dynamics of the 1960s
and 1970s. Of course, concerns about the performativity of identity were
extended into later work as well, most notably that of Judith Butler. Butler’s
understanding of identity is clearly focused on how performance processes
work together to produce an identity (1990). However, in contrast to Goff-
man, there is no sense of inner desires or propensities to act. Butler describes
(gendered) identities as produced over time through repeated performances.
It is an illusion of identity that is produced in a continual acquiescence to
the workings of power.
Butler’s work is in line with and draws on Foucauldian perspectives on
the production of subjectivity. Both Foucault’s understanding of disciplin-
ary power (1979), the focus of his earlier work, and the focus of his later
work on confessionary technologies of the self (1988) indicate that identity
is largely a by-product of the workings of power upon ‘docile bodies’ in the
first instance and self-reaffirming relationships to power through personal
examination in the second.
Though some sociological theory pays attention to materiality—such as
Goffman’s references to ‘props’ and Foucault and Butler’s concerns with
performances on and of the body—in large measure, these descriptions are
primarily discursive in their focus. The focus in this collection on digital
technologies, however, requires an account of both the material and the
discursive. Though some sociological theory points us in this direction, the
incorporation of materiality in understandings of identity and identification
is more explicit and better accounted for in the next theoretical approaches.

SURVEILLANCE STUDIES

Another key discourse on identity informing the chapters in this volume is

that of surveillance studies, an interdisciplinary (though heavily inclined
towards sociology) field of research focussing explicitly on information
 10 Irma van der Ploeg and Jason Pridmore
technologies, in particular on practices of digital monitoring of people in
the broadest sense of these terms. Studies in this field, with topics ranging
from police using closed-circuit television (CCTV) in public spaces, man-
agers monitoring workers in call centers, health organizations collecting
patient data, to big data analysis of consumer data by marketers, and much
more, are deeply concerned with the societal, socio-political, and theoreti-
cal meaning, impact, and significance of such practices of ‘surveillance.’
Taking their cue from the work of, amongst others, Michel Foucault, these
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

studies have tended to emphasise the way in which ICTs are contributing
to phenomena like social sorting, automated discrimination, categorical
suspicion, and panoptic surveillance (Norris and Armstrong, 1999; Lyon,
2003; Monahan, 2006).
The problem of identity is one of the central issues of surveillance studies
for at least two reasons. One is that many surveillance practices come in the
guise of identification practices and technologies, such as described above,
thus putting the notion of identity centre stage empirically. Another is that
a key insight from its Foucauldian discourse-theoretical roots is that sur-
veillance, through its intricate micropolitical power mechanisms, produces
subjectivities and identities.
How these two lines of thought and research are related, however, is not
a straightforward matter at all. The key terms, ‘surveillance’ and ‘identity,;
tend to have different meanings in each, as well as different levels of abstrac-
tion, different relations to contemporary technological developments, dif-
ferent practical articulations, and different socio-political significance. But
they are related in important and complex ways that are rendered even more
opaque by frequent conflations and shifts of meaning in both academic and
general public discourses on ‘surveillance.’
Nonetheless, with its fast-growing body of both theoretically and empiri-
cally rich and robust descriptions, the field of surveillance studies has
become increasingly influential in focusing academic research, policy, and
various publics on the issues, incidents, and inequalities that have arisen
from the proliferation of digital monitoring and registration practices and
technologies.

SCIENCE AND TECHNOLOGY STUDIES

In the conceptualizations of identity described so far, one aspect remains

rather elusive, and that is the role of technology. Because we are concerned
here with the digitization of identities, that is, the involvement of various
forms of information and communication technologies in contemporary
transformations of identity, this is particularly problematic. Of course, digi-
tal technology, in the form of various IT–based identification systems and
online platforms, figures prominently in the approaches discussed so far, but
more often than not, these technologies are represented there as black-boxed
 Introduction: Digitizing Identities 11
tools, whose effects follow inevitably from their intended functions, without
much attention to the materialities and particularities of these systems them-
selves or how these play out in specific, localized practices.
The field of science and technology studies (STS), though not specifically
focussed on issues of identity, does offer an approach that enables inclu-
sion of these material technological, contingent, and local (f)actors in the
analysis. Deeply committed to an anti-essentialist and empirical construc-
tivist epistemology, STS approaches to digital identities and identification
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

technologies emphasize the need to study situated instances of implemented

technologies to make sense of their transformative agency.
In particular, approaches more or less loosely based on actor-network
theory (ANT) provide useful theoretical and methodological resources to
articulate the way various technical systems are involved in contemporary
transformations of identity. If one goes along with the notion that identities
are constructed or enacted rather than pregiven essences, and differently so
with the uptake of digital technologies, it makes obviously more sense to
take an approach that provides an account of how technologies are impli-
cated in this process.
Thus, in this field, ‘identity’ has been theorized as something that is pro-
duced or ‘performed’ in the very interaction among technology, persons/
bodies, discourses, institutional arrangements, and practices or, put differ-
ently, within hybrid socio-technical configurations (Law, 1991; Haraway,
1991; Latour, 1993; Mol, 2002).
ANT distinguishes itself from other approaches precisely by the agency it
grants to technology in these construction processes. Far from resorting to
technological determinism, it stresses the distributed nature of this agency
over the socio-material configurations, the ‘networks’ or ‘hybrid collec-
tives’ through which a range of different ‘actants’ jointly perform identities
(Akrich and Latour, 1992). In order to conceptualize and analyse the active
role of technology in this in a non-deterministic way, the notion of ‘script’
has been introduced, thus highlighting how technology enables and invites
certain actions and behaviours by ‘users’ without fully determining these
(Akrich, 1992). With the help of additional key concepts like ‘enrolment’
and ‘translation,’ this approach makes it possible to analyse the tranforma-
tive role of technologies in today’s digital identification practices.
Clearly, this approach comes with an elaborate vocabulary of its own,
which renders combinations with other approaches less than straightfor-
ward, but it has the huge advantage of providing the best theoretical and
methodological resource available today to account for technologies’ trans-
formative capabilities in an empirically sound, non-reductionist, and non-
determinist way.
Another highly relevant perspective developed within STS is the one
focussing on classification and its consequences (Bowker and Star, 1999).
One curious aspect of identification is that it usually is not just about deter-
mining unicity but also about assigning category. Thus, if we show you
 12 Irma van der Ploeg and Jason Pridmore
our passports, this not only establishes our unique personal identities but
simultaneously our nationalities, our genders, and so forth. In fact, the very
possibility of establishing a unique identity with a passport is based on this
passport’s status as official national document establishing the category of
citizenship. In similar vein, if a police detective identifies a suspect by match-
ing a latent fingerprint with a database-registered one, this simultaneously
shows this unique suspect to be a member of the class of registered criminals.
Thus, identity and classification are intrinsically connected, which makes
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

the analysis of categorization and classification schemes, the ‘scaffolding’ of

most IT systems and databases, highly pertinent to our topic.
This phenomenon has received attention from surveillance studies schol-
ars through the concept of ‘social sorting,’ but there it figures mostly as a
conclusion, where the analysis stops. The STS approach highlighted here
takes classification and its ‘social sorting’ effects as the starting point of
the analysis and subsequently aims to deepen understanding by unravelling
which contingencies went into the design of classification schemes and cat-
egorizing systems in the first place. Moreover, it emphasises that in order to
assess effects, one needs to study how classifications are appropriated and
used in specific contexts, where local, contingent factors shape the way cat-
egorization and classification play out on the lives and fates of those being
subjected to such systems.

ORGANIZATION OF THIS VOLUME

A common element in the contributions to this book is the empirical approach

to the topic: All chapters are based on empirical research of a domain or
practice in which digital identities and identification are, one way or another,
prominent. Moreover, while most contributions focus on some form of iden-
tification in the narrow sense empirically, the over-arching concern is how
new modes of ‘doing identity’ digitally affect identities in the wider sense,
as socio-politically significant and often ethically problematic phenomena.
Thus, throughout, an anti-essentialist perspective on identity is maintained
but one with the serious awareness that ‘if technology defines situations as
real, they are real in their consequences.’
We have clustered these contributions by the type of identification prac-
tice and the types of identities involved. Thus, Part I consists of a cluster of
chapters concerned with today’s wide take-up of social media and ‘sharing’
personal experiences with peers. This Sharing and Connecting section exam-
ines more fully the production of ‘friends’ and ‘consumers’ as identities of
particular interest.
Anders Albrechtslund and Anne-Mette Albrechtslund begin this section
by offering an exploratory and interpretive study analysing the activities on
social media in terms of ‘touristic practices.’ Their text suggests a direction
in which identity performances via online social networking practices are
 Introduction: Digitizing Identities 13
potentially translatable and useful in relation to practices of online sociality.
Touristic practices then become a way of thinking about and through the
expansion of digitized identities on social media.
Following this, Jason Pridmore examines the way in which identity attri-
butions and personal biographies become increasingly sharable. His chapter
focuses on application programming interfaces (APIs) that serve as the basis
for how particular sets of data move from one place to another—specifically
how applications (apps) rely on certain protocols and requests for personal
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

data. As these ‘market devices’ become increasingly prolific, understanding

of the potentials for and dangers and difficulties posed by digital identities
becomes increasingly clear.
Daniel Trottier then looks closely at the work and cultivation of iden-
tities on social media, looking specifically at experiences with Facebook.
His chapter discusses the uncertainty and tensions inherent in caring for a
‘digital self,’ combining issues of exposure with the proactive management
of a public self. The visibility of Facebook produces a convenient platform
to connect, but one rife with concerns about creeping and stalking. This
early look at Facebook adoption continues to resonate and helps frame the
navigation of purposeful identity development and its surveillance. Finally,
Isolde Sprenkels and Irma van der Ploeg look specifically at how children’s
digital identities are produced and enacted in the very process of identifying
these children. This chapter shows how the categorizations and character-
izations of children that online marketing surveillance practices produce are
simultaneously called upon to legitimate these practices. The empirical focus
of this chapter is on ‘advergaming,’ the uncomfortable amalgam of branded
online games blending childhood fun with commercial interests. This chap-
ter’s specific focus on children provides the transition to our next section.
Part II, Growing Up: Children and Guardians, centers around digital
practices involving children. Growing up in a digitized society, most children
today are highly involved in various new forms of online social interaction
and, as such, are often considered much better informed and skilled in using
these new forms of digital communication than are earlier generations. On
the other hand, however, they are also considered naive and vulnerable,
in need of care, surveillance, and protection. At the one end of this spec-
trum, children as users of digital media and systems are seen as the ‘avant
garde’ in whose practices and preferences the future of learning, participa-
tion, and sociability can be gleaned. At the other end, they appear as the
objects of increasingly elaborate, extensive registration and risk-assessment
systems, through which they are preformed at once as potential victims and
potential perpetrators, requiring pre-emptive identification and preventive
intervention.
The chapters in this section focus on this double-edged relationship of
children to digital systems and the identities this brings along. Based on
extensive interview data gathered in the context of the Media Smart’s Young
Canadians in a Wired World research project, Valerie Steeves tells a rich story
 14 Irma van der Ploeg and Jason Pridmore
about adolescents’ experiences with online visibility, surveillance, and their
strategies for ‘identity management.’ Taking issue with a one-dimensional
view on young peoples’ relation to social media and digital environments,
Steeves paints a colourful picture of the aware and complex nature of their
views and negotiations with peers as well as with parents and other authori-
ties to guard their ‘online persona’ and enjoy the freedoms this brings.
The contribution by Isolde Sprenkels and Sally Wyatt traces the emer-
gence and development of ‘media wisdom’ as a policy issue in Europe and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in the Netherlands. The chapter analyses this concept’s translation into a

range of initiatives in which the Internet and social media appear alternat-
ingly as opportunity and as threat to children’s development into healthy,
engaged, and educated citizens. Using the concept of ‘boundary object,’ the
chapter shows how ‘media wisdom’ has been appropriated by various actors
to eventually produce a whole industry of educational initiatives and courses
directed at children. The authors conclude that, even if these initiatives show
variation in the extent to which digital media are seen as threat or opportu-
nity, they tend have in common a particular distribution of responsibilities:
Through a common neglect of the agency and affordances of particular fea-
tures and designs of online media and the way these shape children’s chances
to understand and engage with them, it is the children who invariably get the
task of learning, adapting, and becoming ‘media wiser.’
The chapter by Karolina La Fors-Owczynik and Govert Valkenburg looks
at children’s identities from the opposite direction, that is, at systems in
which they figure not as users but as the objects of data gathering, elabo-
rate registration, and risk assessment. Starting with systems recently imple-
mented in Dutch youth care, the chapter then traces information-sharing
practices and connections to other actors involved in risk identification in
relation to children, such as police departments dealing with juvenile delin-
quency. It argues that, with the proliferation, sharing, and communication
of ‘risk indicators’ and ‘profiles’ and the emphasis on prevention and pre-
emption, risk identities of children come to include ever more aspects and
extend further and further in time and (social) space. In the process, this
renders being seen at risk interchangeable with being seen as risk and these
risk identities increasingly impossible to ever grow free from.
Part III focuses on digital practices connected with potentials for crimi-
nality or deviant behaviour. This section on (Mis)Behaving: Suspects and
Deviants sees the application of digital technologies in the on-going process
of identifying persons or practices deemed subversive or problematic. It is
not surprising then that a fair bit of focus both in practice and in these col-
lected chapters examines children and youth as a key concern. They are of
a particularly intensive involvement in today’s digitization of identification,
and the first two chapters in this section are both on ‘problem youth.’
Francisca Grommé’s description of Burgcity and their pilot study into
data mining examines in context the processes and problems of finding
problem youth through the combination of various data sources. She looks
 Introduction: Digitizing Identities 15
specifically at the process of focusing or ‘zooming in’ on these problem
youth that, as a metaphor, remains problematic. She argues that this meta-
phor both reinforces the need for ever more data collection and maintains
embedded and problematic normativities inherent in data-mining practices.
What emerges clearly from her analysis is how the digital practice of ‘zoom-
ing in’ requires ‘situated improvisations’ to become meaningful, resulting
in the creation rather than (closer) observation of that object of concern,
‘problem youth.’
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Vlad Niculescu-Dinca, Irma van der Ploeg, and Tsjalling Swierstra’s chap-
ter then continues this focus on normative issues by looking at technologi-
cal developments and policy transformations in contemporary policing. For
years now, the Dutch police have prioritized the issue of ‘problem youth,’
and this chapter examines how this increasingly has come to involve moni-
toring and ‘mapping’ of youth groups of concern with geographical infor-
mation systems and social media monitoring technologies. The authors
show how these technological practices play a constitutive role in the ways
in which youth groups are performed as ‘problematic’ and even in the defini-
tion of what counts as a ‘group.’ The authors note that in this context, the
police use the notion of ‘crime displacement’ to explain why they see fewer
‘problematic groups’ on the streets while justifying more intensive monitor-
ing and data collection; thus these practices are seen to perform new forms
of ‘suspects’ and ‘suspicious behaviours.’
Shifting away from the issue of problematic youth, Peter Lauritsen exam-
ines new practices involving the use of CCTV and the identification of
criminal perpetrators. Lauritsen looks at the idea that more crimes will be
solved at a low cost because of the cameras as a ‘folk theory’—something
that, although it lacks empirical support, still serves as an engine for future
action. In his chapter, Lauritsen further questions the effectiveness of CCTV
in police work by using the concept of ‘oligopticon’ to interpret the empiri-
cal findings, noting how the surveillance process envisioned by folk theories
of CCTV is always dependent upon limited and fragile networks. What
emerges from this study of CCTV use is that it does not provide a holistic
view of criminal activity or an effective prevention strategy but is rather
about establishing a vision that occasionally may help the police to identify
a suspect.
Each of these chapters looks at how digital technologies are used to
establish particular persons and groups of persons as suspicious and care-
fully analyses the less-than-straightforward way this happens. ‘Zooming in,’
‘crime displacement,’ and ‘folk theories’ are the concepts used to make sense
of these practices and to highlight their normative implications, given their
significance for police interventions and actions.
Finally, Part IV On the Move: Migrants and Travellers focuses on an area
of digital identification dense with global socio-political, legal, and ethi-
cal tension: (international) mobility. For all the many and diverse reasons
people may travel, be it as tourists, refugees, labor migrants, or displaced
 16 Irma van der Ploeg and Jason Pridmore
by natural disasters, governments are exerting themselves to keep track of
their movements. The desire to control cross-border mobility, to know who
enters the territory, and to separate the trustworthy from potential threats,
the ones needing protection and help from the ones requiring to be stopped
and apprehended, has engendered elaborate digital identification and moni-
toring infrastructures.
The chapter by Peter Adey and Philip Kirby starts with a special form
of mobility relatively neglected in the scholarly literature so far: evacuees
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

from natural disaster–stricken areas. After briefly outlining the historical

emergence of disaster-management systems in the Second World War, they
describe how the catastrophically failed emergency management during the
Hurricane Katrina disaster gave rise to a new generation of evacuation-
management systems, from which emerges ‘the digital evacuee.’ While the
authors document the disastrous results from failing to adequately track and
trace people on the move—an aspect receiving little attention from surveil-
lance scholars—they also point out how a highly abstracted representation
of evacuees within even the new systems fails to perform them as embodied
social beings, with highly problematic consequences.
Next, Dennis Broeders and Huub Dijstelbloem take on the more general
issue of international migration policy, specifically in the European context.
They observe how a range of technological developments, which they char-
acterise in terms of ‘datafication,’ have changed not merely the way policies
are executed but the process of policy making itself, including the nature of
the actors involved in that process, as well as the perception and definition of
the problems to be addressed. Using actor-network theory’s notion of ‘centers
of calculation,’ they address the question how the role of the state, as key
responsible actor in migration and mobility management, changes as a result.
Karolina La Fors-Owczynik and Irma van der Ploeg, finally, look into the
way these same systems play out locally when used by immigration, border
management, and other law enforcement authorities. Describing three spe-
cific risk-assessment systems migrants may come into contact with in the
Netherlands, the authors analyse how their design and operation affect the
needs and rights of those subjected to them. On the basis of these empirical
examples, they conclude that these systems contribute to an erosion of the
distinction between perceptions of migrants as potentially being at risk and
posing a risk, while an over-optimistic belief in these technologies’ function-
ality simultaneously exposes these already vulnerable groups to new types
of hazards.
The overall aims of this book are to increase understanding and aware-
ness of the more problematic aspects of digital identification practices and
to contribute to the quality of academic and policy debates about social
and ethical acceptability of these technological developments. We hope to
achieve these goals by bringing together the work of a set of exemplary
scholars, using insights gained from several disciplines to bear on contem-
porary developments in digital identification, thus offering novel ways to
 Introduction: Digitizing Identities 17
identify and articulate the issues concerned. With this series of chapters
reporting empirical studies on a range of domains in which digital identifi-
cation is at the centre of innovation efforts, the book offers state-of-the-art,
fine-grained knowledge of the ways digital technologies of identification are
implicated in contemporary transformations of identity.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

REFERENCES

AKRICH, M. (1992) The De-scription of Technical Objects. In Bijker, W. E. & Law,

J. (eds.) Shaping Technology/Building Society—Studies in Sociotechnical Change.
Cambridge (MA): The MIT Press, 205–224.
AKRICH, M. & LATOUR, B. (1992) A Summary of a Convenient Vocabulary for
the Semiotics of Human and Nonhuman Assemblies. In Bijker, W. E. & Law, J.
(eds.) Shaping Technology/Building Society—Studies in Sociotechnical Change.
Cambridge (MA): The MIT Press, 259–264.
BARNARD-WILLS, D. (2012) Surveillance and Identity. Discourse, Subjectivity,
and the State. Farnham: Ashgate.
BLUMER, H. (1969) Symbolic Interactionism: Perspective and Method. Englewood
Cliffs, NJ: Prentice-Hall.
BLANCHETTE, J.-F. & JOHNSON, D. G. (2002) Data Retention and the Pan-
optic Society: The Social Benefits of Forgetfulness. The Information Society,
18(1). p. 1–13.
BOWKER, G. C. & STAR, S. L. (1999) Sorting Things Out. Classification and Its
Consequences. Cambridge, MA, London: MIT Press.
BRUBAKER, R. & COOPER, F. (2000) Beyond Identity. Theory and Society,
29(2000). p. 1–47.
BUTLER, J. (1990) Gender Trouble: Feminism and the Subversion of Identity. New
York: Routledge.
CLARKE, R. (1994) The Digital Persona and Its Application to Data Surveillance.
The Information Society, 10(2). p. 77–92.
COOLEY, C. H. (1902) Human Nature and the Social Order. Scribner: New York.
FOUCAULT, M. (1979) Discipline and Punish: The Birth of the Prison. Harmond-
sworth: Penguin.
———. (1988) The History of Sexuality, Vol. 3: The Care of the Self. Vol. 3. New
York: Random House LLC.
GANDY, O. H. (1993) The Panoptic Sort: A Political Economy of Personal Informa-
tion. Boulder (CO): Westview Press.
GIDDENS, A. (1991). Modernity and Self-Identity: Self and Society in the Late
Modern Age. Stanford (CA): Stanford University Press.
GOFFMAN, E. (1959) The Presentation of Self in Everyday Life. New York: Anchor
Books.
GUTWIRTH, S. (2008) Beyond Identity? Identity in the Information Society, 1(1).
p. 123–133.
HARAWAY, D. J. (1991). Simians, Cyborgs, and Women: The Reinvention of Nature.
London: Free Association Books.
LYON, D. (Ed.) (2003). Surveillance as Social Sorting: Privacy, Risk, and Digital
Discrimination. London/New York: Routledge.
LATOUR, B. (1993) We Have Never Been Modern. Cambridge (MA): Harvard Uni-
versity Press.
LAW, J. (Ed.) (1991) A Sociology of Monsters: Power, Technology and the Modern
World. Oxford: Basil Blackwell.
 18 Irma van der Ploeg and Jason Pridmore
MEAD, G. H. (1932) Mind, Self and Society from the Standpoint of a Social Behav-
iorist. Chicago: University of Chicago Press.
MOL, A. (2002) The Body Multiple: Ontology in Medical Practice. Durham (NC):
Duke University Press.
MONAHAN, T. (Ed.) (2006) Surveillance and Security: Technological Politics and
Power in Everyday Life. New York: Routledge.
NORRIS, C. & ARMSTRONG, G. (1999) The Maximum Surveillance Society. The
Rise of CCTV. Oxford: Berg.
PFITZMANN, A. (2007) An Introduction to Digital Identity. Keynote lecture. OECD
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Workshop on Digital Identity Management. Trondheim, Norway, May 8–9.

RICOEUR, P. (1990) Time and Narrative. Vol. 1. Chicago, London: University of
Chicago Press.
SCHECHTMAN, M. (1990) Personhood and Personal Identity. Journal of Philoso-
phy, 87(2). p. 71–92.
TURKLE, S. (1984) The Second Self: Computers and the Human Spirit. New York:
Simon and Schuster.
TURKLE, S. (1995) Life on the Screen: Identity in the Age of the Internet. New York:
Simon and Schuster.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Part I

Friends and Consumers

Sharing and Connecting
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 1 The Touristic Practice of Performing
Identity Online
Anders Albrechtslund and Anne-Mette
Albrechtslund
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

Since the early days of the social Internet in the 1990s, scholars have been
concerned with how we perform identity and make sense of ourselves online.
Observing how communication in online settings is always dependent on
some form of symbolic mediation, usually text and images, it appears that
identity needs to be created and “composed”, as Sherry Turkle has put it (see
2002, 101; see also Albrechtslund, 2010). The idea of multiple, fragmented
identities which online personas permit as something useful and liberating,
undermining many traditional ideas about identity, connects to Donna Har-
away’s famous myth of the cyborg as a hybrid identity in a utopian reality
where people are “not afraid of permanently partial identities and contra-
dictory standpoints” (1991, 154).
However, although Turkle argued for the intimate connection between
what we do online and what we do in real life, she still represents a view on
the Internet, especially prevalent in early Internet studies, which separates
virtuality and reality, when writing, for example: “We can use the virtual
to reflect constructively on the real” (2002, 109). This tradition sees pos-
sibilities for expressing and exploring multiple identities in ‘cyberspace’, the
famous term coined by William Gibson in his novel Neuromancer (1984).
Cyberspace as metaphor designates a frontier-like abstract space, removed
from the real, physical world as we know it, an exotic, mythologized space
where users could be free of the limitations of the body, time, geographical
boundaries, and so on (Finnemann, 2005, 11). Such utopian perceptions
of cyberspace have been criticized from several angles as different Internet
researchers have taken a more nuanced view on cyberspace. For example,
Annette Markham has demonstrated an understanding of the many differ-
ent ways people perceive what it is to be online, identifying a continuum of
experiencing cyberspace as a tool to a place to simply a way of being (1998,
87). Following Markham’s distinctions, we might say that today, the Inter-
net has become a way of being, or perhaps even more accurate, the Internet
has become an integrated element in our way of being and cannot be isolated
in the form of ‘cyberspace’.
 22 A. Albrechtslund and A.-M. Albrechtslund
Even though communities and social networks are relatively new phe-
nomena in relation to the Internet, there already exists a large volume of
studies about activities, technologies, services, communication, and criti-
cal issues. These Internet studies are diverse in approach, scope, and focus.
However, most studies share an interest in concrete practices of Internet
users. Our current study connects to this interdisciplinary field, as it draws
on qualitative and analytical methods to investigate online social network-
ing practices. Here, we offer an exploratory and interpretive study focusing
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

on the typical activities related to the use of social media. By analysing these
activities in terms of touristic practices, we hope to contribute to the under-
standing and conception of identity performance via online social network-
ing practices.
In this chapter, we explore and analyse online sociality as a touristic prac-
tice to provide a new way of understanding how identity is produced and
performed in a network of diverse and dynamic actors. Online sociality
mainly unfolds within peer-to-peer social networking services, such as
Facebook, Twitter, and Foursquare, as well as Internet communities centred
on common interests, such as online games, music, or sports. There are at
least two reasons for applying this approach. First, a number of similarities
between practicing online sociality and practicing tourism can be readily
identified. These include both technologies such as the camera and practices
such as ‘postings’ (as in Facebook status updates and postcards, respec-
tively). Second, by exploring these similarities further, our ambition is to
emphasize certain dynamics in online social practices which are otherwise
not immediately recognizable. For instance, the tourist engages in a range
of activities (sightseeing, photographing, sending postcards, etc.) which
appear “in some sense unnecessary” but can be seen as part of a meaning-
ful exchange between everyday life and the vacation (Larsen, 2008; Urry
and Larsen, 2011, 1). Correspondingly, the user of social media typically
performs seemingly unnecessary activities (sharing mundane moments, pho-
tos, thoughts, and links) which nonetheless can be interpreted as a catalys-
ing practice producing meaning and identity (A. M. Albrechtslund, 2011).
This approach can help us better understand the popularity of social media
and the easy incorporation of such practices into everyday living in spite of
the many perceived dangers (e.g. privacy invasion, online predators) and
negative consequences (e.g. time waste, corporate exploitation; A. Albrecht-
slund, 2008). As such, our purpose is to situate social media practices and
the performance of digital identity in existing cultural practices by pointing
out similar dynamics.
In the following, we introduce an analytical framework by character-
izing a number of key features of the dynamics of touristic practices, draw-
ing on insights from the academic literature and observations of tourism in
Paris. The complementarity of online sociality and touristic practice then
forms the basis for our key analytical points. Finally, we use these points
to explore the production and performance of identity across different sites
 Performing Identity Online 23
of online sociality, and we conclude with a discussion of some of the issues
emerging from the results of our approach.

DYNAMICS OF TOURISTIC PRACTICES

The study of modern mass tourism includes a wide variety of theories, con-
cepts and empirical focal points, and in this section, we draw especially on
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tourism studies influenced by cultural and sociological theories, in particular

John Urry’s The Tourist Gaze (Urry and Larsen, 2011) as well as Roland
Barthes’s The Eiffel Tower (1983). We will elaborate on four selected phe-
nomena of special interest: the sight, the tourist, the camera, and the post-
card. These constitute key elements of the touristic experience, which are all
interrelated and very useful for building our analytical framework.

The Sight
Surfacing from the Trocadéro underground metro station in Paris’s 16th
arrondissement, the first thing which catches the eye is a colourful kiosk
selling tourist artefacts, pancakes, and ice cream, surrounded by people
with cameras and faces generally turned in one specific direction; it is
actually possible to see the tourist attraction prior to it appearing before
your eyes. After turning left around one wing of the Palais de Chaillot
onto the crowded paved platform, the Eiffel Tower, more than 500 meters
away, appears as a spectacular and overwhelming sight. A closer look
reveals that many of these people are in fact not looking at the tower but
have their backs turned towards it; they are posing in front of a photog-
rapher who is working to include both the person and the tower in an all-
encompassing picture. As one continues towards the attraction down the
stairs and through the Jardins de Trocadéro, passing more tourists and ice
cream stands, the tower grows before the eyes, and it becomes increasingly
difficult to capture it in a single gaze. Finally, the tourists arrive at the foot
of the Eiffel Tower.
The curious thing is that there is no real inside to visit, as Roland Barthes
concludes (1983). In fact, it might be said that the monument is “utterly use-
less” (238), as it does not serve any reasonable purpose other than to attract
tourists. This has been the case since its construction in the 1880s, when
Gustave Eiffel himself found it necessary to defend the contested usefulness
of the tower, including different kinds of scientific research and meteoro-
logical observations. However reasonable these claims are, “they seem quite
ridiculous alongside the overwhelming myth of the Tower, of the human
meaning which it has assumed throughout the world” (239). In other words,
the Eiffel tower and other sights have no useful function such as a road or an
airport, but this does not imply that the tourist sight is meaningless. The dis-
tinction between usefulness and meaning is important, because it anticipates
 24 A. Albrechtslund and A.-M. Albrechtslund
the point that the tourist attraction is more than just an object to see; rather,
the sight is a catalyst for a more complex tourist experience.
This corresponds to John Urry’s characterization of modern mass tour-
ism as essentially unnecessary (Urry and Larsen, 2011). Urry compares the
relation between everyday life and tourism with Michel Foucault’s study of
the relation between normality and illness in The Birth of the Clinic (1973),
where studying deviation at the same time provides a perspective on normal-
ity. In a similar way, the study of tourism—where people are separated from
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

their ordinary circumstances—gives insight into everyday life as well as soci-

ety, since tourism is at the same time in opposition to and a product of every-
day life. Accordingly, tourism is meaningful as a part of modern experience,
even though it is unnecessary, and tourist attractions are essentially useless.
“Then why do we visit the Eiffel Tower?,” Roland Barthes asks, and he
answers:

No doubt in order to participate in a dream of which it is (and this is its

originality) much more the crystallizer than the true object. The Tower
is not a usual spectacle; to enter the Tower, to scale it, to run around
its courses, is, in a manner both more elementary and more profound,
to accede to a view and to explore the interior of an object (though an
openwork one), to transform the touristic rite into an adventure of sight
and of the intelligence. (Barthes, 1983, 241)

Tourism thus encompasses a ritualistic practice in the sense that sight-

seeing can be a form of appropriation. In the case of the Eiffel Tower, this
attraction is an “obligatory monument” (247) for the tourist to see Paris.
As Barthes puts it, the tourist must perform an “initiational tribute” (248)
when visiting the city for the first time to be able to claim to have seen Paris.
There is a double meaning to ‘seeing Paris’ here, as the view from the tower
obviously provides a great panorama of the city; however, the ritualistic
practice of visiting the tourist attraction involves a translation into knowl-
edge on a more symbolic level of apprehending a certain meaning of Paris.
The Eiffel tower, one of the most-visited and famous monuments in the
world, can be understood as the archetype of a tourist attraction. What con-
stitutes this as a sight—irrational, useless, but certainly meaningful—is its
facilitation of the touristic experience, the tourist’s participation in a dream,
as Barthes puts it. Thus the sight is nothing without the sightseeing; the tour-
ist’s job is to perform it through a specific set of actions.

The Tourist
When you enter the Rue des Francs Bourgeois in Paris’s Marais district, it
feels like stepping onto a trail towards something. The intimate, quiet atmo-
sphere of the small streets of this neighbourhood are intersected by more
populated stretches connecting to the Place des Vosges, the oldest square in
 Performing Identity Online 25
the city, celebrated for its architectural beauty, its famous historic residents
and literary associations, including Alexandre Dumas’s three musketeers.
People yielding cameras and backpacks are immediately recognizable as
tourists, many studying maps and guidebooks. The tourist trail towards the
Place des Vosges is not an official, designated route to the tourist attraction.
However, the crowds seem to follow some invisible road signs that organize
the Marais into sections that are crowded and others almost left alone by
tourists.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The figure of the tourist is a well-known stereotype with a broad variation

of local flavours. For example the ‘American tourist,’ loud-speaking adults
wearing shorts, baseball caps, and Hawaiian shirts, or the ‘Japanese tourist,’
travelling in large groups, always snapping photographs. Although these
stereotypes of course rarely correspond directly to reality, tourists do tend
to stand out in street life due to certain conducts, appearances, and activi-
ties. Not to imply that any single feature immediately identifies a person as
a tourist; however, a number of little things often indicate something sepa-
rated from the mundane practices of locals. For example, a man entering a
Parisian bakery wearing his backpack on his front so as to avoid pickpockets
might display a certain unfamiliarity with his surroundings. Or a family
moving slowly, side by side, down the sidewalk of a crowded street, not
adapting to the polite, observant, and fast-paced rhythm of the Parisians.
This is, of course, not to say that tourists are unaware of appearing ‘tour-
istic’ and, essentially, not being local. Self-conscious and playful attitudes
towards the practice of sightseeing itself can often be observed to the point
at which it is possible to talk about a kind of post-tourism (Urry and Larsen,
2011, 13). This connects to another characteristic of the modern tourist
experience: that it can hold both the genuine and the tacky. On the one hand
the tourist’s ritualistic practice is a search for the experience of something
true or authentic, for example ‘the romance of Paris’; however, the very
same dream can just as well become part of the cliché of Paris, reproduced
in tourism advertisements. The authentic and the inauthentic can be sought
after or enjoyed by the tourist even as part of the same experience, and both
can be part of the production of meaning.
When tourists gather on the Place des Vosges or take in the view of the
Eiffel Tower from the Palais de Chaillot, they tend to do a number of things:
They either move slowly or linger in the same spot, they study guidebooks
and maps, interact with their travel companions, and display a blend of
focused seriousness and light-hearted playfulness. A noticeable activity is,
of course, the posing and photographing—or, to put it another way, the
searching for and selecting motifs and angles that adequately capture the
experience. In his book, Urry develops an understanding of the “tourist
gaze” as a particular way of looking at the world which translates certain
things into sights (Urry and Larsen, 2011). This gaze is not an individual
perspective but is a product of many working actors, including guidebooks,
others’ experiences, local presentation, cultural signs, and general social
 26 A. Albrechtslund and A.-M. Albrechtslund
patterning (4). Importantly, it is also a product of its relation to its opposite,
such as the mundane activities of everyday life, indicating that the tourist
gaze is always dependent on historical and societal circumstances.

The Camera
When one steps onto the Pont de la Tournelle towards the Ile Saint-Louis
between the Latin Quarter and the Marais district, a beautiful view of the
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Notre Dame cathedral appears. This is a spot where especially tourists stop
to take in the view. It seems as if the bridge has dynamic invisible fields into
which tourists and accidental passers-by try to avoid stepping. These fields
are constituted by the line between a photographer and one or more tour-
ists posing in front of the view towards the attraction. Interestingly, such
invisible fields seem to be generally respected as a private space, indicating
that this is an important moment not to be disrupted. The photographer,
the sight, and the poser(s) are all part of a network performing a ritual con-
sisting of three acts. First, the spectacle is composed, poser(s) are arranged,
and the right angle is found. Second, the photo is taken, mostly including
several clicks, and third, the photographer and poser(s) gather to look at
the picture(s) just taken, as if previewing the vacation photo album in the
making. However, when this ritual has been performed, the invisible field
disappears, and everything moves on again.
Modern mass tourism has been characterized as being a visual culture
(Urry and Larsen, 2011, 14). Thus, photography has always been a central
practice in the tourist experience, with the classic Kodak camera as iconic
device (West, 2000). Like any other technology, the tourist camera has
developed significantly through the decades; however, the change to digital
photography in the last few years has arguably been the most important.
At least, the third act of the ritual mentioned above is only made possible
by the viewing display of the digital camera, just as it generally facilitates
a wide range of sharing and distribution opportunities. Accordingly, the
advent of digital photography in tourism has further emphasized the social
aspect of taking pictures, making it a group activity. Interestingly, the recent
explosion in the use of different types of camera phones and smartphones
has apparently not yet eliminated the classic companion of the tourist, the
single-purpose camera around the neck, although the types of devices are
many, ranging from the semi-professional SLR models to the more compact
pocket versions. This choice of technology seems to be reflected in the prac-
tice of photography, as taking tourist pictures does not seem to have the
light, casual air to it as everyday smartphone use.
The tourist experience in all its phases is filled with pictures: studying
guidebooks, brochures and online resources when planning where to go or
researching the chosen destination. At the vacation spot, personalized pic-
tures of tourist attractions and leisure activities are taken. After returning
home to everyday life, the pictures are arranged and displayed in albums
 Performing Identity Online 27
(Larsen, 2005, 431). An obvious question to ask might be: Why take pictures
of tourist attractions that have been photographed countless times already
and are perhaps even on the cover of the brochure that launched the idea
of going in the first place? Susan Sontag (1977) argues that photography
is a way of grasping the world: “To photograph is to appropriate the thing
photographed. It means putting oneself into a certain relation to the world
that feels like knowledge—and, therefore, like power” (4). Taking pictures
of the Notre Dame cathedral is therefore rarely about discovering a novel
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

perspective or creating an original aesthetic work. Rather, the picture taking

is a necessary act of appropriation both as a more abstract knowledge/power
relation and as a concrete documentation of a moment for later retrospection.
The tourist photo is usually part of a collection of photos chronicling the
vacation, and this collection always has a potential audience. The photo
album thus serves as a point of departure for a narrative structured around
the events or sights which the selection of photos has emphasized as the
most important. As such, the individual photo is always part of a larger
network consisting of other pictures, memories, stories, other people, and
so forth. This also means that the tourist vacation is connected to everyday
life through pictures:

“Tourist photography and everyday life are not separate worlds but
bridges constantly traversed by photographing tourists on the move. It
is a form of photography intricately bound up with performing social
relations and picturing co-travelling ‘significant others’, which also
means that many otherwise ‘ordinary’ places are transformed into dra-
maturgical landscapes” (Larsen, 2008, 431).

The tourist album is the key element in this bridging between vacation
and everyday life, and, today, it comes in a variety of forms. As another
consequence of the change to digital photography, the photo album can be
shared with selected friends, family, or networks where a co-narration can
take place through comments, shares, and ‘likes’ (Facebook). Accordingly,
the social aspect of tourist photography has been brought to the foreground
by the emergence of online social networks.

The Postcard
Arriving at the Funiculaire Gare Haute, the upper station of the little tram
line that helps you get up the many stairs of the Montmartre, the Sacré-
Coeur Basilica appears on your right hand. The imposing church is one of
the major tourist sights in Paris, but, like the Eiffel Tower, it also attracts
tourists because it offers a great panoramic view of the city. After visiting
this site, many tourists pass along one side of the church and then turn left
on the Rue du Chevalier de la Barre, where a noticeable tourist trail towards
the Place du Tertre begins. This surrounding area is densely packed with
 28 A. Albrechtslund and A.-M. Albrechtslund
restaurants and shops selling tourist merchandise, collectibles, souvenirs,
and postcards. Most often, the postcards are displayed on stands outside
the shops and are, thus, a very eye-catching part of what tourists experi-
ence walking in these streets. Some stop to browse through these postcards,
which are mostly inexpensive and can be hand picked from the stands or
bought in preselected packages at a discount.
The postcard is not confined to being a phenomenon of mass tourism, as
it has been an important aspect of modern communication needs since the
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

early 20th century (Rogan, 2005, 3). A postcard is a single piece of paper or
thin cardboard meant to be written and mailed without an envelope. How-
ever, the touristic postcard most often features a glossy photograph of a sight
or location at the destination from which the tourist is mailing the card: for
example, at the Place du Tertre many postcards present a photo of the Sacré
Coeur Basilica. Other postcards feature images more symbolically connected
to the destination, for example by alluding to the city’s colourful past with
reproductions of famous Toulouse-Lautrec posters such as ‘Le chat noir.’
Yet another type of postcard can offer more playful motifs, such as a simple
black background with the generic text ‘Paris by night.’ The postcard is thus
obviously a part of the visual culture of mass tourism but is also distinguished
from photos, souvenirs, and so on by its emphasis on written communication.
As a communicative object, the touristic picture postcard is remarkable
for its blending of private and public spheres. It contains a private message
between a sender and a receiver familiar with each other, usually friends
or family, but the open format makes this message potentially readable for
anyone handling the card between the tourist and the recipient—for exam-
ple, mail workers, other household members, and so forth. Moreover, the
recipient conventionally displays the postcard in the home for guests to see
and even sometimes read, emphasizing the semi-public nature of this type of
communication. Like the tourist experience in general, the postcard is essen-
tially unnecessary. The genre does not invite an intimate correspondence
and, therefore, the content is often generic. However, sending a postcard is
a meaningful action for the tourist as well as the recipient by being mailed
from the vacation spot to friends or family, thus connecting the tourist to his
everyday life and maintaining this relation as a communicative performance.
This location-specific mailing of the postcard is what makes it meaningful,
as the act of conveying the message that the tourist is currently vacationing
and having experiences makes the communication phatic in nature (Jakob-
son, Pomorska, and Rudy, 1987; Rogan, 2005).

TOURISTIC PRACTICES AS ANALYTICAL FRAMEWORK

In the previous section we studied sights, tourists, cameras, and postcards and
found certain characteristics that prevailed throughout these dynamic tour-
istic practices. As Urry (and Larsen, 2011) has noted, tourism is essentially
 Performing Identity Online 29
unnecessary but maintains a complicated and meaningful relationship with
the ‘real’ life of the everyday (1). In this sense, the tourist vacation is not
so much seen as an escape from reality or from what really matters but as
something which is distinguished from but also created by everyday life. The
tourist sight is rarely in itself a meaningful object but rather has a catalysing
function similar to the ‘MacGuffin,’ a narrative plot device used especially
in thrillers and whodunits (Truffaut, Hitchcock, and Scott 1985, 138). In
such stories, this is usually an object desired and pursued by the protagonist;
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

however, the reason for its desirability is rarely explained in detail,1 but it
remains an obligatory object motivating the moving forward of the plot.
Like the Eiffel Tower in Barthes’s analysis, the tourist sight is often in itself a
‘useless’ monument, but it plays a central role as a catalyser for the ‘dream’
in which the tourist wants to participate, that is, the modern mythology or
popular story of Paris as the city of romance, art, and good living.
Just as the film spectator may be aware of the emptiness of the MacGuffin
while still enjoying the thrill of the plot, the tourist often displays a quite
self-conscious and sometimes even playful attitude towards sightseeing and
the sight itself. At the same time, the tourist makes sure to uphold the acts of
the touristic ritual, that is, taking pictures, lingering, and gazing at the view.
Indeed, the photographing activities at the tourist site are rarely connected
to aesthetic ambitions but are part of this catalysing process. Tourist pho-
tography, for instance, can function as a way of reaffirming significant social
relations, both in the act of taking pictures of each other at the destination
and of sharing them after having returned home (Larsen, 2005).
Another characteristic feature of touristic practices is that they serve as
bridging mechanisms between seemingly distinct spheres. The photo album
links the vacation experience with the context of everyday life by being
a medium both for own reminiscence and for communicating the experi-
ence to others. The photo album is thus not only an example of tourism as
constituting a primarily visual culture, it also highlights the social nature of
most touristic practices. Further, the bridging mechanism can also produce
a blending of public and private, as the semi-public nature of the postcard
illustrates. When tourists appropriate attractions by taking personalized pic-
tures, this practice is an example of something public becoming private by
the creation of ‘invisible fields’ between poser and photographer which are
respected as tiny private spheres in an otherwise very public space. These
photos are again made semi-public when the photo album communicates
the private experience to a broader audience through sharing with family,
friends, or a network.
These two mechanisms, catalysing and bridging, are not simply acciden-
tal side effects of tourism but are central to the dynamics of the touristic
practices we have studied and their production of meaning. In the following,
they are used as analytical concepts to explore the practice of performing
identity in online social settings with a special focus on posting, photo shar-
ing, and location check-ins.
 30 A. Albrechtslund and A.-M. Albrechtslund
PERFORMING DIGITAL IDENTITIES

Engaging in online social activity can be said to always involve posting as a

general activity which includes text communication, sharing of photos and
other images, and, as a newer development, indicating a specific location
at the time of posting. As such, posting something is the basis of online
communication on diverse platforms such as blogs, message boards, and
social network sites—the latter seeing a dramatic increase in user activity
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in recent years (Madden and Zickuhr, 2011). Focusing on the practice of

posting rather than a more specific social network site activity is therefore
productive for analysing the performance of online identity, as it offers a
way to maintain a general perspective on online communication. We struc-
ture the following study around the two central features of our analytical
framework, catalysing and bridging, as it is difficult to separate these diverse
types of online activities.

The Useless as Catalyst

In their influential article from 2007, boyd and Ellison developed a defini-
tion of social network sites which emphasizes a semi-public profile, a list
of connected users, and ability to browse and interact with this egocentric
network of users. The personal profile thus plays a central role in this defi-
nition, as it was typically the point of reference in the design of the site, but
today, it seems that the importance of the profile has diminished in favor of
the flow of live updates from users in the network (A. Albrechtslund, 2011).
This live content stream of ‘statuses,’ ‘tweets,’ or ‘check-ins’ is the de facto
interface to social network sites such as Facebook, Google+, Twitter, and
Foursquare.
In the section on tourist practices above, we discussed how the tourist
engages in a range of activities (sightseeing, photographing, sending post-
cards, etc.) that seems essentially unnecessary but is part of a meaningful
exchange between everyday life and the vacation. The tourist sight, as epito-
mized by the Eiffel Tower in Barthes’s essay, is characterized by its inutility,
something which seems to “escape reason” but is “something other and
something much more than” (1983, 238) the object in itself. The question of
why we visit these ‘useless’ monuments can be seen as addressing the same
consternation regarding the use of online social services, where the user of
social network sites typically performs seemingly unnecessary or superficial
activities (sharing mundane moments, photos, thoughts, and links). While
the type of content may vary across sites, postings are typically character-
ized by being compact, quick observations, descriptions, or opinions meant
for continual reading. Although longer texts are allowed and produced on
social network sites, Twitter’s famous 140-character limit sets a standard
for the genre of the status update. The content of the live stream, produced
by users posting to the network, is rarely profound or thought provoking in
 Performing Identity Online 31
itself. However, performing these activities can be interpreted as a catalysing
practice producing meaning and identity.
If we understand identity as something performed (Butler, 1990) or con-
tinuously narrated (Ricoeur, 1988), a person’s actions and statements in all
aspects of life, such as clothing, hairstyle, general appearance, and behav-
iour, all contribute to both self-understanding and outward personality. Per-
forming identity online is thus not to be seen as isolated from the general
identity work but rather as an integral part of it. However, online social net-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

working does offer a distinct way of connecting these identity performances

that, in turn, produces something new.
To read and contribute to the live content stream involves a set of prac-
tices that produce a certain way of ‘seeing’ and understanding. Like the tour-
ist gaze frames the experience of something as a sight to visit, online social
networking produces a certain ‘gaze’ which translates experiences into
something sharable which might attract ‘likes’ and comments. For example,
a parent sharing a baby picture with a cute caption is an active and selective
construction of a certain moment, which is different from a passive regis-
tration of the complete activities during a certain timespan. This is not to
say that such postings are necessarily a rosy presentation of a moment, for
example leaving out the baby screaming, diaper changes, and the like. The
capture is an actively constructed depiction of this moment, produced not
only by the individual perspective of the poster but rather as the result of
a co-creation involving, for example others’ comments, likes, the design of
the site, and general experience with the genre of (baby photo) posting. Just
as tourists can exhibit an ironic, self-aware attitude towards practicing the
tourist gaze (cf. the so-called post-tourist), it is entirely possible—and quite
common—to play around with the conventions of posting, but this aware-
ness is also an affirmation of the existence of a certain ‘gaze,’ or what could
be called “The Facebook Eye” (Jurgenson, 2012).
The social Internet space consists of diverse discourses and data. In the
context of social network sites, the live content stream is primarily2 con-
stituted by postings from a given user’s network of contacts, shaping the
communicative space in which the user participates. While other kinds of
social sites such as web forums and blogs are not usually customized to the
individual user in this way, they similarly provide a specific framework for
social interaction. It is a distinctive feature of social network sites that, for
example, baby photos, links to news articles, photos of silly cats, discussions
about sports, and location specifications blend together in a continual flow.
Meaning is produced across these various discourses as the practice of online
sociality unfolds. Postings from both private and professional life may form
part of an ongoing production of identity in which each contribution to the
content stream can have a catalysing function. The specific framework for
such a meaningful identity online can be described as a co-constructed way
of understanding and performing sociality (Larsen and Ryberg, 2011). In
other words, our social life is also experienced under the conditions of a
 32 A. Albrechtslund and A.-M. Albrechtslund
certain gaze resembling that of the tourist. This gaze is totalizing because it
assembles a variety of actors and circumstances into a composite vehicle for
continuous identity performance and narration.

Bridging the Gaps

Reminiscent of the way in which the tourist leaves behind everyday life and
goes on vacation, online socializing involves a range of conceptual divisions
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and dualities such as online/offline, private/public, real/virtual, professional/

personal—an issue which has remained a central matter in Internet research
for many years (Silver, 2000; boyd & Ellison, 2007; Baym, 2010). As men-
tioned, identity is produced through many different elements in the practice
of online sociality, as both posting updates, photos, and links and taking
part in different interactions constitute a continuous circulation of informa-
tion. Identity performance online thus always unfolds in relation to a hybrid
and transient totality. In the following, we will elaborate on this dynamic
and discuss how the individual engages in a constant interpretation and
negotiation, bridging the gaps experienced as a result of these perceived
divisions and dualities.
Seemingly distinct spheres are constantly bridged in the practice of online
sociality. The division between public and private spheres, for instance, is
at the heart of much discussion, confusion, and concern relating to Internet
use. The more or less extensive sharing of personal information on social
websites—activities, beliefs, preferences, whereabouts, family relations, and
so on—creates a hybrid space in which public and private life is continually
traversed and negotiated (Albrechtslund, 2008). Similarly, distinguishing
between online and offline domains often entails a complex process of sort-
ing out yet other divisions and dualities, including real/virtual, authentic/
inauthentic, and natural/technological. The separation of such concepts
often proves to be difficult when engaging with actual use, as Petersen (2007)
has demonstrated in his study of the interweaving of materiality and online
life, describing it as mundane cyborg practices in reference to Haraway
(1991). While engaging in online sociality does not involve departing from
the physical circumstances of everyday life as it does for the tourist on vaca-
tion, both practices share the persistent involvement with devices that main-
tain the social relations, context, and significance of everyday life and thus
bridges these gaps.
The central touristic device of the postcard embodies the blending of
spheres that are often perceived as separate, connecting vacation and every-
day life as well as public and private communication. Reminiscent of this
dynamic is the key practice of posting updates to online social sites. Seen
as a bridging activity, the posting of any kind of content to, for example,
Facebook’s News Feed is an act of performing identity within the context
of a network consisting of various social relations. Like the postcard, the
 Performing Identity Online 33
status update is rarely about profound content but has more to do with the
act of communicating and connecting in itself. Just as the postcard functions
as a device to embed the vacation experience in the context of everyday life,
the status update achieves meaning because it is part of a person’s reality as
a whole, not removed from it. Furthermore, the importance of mailing the
postcard on location, and, correspondingly, posting the thought, comment,
anecdote, and so on to the online network in a more or less immediate con-
nection to the given situation is a testament to the way the meaning created
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

by these acts results from a hybridity of physical and virtual space and of
present and absent actors.
The same dynamics can be said to be at work in the practice of online
photo sharing. This activity has become an increasingly significant element
of posting in recent years and is manifesting in various interesting and play-
ful ways, as the rise of services such as Instagram, Pinterest, and Snapchat
show. Sharing personal photos online is thus a key bridging activity and
an important catalyst for meaning production as well. The coalescence of
intimacy and exhibition is similar in the practices of tourist photography
and photo sharing, which are both characterized by transforming “configu-
rations of the fields of cultural production in the context of new media, for
which ‘art’, ‘folk’ and ‘popular’, as well as ‘artist’, ‘professional’ and ‘ama-
teur’ are inadequate” (Burgess, 2007, 29). The creation of photo albums is
at the same time a construction of a personal archive, driven by a sort of
“nostalgia for the present” (Jurgenson, 2011) in which meaning is produced
by “slicing out this moment and freezing it” (Sontag, 1977, 15), and a cen-
tral device in the social dynamics of identity performance, producing and
displaying social relations (Larsen, 2005).

CONCLUSION

It is important to stress that we do not claim any direct identification between

online social networking and tourism such as a superficial statement that
online social networking is ‘the same’ as tourism. Nor do we pretend that
users of social media are ‘everyday tourists’ or even that tourism and online
social networking can be reduced to a single type of practice carried out by
a specific type of actor. Here, tourism purely serves as an analytical frame-
work providing a point of departure that allows for a productive way of
thinking about the everyday practices of social media.
What we have set out to do in this chapter is to identify the dynamics
that are at play in the cultural practice of tourism in general and then to
show how these dynamics are translatable and useful in relation to practices
of online sociality. As these new practices emanate from and are embedded
in our traditional cultural habits, it is hardly surprising that we are able to
connect online sociality to a phenomenon such as mass tourism. Facebook,
 34 A. Albrechtslund and A.-M. Albrechtslund
Foursquare, and other social networking applications are new technologies
that let us connect to our social networks in new ways. However, as we have
shown, the dynamics driving these services look very familiar. There are a
number of immediately apparent similarities between tourism and social
media practices which inspired us to explore the connection further, namely
the interesting blending of public and private spheres and the complex role
of gratuitous activities in the mundane context of everyday life. Further
analysis led us to articulate two concepts, catalysing and bridging, as a way
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

to grasp how the seemingly unnecessary or even useless activities of online

socializing are integrated in a meaningful performance of digital identity. In
other words, when users of social media engage in what appears to be time-
wasting, gullible, or escapist activities, they are also actively taking part in a
meaningful enactment of identity in a network of social relations. This also
means that identity performance is not understood as a fabrication or stag-
ing of the self,3 where the ‘real’ or ‘authentic’ is somehow located behind the
mask of social conduct.
Another pertinent similarity when considering touristic practices and
online sociality is the political and economic issues; tourism is a major com-
mercial industry, and so are social media. As such, the experience of vaca-
tion as well as online social networking is obviously embedded in a larger
ideological framework which enables and shapes these practices. However,
these structural issues alone do not answer the question of why we engage in
these practices. Our study offers a way to understand the motivation behind
them and thus what makes the industries possible in the first place. A nar-
row focus on ideological issues can have the unfortunate consequence of
infantilizing the subject taking part in these activities and reducing the com-
plexity of the meaning-producing practice of performing digital identities to
an exploitative power relation. Rather, a closer look at these practices as an
experience which is meaningful for the individual opens up a trajectory of
thinking that provides new insights into the everyday life with the Internet.

NOTES

1. Famous examples include the ‘secret microfilm’ in Alfred Hitchcock’s North

by Northwest (1959), and the ‘Maltese Falcon’ in John Huston’s film (1941).
2. Other content includes ads, promoted posts, and site updates.
3. Only if understood in Goffmanian terms (1959). Our understanding of iden-
tity in this chapter corresponds to Goffman’s dramaturgical model of social
interaction, which describes identity as a performance with varying degrees of
social strictures and codes.

REFERENCES

ALBRECHTSLUND, A. (2008) Online Social Networking as Participatory Surveil-

lance. First Monday, 13(3).
 Performing Identity Online 35
ALBRECHTSLUND, A (2011) Socializing the City: Location Sharing and Online
Social Networking. In Fuchs, C., Boersma, K. & Sandoval, M. (eds.) Internet and
Surveillance The Challenges of Web 2.0 and Social Media. New York: Routledge,
187–197.
ALBRECHTSLUND, A. M. (2010) Gamers Telling Stories: Understanding Narrative
Practices in an Online Community. Convergence: The International Journal of
Research into New Media Technologies, 16(1) (February). p. 112–124.
ALBRECHTSLUND, A. M. (2011) Online Identity Crisis: Real ID on the World of
Warcraft Forums. First Monday, 16(7).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

BARTHES, R. (1983) A Barthes Reader. New York: Hill and Wang.

BAYM, N. K. (2010) Personal Connections in the Digital Age. Cambridge: Polity
Press.
BOYD, D. & ELLISON, N. (2007). Social Network Sites: Definition, History, and
Scholarship. Journal of Computer-Mediated Communication, 13(1). p. 210–230.
BURGESS, J. E. (2007) Vernacular Creativity and New Media. Brisbane: Queensland
University of Technology.
BUTLER, J. (1990) Gender Trouble: Feminism and the Subversion of Identity. New
York: Routledge.
FINNEMANN, N. O. (2005). Internettet i mediehistorisk perspektiv. Copenhagen:
Samfundslitteratur.
FOUCAULT, M. (1973) The Birth of the Clinic. London: Tavistock.
GOFFMAN, E. (1959) The Presentation of Self in Everyday Life. Garden City (NY):
Doubleday.
HARAWAY, D. (1991) A Cyborg Manifesto: Science, Technology, and Socialist-
feminism in the Late Twentieth Century. In Simians, Cyborgs and Women: The
Reinvention of Nature. New York: Routledge. p. 149–182.
HITCHCOCK, A. (1959) North by Northwest. MGM Pictures.
HUSTON, J. (1941) The Maltese Falcon. Warner Bros.
JAKOBSON, R., POMORSKA, K. & RUDY, S. (1987) Language in Literature.
Cambridge: Harvard University Press.
JURGENSON, N. (2011) The Faux-Vintage Photo. Cyborgology. Available from:
http://thesocietypages.org/cyborgology/2011/05/14/the-faux-vintage-photo-full-
essay-parts-i-ii-and-iii/. [Accessed 10–10–2012]
JURGENSON, N. (2012) The Facebook Eye. The Atlantic. Available from: www.
theatlantic.com/technology/archive/2012/01/the-facebook-eye/251377/.
[Accessed 12–10–12]
LARSEN, J. (2005) Families Seen Sightseeing: Performativity of Tourist Photogra-
phy. Space and Culture, 8(4). p. 416–434.
LARSEN, J. (2008) Practices and Flows of Digital Photography: An Ethnographic
Framework. Mobilities, 3(1). p. 141–160.
LARSEN, M. C., & RYBERG, T. (2011) Youth and Online Social Networking:
From Local Experiences to Public Discourses. In Dunkels, E. & Fraanberg, G. M.
(eds.) Youth Culture and Net Culture: Online Social Practices. Hershey (PA): IGI
Global, 17–40.
MADDEN, M. & ZICKUHR, K. (2011) 65% of Online Adults Use Social Net-
working Sites. PEW Internet & American Life Project. Washington (DC): PEW
Research Center.
MARKHAM, A. N. (1998) Life Online: Researching Real Experience in Virtual
Space. Lanham (MD): Rowman Altamira.
PETERSEN, S. M. (2007) Mundane Cyborg Practice: Material Aspects of Broad-
band Internet Use. Convergence: The International Journal of Research into New
Media Technologies, 13(1). p. 79–91.
RICOEUR, P. (1988) Time and Narrative. Vol. 3. Chicago: University of Chicago
Press.
 36 A. Albrechtslund and A.-M. Albrechtslund
ROGAN, B. (2005) An Entangled Object: The Picture Postcard as Souvenir and Col-
lectible, Exchange and Ritual Communication. Cultural Analysis, 4(1). p. 1–27.
SILVER, D. (2000) Looking Backwards, Looking Forward: Cyberculture Studies
1990–2000. In Gauntlett, D. (ed.) Web.studies: Rewiring Media Studies for the
Digital Age. Oxford: Oxford University Press. p. 19–30.
SONTAG, S. (1977) On Photography. New York: Macmillan.
TRUFFAUT, F., HITCHCOCK, A. & SCOTT, H. G. (1985) Hitchcock. New York:
Simon and Schuster.
TURKLE, S. (2002) Our Split Screens. Etnofoor, 15(1/2). p. 5–19.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

URRY, J. & LARSEN, J. (2011) The Tourist Gaze 3.0. London: Sage.
WEST, N. M. (2000) Kodak and the Lens of Nostalgia. Charlottesville: University
of Virginia Press.
 2 A Social API for That
Market Devices and the Stabilisation
of Digital Identities
Jason Pridmore
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

The mechanisms for transmitting and receiving data on the Internet have
shifted. While early protocols and the use of hypertext transport protocol
(http) remain in use, web browsers to mobile devices to operating systems
increasingly rely upon applications to function. This chapter focuses on
the process whereby data has been made accessible and usable by exter-
nal applications—that of application programming interfaces (APIs). At the
most basic level, APIs are the means by which different technologies are
enabled to communicate with each other. They have been in use for many
years, as they standardize the protocols for the information that can be
retrieved, used, and amended by applications from an external resource.
In a number of cases, and more importantly for the context of this book,
increasingly this resource comes in the form of social media data. This chap-
ter argues that these relatively obscure and unseen APIs, particularly those
connected with social media, have significant implications for the conceptu-
alization and development of digital ‘identities’ in our contemporary world.
As social media is infusing itself into numerous aspects and domains of
everyday life, these social media APIs are arguably becoming default arbitra-
tors and mediators of ‘identity.’
The increased popularity and availability of social media itself can in
part be attributed to publicly available APIs, or rather, the applications that
have been made to work on or through these interfaces (see Dix, 2011).
The relatively recent orientation to applications, or apps as they are better
known (popularized by a 2009 Apple iPhone advertisement with the phrase
“There’s an app for that”), have received significant attention from consum-
ers and the media. However, the APIs they are based upon have received
significantly less attention, despite the fact that an increasing number of
APIs are available for a divergent number of sites, services, and resources
(see Bodle, 2011). APIs are focused on potentials for interoperability and
include commonly used ‘tools’ such as Google search, resources such as
videos with YouTube or photos with Flickr, or social media like Twitter and
LinkedIn. All of these allow their content or data to be ‘embedded’ onto
 38 Jason Pridmore
an independent website or application through the use of their open API.
One of the first and longest-lasting API successes, the Google Map API,
became very popular as it allowed a number of different applications to
integrate Google maps as a resource for their own purposes (Dix, 2011).
More recently, Facebook’s success as a platform (at least in historic terms of
growing its user base and popularity over MySpace) has largely been attrib-
uted to its public API, particularly as it allowed for the implementation of
third-party applications and the more recent development of what they call
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

an ‘Open Graph’ protocol (Bodle, 2011; Jung and Lee, 2010; Dix, 2011).
The shift towards the development and use of social media APIs has hap-
pened more recently (concurrent with the growth of social media). The pre-
dominant use of APIs previously centred on developing tools and games
that relied on data separated to some degree from individual persons. These
include maps, images, and keywords. Yet the increased presence of social
media and their publicly available APIs have allowed for the development
of applications that can extract a variety of more personally identifiable
information. APIs—as market devices (see what follows)—provide access
to data in various forms, from more “transactional” data to “archives of
the everyday” (see Beer and Burrows, 2013). They give certain but easily
measurable and quantifiable sets of attributes, characteristics, practices, and
more of those persons engaged in social media. These are records connected
to a social media ‘profile,’ which can be seen as tantamount to a ‘bricolage’
of digital identity: status updates, relationships, friends, location, tagged
photos, searches, reposts, ‘likes,’ and more are the ‘materials at hand’ in the
intermingling of ‘identity’ production with digital technologies. This con-
nection between identity and technology is inclusive of the idea that at the
very least, identity is seen as some form of a fluctuating collection of attri-
butes (voluntarily) indicated to others and those characteristics attributed
to and expected of us by other agents (see Jenkins, 2008). Identity in this
form is seen as contingent and relational, marking a transition from rather
static notions of identity inherent in identity documents and identity provi-
sions. Instead, social media allow for both indications and attributions to
be collected and analysed in digital form, and it is APIs that have made this
increasingly accessible.
The argument of this chapter is that APIs are market devices that sta-
bilize the dynamic and constantly updated data of social media and serve
to connect a number of different actors. They allow applications to access
significant details about social media users and can act in a person’s place,
extending and amplifying their actions, relevance, and relationships in ways
previously unknown. They make the dynamic constitution of identity mea-
surable and quantifiable and continually updatable. Yet all the while these
APIs remain out of sight and are only hinted at within the confines of a
potentially singular encounter in which a social media participant grants
permission to an application to access their personal profile. The chapter
explores the relationship between the use of APIs and how this connects
 A Social API for That 39
with the personal production or performance of identity through applica-
tions. APIs are key components in the socio-technical configuration of per-
sons, social media databases, and applications that define and determine
the technological transfer of data in ways hidden from people that use these
various social media–‘powered’ applications. Though applications are most
often convenient, engaging, and entertaining tools or resources, the rela-
tionship between information available through APIs and the production of
contingent and relational digital identities needs to be made clearer.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The following sections in this chapter begin to do just this. First, this
chapter describes APIs as ‘market devices’ that bring together people, activi-
ties, and technologies. Second, the permission process on the side of both
developers and social media users of these market devices is shown to indi-
cate the extent of identifiable data that becomes accessible through a singu-
lar acceptance ‘click.’ Though there are a number of relevant social media
APIs, this section focuses specifically on the permission process by two of
the largest social media providers, those of Facebook and Twitter. Next,
given the availability of data from this process, the chapter highlights some
of the analytical convergence of technology, consumption, social media, and
identity in contemporary discussions. It notes that social media are impor-
tant platforms upon which both self-identified and externally identified cat-
egorisations are created, produced, and imposed. This information yields
a number of (partial) digital ‘identities’ that serve as representations for
particular persons in particular contexts. The chapter then concludes by
arguing that APIs are increasingly important and necessary components in
socio-technical configurations of identity. The stabilize access to continually
shifting and augmented digital ‘identities’ by allowing access to the dynamic
and rich mixture of data and associations of social media participants. These
can be seen as neither fully owned nor controlled by persons or a given sys-
tem. Instead, a digital identity—or perhaps digital identities—are themselves
contingent and relational and enacted in various social and technical ways
that are beyond the range of either human or technical agency.

CONSUMPTION AND MARKET DEVICES

Social media are objects of consumption, yet their use has significantly
shifted understandings of consumption towards cultural forms being seen as
processes of ‘prosumption.’ This concept is at the very least a reaction that
“the analytic distinction between consumption and production has often
been overdrawn by social scientists” (Beer and Burrows, 2013, 49). Histori-
cally, descriptions of material or cultural production and consumption have
reinforced rather mechanical and binary descriptions of markets—such as
depictions of ‘cultural dupes’ and ‘consumer choice.’ The former—in the
vein of Frankfurt School theorists—have focused attention on culture indus-
tries and the latter—in a more ‘post-modern’ vein—viewed consumption as
 40 Jason Pridmore
a means of producing personal identity. These have largely ignored the co-
constitution of consumption subjects and objects (see Miller, 1987) in their
focus either on corporate power (through advertising, for instance) or asser-
tions of consumer expression through the use of goods (in either anticipated
or unanticipated ways). These descriptions have often disregarded the fluid
and nuanced ways in which consumers and producers interact. Prosump-
tion, in part, begins to point out the overlap already present in many areas of
consumption and production, but most specifically it does so in connection
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

with social media (Ritzer and Jurgenson, 2010).

More importantly for this chapter, there has been very little recognition
of the role of technology in these ‘prosumption’ interactions. This chapter
draws from an actor-network theory approach that “invites us to notice
the broader networks that tend to include myriad sociotechnical devices”
(Bajde, 2013, 10). Within networks, devices serve to stabilize a given con-
figuration and reinforce certain sets of practices as valued and valuable. In
this case, APIs as devices are better understood as market devices. Market
devices “play a pivotal role in deploying and stabilizing special and tempo-
ral configurations” through which “consumption subjects and objects can
come into being and relation” (ibid., 10). That is, in this case, APIs set the
parameters around how information about social media participants can be
accessed and produced within the ‘marketplace’ of social media data. They
make this information both quantifiable and useable by simplifying interop-
erability and standardizing categories of activities in order to ease applica-
tions access and to reinforce and perpetuate social media use. Gerlitz and
Helmond (2013) describe ‘like’ buttons in similar terms, noting how these
have the capacity “to both metrify and intensify user affect and engagement
by turning them into numbers” (1361).
Though market devices have increasingly been looked at in relation to
market studies (Araujo et al., 2008), there have been limited analyses of
these devices in connection with studies of consumption (Badje, 2013). By
describing APIs as market devices, the focus here is on how these sets of
protocols and categories can be seen as technical instruments that serve
to bring together technologies, people, and activities. APIs are part of a
configuration of activities that serve to construct an economic, social, and
technical space. Within this space, everyday practices are documented and
digitized, and APIs make this process accessible. This is largely for com-
mercial purposes, if not on the part of the application that uses this data,
certainly on the part of the social media corporation. As market devices,
APIs are “material and discursive assemblages” that, given the position
social media occupies in the various forms of contemporary capitalism,
“intervene in the construction of markets” (Muniesa et al., 2007, 2). APIs
can be seen to have fundamentally shifted the usefulness of social media
by creating an accessible and systematically categorized dynamic archive
of data. Some applications themselves are material and discursive assem-
blages and, as such, can be seen as market devices as well. However, they
 A Social API for That 41
are predicated on APIs that allow data to make sense—to be categorized
and classified in ways deemed useful.
For example, an application relies on an API to gain access to a user’s
posts or status updates on social media. This allows the owner of the appli-
cation to learn the activities of social media users—whether they take notice
of promotions, make references to a particular brand or company, or indi-
cate an activity in which they regularly participate. A fitness company, for
instance, can gain data through an application that tracks a participant’s
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

exercise routine and the posting of this on a social media. Or a media com-
pany can quantify the number of times a subscriber references their publi-
cations and begin to quantify the spread of a given story. This information
is determined by the classification and categorization of activities deemed
relevant by the social media, and its APIs ‘ascribe’ (see Akrich, 1992) a
particular way in which this information is conceived of and retrieved. This
information is constantly updated and made accessible so that the social
media participant can be assessed historically and dynamically. Additionally,
given their interactive capabilities, APIs similarly ascribe how an application
may produce data on behalf of the user, such as posting updates, from arti-
cles read to exercise activities to personal location. As a marketing device,
the ascription process for an application allows that application to see and
produce “certain (but not all) things very well” (Badje, 80).
As market devices, APIs are objects with agency—these articulate the
means by which applications can act and set the parameters around those
actions (Muniesa et al., 2007, 2). They create an open architecture for
sharing content and data between communities and applications that are
increasingly seen as valuable. More specifically, in a context in which social
media is increasingly integrated into everyday life, APIs stabilize the interac-
tions among organizations, applications, and the dynamic practices of social
media users. They are not simply intermediaries, but devices that determine
the architecture of these interactions. In order to understand a bit better
what this architecture looks like, we will look in the following section more
specifically at the developer options and the permissions process seen by
users.

APIs IN PRACTICE

As market devices, APIs sit at the intersection of social media participants

(their activities, posts, and aspirations), technology developers and their
platforms, and those that seek to use the information being produced within
social media for particular goals. This latter group includes marketers, poli-
ticians, social activists, social scientists, and more. The normative expecta-
tions embedded into system design (see Madrigal, 2013), anticipated by users
and shaped through everyday consumption of and within social media, are
becoming an increasingly crucial resource for understanding and arguably
 42 Jason Pridmore
guiding people’s everyday lives. This resource allows for the interpretation
of social media user profiles into various forms of social, economic, politi-
cal, and geographic categories (as well as many others). APIs are the means
by which this information is made accessible and continually updateable.
There is a proliferation of APIs across a variety of technologies—currently
more than 11,000—with a number of applications or ‘mashups’ that rely
on multiple APIs to accomplish certain integrated technology and software
goals (see www.programmableweb.com). Not all of these APIs are forms of
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

social media; in fact, most are focused on less ‘social’ data, such as maps or
statistics, for instance (the Google Maps API remains the most widely used
API). However, this chapter focuses on social media and, specifically, two of
the largest social media providers, that of Facebook and Twitter.
While a majority of social media ‘enhanced’ interactions are dependent
upon APIs, some are simply part of the internal social media environment
itself. Along with and integrated into APIs, these tools allow for extended
information gathering and dissemination through social media. They make
it possible to allow people to keep up with the latest trends, information,
products, and services offered by a number of brands, for instance via
the ‘like’ button on Facebook, ‘following’ an organization on Twitter, the
“check in” feature of Foursquare, the ‘pin-it’ feature for Pinterest, and more
generally the ability to share content with others. The information produced
and distributed by these are dependent on the way in which social media
have set up the interface between users and these organizations, but most of
this occurs internally.
However, to access and use relevant data produced by and about social
media participants (including the internal features) beyond the enclosed
social media environment, websites, organizations, and marketers rely on
applications that interface with social media data. The API determines the
mechanics of accessing data.
For most APIs, after an initial agreement, the more ‘technical’ aspects of
the connections among the social media participant, the application, and
the social media provider, quickly disappear. None of the technical complex-
ity that underpins these devices is evident. Once activated, the agreement
to ‘connect’ or ‘log in’ (or other mechanisms that rely on APIs) permits
consumer access to specific branded content, features, and services that are
available both on social media and proprietary websites. In many cases, this
permission process also allows these organizations significant and continual
access to the user’s social media profile and the ability to alter these profiles,
post ‘messages’ or ‘status updates’ for them, geographically locate them, and
more. It is this initial moment of application request and user acceptance
that is of particular interest for this chapter.
The singular ‘click’ that indicates a user accepts the terms and conditions
for the application, including what an application can access or do in rela-
tion to the social media profile, is the initial moment through which software
components and data are enabled to communicate with each other. These
 A Social API for That 43
are based on the protocols or rules set out within an API (for further infor-
mation, see 3 Scale 2011: “What is an API? Your Guide to the Internet
(R)evolution” and Anonymous 2014: “API versus Protocol”). For social
media, the relevant data held by providers is connected with external appli-
cations and systems that can make use of this information. The permission
process through which an application can access and produce data is crucial;
from this point forward, applications appear to be seamlessly integrated
with social media resources, and the permission request only appears again
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in very limited circumstances.

So what does this primary access look like—how does it appear? What is
the process by which an application obtains a social media participant’s per-
mission to access and amend his or her data? And what data is available? As
the API is predetermined by the social media platform, there are two main
processes through which an application begins to interact with social media
data—either accessing user data or amending their profile. On the one side,
the application developer has to select certain features and requirements
for the social media participant to accept, and on the other side this person
must accept these requirements by granting permission to the application
to access their data. What follows details both the application developer’s
view and the consumer acceptance process for both Facebook and Twitter.
Though the uses of their APIs are not always the largest by volume, they are
perhaps the best-known social media APIs.

Application Developers’ View

In the case of Facebook, an application may require the social media to act
as a form of ‘identity’ verification (sign-in), be used for a particular game
portal, allow for personalized content, or produce certain consumer analyt-
ics. By default, applications that are given access (permission) receive the
following pieces of basic data: ‘id name,’ ‘first_name,’ ‘last_name,’ ‘link,’
‘username,’ ‘gender,’ and ‘locale’—this is described as ‘public data’ in what
follows. More information can be requested of the user through specific
permissions through the Facebook Open Graph API. Figures 2.1, 2.2, and
2.3 indicate the additional information an application developer can request
from users through Facebook.
These additional permissions encompass more than 70 data fields beyond
a user’s basic public profile information. Given the increased reliance upon
social media for personal expression and connections to others, these fields
are intimately connected to the ‘who’ a person is in terms of characteristics
and attributes, the way a person is identified, and the relationships she may
have. Based on the data in these fields, an application can gather rather sig-
nificant amounts of information that is available about its users.
Twitter’s current REST (representational state transfer) API does not pack-
age the data that can be accessed by using their interface in the same way
as Facebook. Rather, all the accessible data by applications can be retrieved
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Figure 2.1 Data permissions for users

Source: publicly available developer interface form from Facebook

Figure 2.2 Data permissions regarding friends of users

Source: publicly available developer interface form from Facebook
 A Social API for That 45
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Figure 2.3 Extended data permissions requested from users

Source: publicly available developer interface form from Facebook

or posted for the Twitter user based on data from the following categories,
which contain some 80+ ‘GET’ or ‘POST’ commands: Timelines, Tweets,
Search, Streaming, Direct Messages, Friends & Followers, Users, Sug-
gested Users, Favorites, Lists, Saved Searches, Places & Geo, Trends, Spam
Reporting, OAuth (Open Authorization), and Help. For instance, under the
Friends & Followers category, the GET friends/list request can be used to
return “a cursored collection of user objects for every user the specified user
is following (otherwise known as their ‘friends’)” https://dev.twitter.com/
docs/api/1.1/get/friends/list). That is, information on a limited number (cur-
rently 30) of a Twitter user’s followers can be accessed by an application
once that Twitter user has given permission to the application. Or using the
POST statuses/update allows the application to “update the authenticating
user’s current status, also known as tweeting” (https://dev.twitter.com/docs/
api/1.1/post/statuses/update). This includes the possibility of geolocating
the user if geolocation tags are enabled. For the application developer, the
process of being able to obtain permissions for all of the ‘GET’ resources
or ‘POST’ possibilities (though there is a substantial amount of overlap for
these) appears through a rather simple read/write developer request screen
shown in Figure 2.4.
What is readily visible in Figure 2.4 is the requirement that access tokens
need to be renegotiated if in fact there is a change in the permission requests.
However, if there are no changes to permission requests, once granted,
acceptance on the part of the user is given in perpetuity. One of the biggest
 46 Jason Pridmore
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Figure 2.4 Data permissions for Twitter

Source: publicly available developer application manager interface form from Twitter

issues for Twitter appears to be the ability of an application to read, write,

or delete direct messages—those that are seen only by the user. This is the
focus of the application permissions model that is hyperlinked in Figure 2.4.
It seems that public tweets (or followers and retweets) are far less a concern
than those messages that are not made public by default. However, as much
of what happens with Twitter is done on and through a public forum, the
API focuses on allowing an application to quickly and easily access this more
public data.
This is a rather significant difference in the two APIs. Twitter’s API for the
most part acts to collate significant pieces of publicly available data about a
user as well as offer a form of user identification and authentication (largely
by relying on OAuth or Open Authorization protocols). It also allows an
application to ‘tweet’ for the user, but this is connected to the ‘buttons’
on websites or applications that say ‘tweet this’ in connection with a news
article, achievement in a game, recent purchase through a webshop, and
so forth. Facebook’s API offers identity verification as well but also allows
applications access to data that is not public—that goes beyond the user’s
public profile and that is arguably more robust in terms of indicating social
connections more directly. Both of these resources allow the application to
gain continual access to a ‘profile,’ updating the data that is produced in the
everyday use and interactions with the social media.

SOCIAL MEDIA PARTICIPANT PERMISSIONS

The above description and figures are indicative of how Facebook and Twit-
ter (differently) organize and standardize access to activities done within their
social media platforms. These categories are made available for application
 A Social API for That 47
developers, who then can choose which features of the API they wish to
include for their API. This is the basis for knowledge about personal pro-
files, statuses, connections, and the relevant activities of people who partici-
pate on Facebook and Twitter. Though several of these items are oriented
to more qualitative descriptions than easily quantifiable data (for instance
Facebook’s categories of ‘about me,’ ‘interests,’ ‘religion_politics’), many
of these are oriented toward pre-formulated responses that appear once a
person begins to fill these out in their profile (that they can choose to use or
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

not). Likewise, the GET statuses function for Twitter is likewise qualitatively
oriented, but the GET statues/retweets and GET friends/id can be compiled
and used for more quantitative and systematic purposes. The details about
‘who’ a person is, how they are connected to others and what they deem
worthy of ‘sharing’ can be quite significant here.
However, a brief examination of the application requests of several major
international corporations reveals that a number of these do not request too
many additional permissions from Facebook, with the exception of email
addresses and birthdates. The open possibilities for Twitter as read/write/
access permissions are, however, often present, but these are again connected
to the more public data that is central for Twitter and seem to be rarely used.
This may shift or change, but at present Figures 2.5, 2.6, and 2.7 indicate the
process by which applications ask permissions from users to use their data.
Figure 2.5 indicates a minimal request for data from Facebook, and Figure 2.6
shows a substantial request from the application for Facebook data (the black
square in each figure is where the user’s profile picture is inserted):

Figure 2.5 User permission request on Facebook from Oprah.com

Source: publicly available user authorisation form from Facebook
 48 Jason Pridmore
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Figure 2.6 User permission request on Facebook from Skype

Source: publicly available user authorisation form from Facebook

Facebook actually notes that “the more permissions an app requests, the
less likely it is that people will use Facebook to log into your app” (see https://
developers.facebook.com/docs/facebook-login/permissions). However, click-
ing on the public profile hyperlink indicates that part of this permission
is to give a somewhat extensive set of data, including your “name, profile
picture, age range, gender, language, country and other public info.” Again,
the preceding two figures only appear when a user chooses to ‘connect with
Facebook’ or ‘sign in with Facebook’ on the website or the application, and
the ‘OK button’ approves the process by which these applications (or sites)
gain initial and then subsequent access to personal and personalized infor-
mation, data, and features. This is the moment at which the data transfer
following what is set out by Facebook’s Open Graph API is begun and con-
tinues until the terms, conditions, or permissions requested change or until
the user changes his or her password.
 A Social API for That 49
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Figure 2.7 User permission request on Twitter from The Economist

Source: publicly available user authorisation form from Twitter

A similar process happens in the use of applications for Twitter. While

Figure 2.4 indicates the process on the application developer’s side for
Twitter, Figure 2.7 indicates the authorization request made of Twitter
users. This gives permission to a particular site to access certain pieces of
Twitter data so the user may continue to use the site. Beyond accessing
basic information from Twitter such as the name and ‘handle’ of the user,
the application is further able to update consumer profiles, learn whom
the user follows and is followed by, and even post information (tweets)
for that user from their site (for example, products purchased or articles
recommended).
This permission process indicates very specifically where a Twitter user
can find the applications for which he or she has given access and specifies
what an application can do. The most significant difference between per-
mission requests for Twitter and Facebook is the indication that the Twitter
application may post on a user’s behalf. Facebook’s Open Graph API focuses
on the information that external applications can receive rather than their
ability to post information for the user. Twitter, however, with its largely
publicly available data, has combined a simplified collection of this data by
an application with permission for that application to mediate particular
posts for the user. Facebook’s approach to this is different, relying on sepa-
rate processes known as ‘social plugins’ to post on behalf of the user, most
notably the like and the share buttons.
 50 Jason Pridmore
With the basic information of a user available in their ‘public profile,’
and presuming an on-going use of and/or interaction with these particular
applications, both these APIs facilitate an extensive collection of personal
data. This collection may be a bit myopic in nature, relying on a fragmen-
tary set of ‘identity attributes’ that make up the participant profile, but for
the purposes of evaluating and analysing use patterns and engagement, this
data can become quite significant. Of course, while Facebook and Twitter
might be the most prolific social media offering APIs with which application
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

developers are most likely to interface, other social media and services like
LinkedIn, Yahoo, Google, and Microsoft Live alongside Tumblr (owned by
Yahoo), Pinterest, and Instagram (owned by Facebook) also provide these
services. All of these set particular parameters around which social media
participants and their lives can be made ‘known’ and traceable. Further,
these figures are only indications of web-based processes, but applications
are predominantly found in connection with mobile devices, namely phones
and tablets. As numerous applications connect with and rely upon data
derived through both Facebook and Twitter APIs, the significance of these
bits of personal attributes that have been translated into specific attributes
and actions on social media only escalates, particularly in terms of its rela-
tion to ‘markets.’
Given the availability and frequent use of social media, seeing APIs as
market devices allows us to understand how applications connect to the
inner workings of social media and its users in order to gain significant infor-
mation about these users. Precisely how this connects with contemporary
understandings of identity is the focus of the next section.

APIS AND CONCEPTIONS OF IDENTITY

The request and permissions processes indicated earlier show how catego-
ries of data—predefined by social media—become available to external
applications for both the production and consumption of user data. This is
predicated on APIs. They determine the access of applications to that data.
This data is representative of the lives of social media participants, who use
the platform to communicate and relate to others. They both indicate to oth-
ers ‘who’ they are and are perceived in certain ways by the ‘connections’ the
social media platform has provided for them. In the earlier text, the descrip-
tions of requests and permissions are focused on a social media participant’s
profile, while the more theoretical concerns in the descriptions preceding
this were focused on conceptions of identity. In large measure, the two terms
are used interchangeably in common discussions, but here it is important to
see the profile as a digital collection of attributes and a means of producing
and ‘consuming’ identity. As such, social media are in this sense identity
intensive environments, with profiles as the means for both the production
of identity (or identities) and their consumption.
 A Social API for That 51
There have been more than two decades’ worth of work examining the
relationship between technology and the production and consumption of
identity, particularly in an online environment (see Turkle, 1995; Zhao
et al., 2008; Trottier in this volume). This has largely been focused on the
participants—their experiences and practices—rather than on the processes
of how they are identified. Identity, as has been argued, is a central concern
in contemporary life. Zygmunt Bauman suggests that identity is experienced
as the new problematic—that is, it is what we are all focused on in part
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

because the contemporary world is unable to present or legislate the world

as cohesive, continuous, and consistent (Bauman, 2000; see also Elliott,
2007). While it is not clear that this cohesive, continuous, and consistent
world ever genuinely existed (see Latour, 1991), Bauman’s sociological per-
spective aligns with others that suggest that we are no longer ‘given’ an iden-
tity by the (social) world(s) that surrounds us; rather, identity is predicated
on symbolic constructions based on the choices we make or are made for us
(see Giddens, 1991). Increasingly, these symbolic constructions can be seen
in terms of associations made present in social media profiles that have in
turn become more accessible through APIs.
Interactional and relational conceptions of identity, drawn from works
such as those from symbolic interactionists like Charles Horton Cooley and
George Herbert Mead and dramaturgists such as Erving Goffman,1 have
largely ignored the role of technologies more substantively in their abil-
ity to determine or direct identity, particularly in relation to social media.
There are some exceptions, as for instance Katherine Walker draws directly
upon Goffman’s work to describe identity production through the Inter-
net (2000), and Zhao and colleagues do so similarly in relation to early
Facebook use (2008). Even in these exceptions, technology is largely seen
as a tool in the identity-construction process. That the interactional basis
for the development of identity can occur in conjunction with technology
was something Anthony Giddens suggested to some degree long before the
advent of the Internet. He noted that with the “development of mass com-
munication, particularly electronic communication, the interpenetration of
self-development and social systems . . . becomes ever more pronounced”
(1991, 4). It is this interpenetration that becomes evident in the use of APIs,
as these technologies allow for desirable applications to continually retrieve
and produce personal data for the user—thus they are about both the devel-
opment of self and the (social) systems in which this occurs.
This interpenetration has also been readily seen in very early texts about
the Internet. Sherry Turkle’s descriptions in Life on the Screen: Identity
in the Age of the Internet (1995) explores the relationship between new
information and communication technologies (ICTs) and their effects upon
and production of personal identity. The point of her work was to uncover
“the story of constructing identity in the culture of simulation” that in her
terms was seen as “evidence of fundamental shifts in the way we create and
experience human identity” (ibid., 10). However, her work was focussed on
 52 Jason Pridmore
the early days of the Internet, specifically interactions that occurred within
multi-user domains (MUDs)—online role-playing games in which text on
the screen was used to choose and describe characters (avatars), actions,
and environments. It was limited in scope and focused on a limited number
of persons who participated in these communities. With the expansion of
social media and its pervasiveness in everyday life, the construction of iden-
tity and its experience through and with technology are no longer clearly
limited within contexts of ‘simulation.’ Rather, the technological foray of
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

early Internet users in producing and consuming ‘identities’ through social

media profiles has become an almost universal experience.
The entirety of these approaches to identity and technology stands largely
in contrast to ascribed notions of identity, namely the means by which persons
are given an ‘identity’ in terms of identification documents and the develop-
ment of identity-management systems. The process of the identification of
citizens, for instance through technologies of identity cards, has elsewhere
been more explicitly seen as a form of surveillance (see Lyon, 2009). Without
delving into the sometimes conflicting and sometimes complementary ways in
which identity in the form of personal subjectivity and sense of self intersects
with identity in the form of identification by external structures and technolo-
gies (see, for instance, Jenkins, 2008), it is sufficient to say that APIs connect
with both notions of identity through participant profiles. They are increas-
ingly the way in which ‘identity’ is both provided for and produced on a daily
basis. APIs give access to social media as an ‘identity provider,’ structuring
this identity-provision process on the basis of a singular notion of identity (see
van Dijck, 2013). They also give access to prescribed lists of ongoing activities
and associations in the form of uniform fields of data. Yet the ‘provision’ of
identity and the changing identity attributes found within social media profiles
are also contingent and relational—they stem from voluntary practices and
connections chosen by users. These profiles are based on an accumulation of
everyday choices and experiences that have been made digitally accessible.
APIs render readable the active production of identity on a day-to-day
basis as it is increasingly intermingled with consumption (see Slater, 1997;
Bauman, 2005; Lury, 2011) or perhaps further, the notion of ‘prosumption’
(Beer and Burrows, 2010; Ritzer and Jurgenson, 2010). This production
of personal identity can be seen to take place in the consumption choices
people make (or not) in order that they may “construct, maintain, interpret,
negotiate [and] display who they are to be or be seen as” (Slater, 1997,
84). These choices have become more visible through social media profiles
as their ever-extending relevance and availability within everyday life has
grown. ‘Consumption’ in this sense should be seen more broadly in terms of
daily practices or use rather than simply ‘purchases’ (see Shove et al., 2007).
As such, it includes the significant amount of content produced by partici-
pants within social media (which is what has given rise to the ‘prosumption’
neologism), even as this is guided and directed by the structures and everyday
practices that make up these media (see Madrigal, 2013). Prosumption here,
 A Social API for That 53
should be seen as a ‘sensitizing concept’ intended to reiterate the observation
that people are actively involved in the production of media content (Beer
and Burrows, 2013, 3). This is the point at which technology, consumption,
and identity become increasingly intertwined.
Social media are predicated on content that is ‘prosumed’ every day in
increasingly large amounts—personal information, status updates, connec-
tions, pictures and videos, notes, shared links, and so on. The mixed set of
content, often significantly guided by social media that attempt to ‘get out
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

of the way’ and allow participants to produce and receive this content (see
Madrigal, 2013), becomes relevant for a variety of purposes. Making this
content available, editable, measurable, and quantifiable through and from
social media is what APIs allow. They are important for understanding and
conceptualizing the fluctuating collection of attributes indicated to others
and those characteristics attributed to and expected of us by others within
social media profiles. In consumption terms, “the anthropology of consump-
tion has . . . clearly shown classifying products, positioning them and evalu-
ating them inevitably leads to the classification of the people attached to
those goods” (Callon, Méadel and Rabeharisoa, 2002, 212). By extending
this to more proactive forms of consumption, what gets ‘done’ with social
media is both a production and classification of participants and their con-
tent. Social media profiles are entities in which ‘identity’ is both produced
and provided, and APIs stipulate and enable access to these identities.

STABILIZING DIGITAL RELATIONS

Given the understandings of identity noted earlier, from its provision to its
production through social media profiles, what precisely is it that social
media APIs can be seen to ‘do’ in terms of identity? And what is it that these
identities do to APIs? APIs are very much about ‘identities’ in use. They can
be seen to make more accessible the socio-technical productions of iden-
tity in many facets. First, APIs both show and reinforce how applications
have become a central part of people’s everyday practices with technology
and their increased reliance on social media. Second, they reiterate the self-
referential but also contingent processes that are a part of the expression of
identity by producing both flexible and persistent categories and classifica-
tions that connect with personal profiles. Finally, these social media APIs
have made accessible data profiles indicative of the flexible and mutable
identities provided for and produced in relation to social media.

The Shift to Applications

Over the past several years, we have seen the transformation of the Inter-
net into interconnected and (mostly) smoothly functioning sets of software
rather than through the traditional ‘window’ of a browser. These softwares
 54 Jason Pridmore
are applications that make use of a variety of resources and are integrated
into multiple technological devices, from smartphones to tablets to laptop
computers. They are part of the increasingly ‘customizable’ range of devices
that rely more on applications and less on browsers to access information
or perform certain functions for the user (see Khalaf, 2013). In the case of
social media, the data made available through the APIs of Facebook and
Twitter provide a significant resource for a variety of purposes, from track-
ing activities to conversing with friends or acquaintances to meeting some-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

one new to playing a game and numerous other possibilities.

It is this shift to applications that is significant for identity. Following
Anthony Giddens’s understanding that in terms of developing our identity
(or self more specifically), “we have no choice but to choose” (1991, 81),
and given the proliferation of portable devices that can access the Internet,
applications provide a variety of possibilities to enact the “routine prac-
tices” that are part of a “lifestyle” (ibid.). These lifestyles are very much
a part of the ‘choices’ people are making in everyday life, and myriads of
choice become expressible and embedded in bits of technology. Applications
connect the users of these technologies into a socio-technical configuration
of practices, people and information that they deem important and, with
increasingly directed advertising and marketing, the people and information
deemed important for them.
The relationship of applications to social media is mixed. On the one
hand, social media applications are themselves fairly robust technologies
that attempt to serve a number of ‘functions’ for their participants. For
instance, both Twitter and Facebook as applications and websites promote
particular social connections that can include indications of location and
messaging services, for which Facebook in particular has introduced its
own stand-alone messaging application. On the other hand, a number of
non-social media applications promote the connection with social media
APIs like Twitter and Facebook as part of their identity verification pro-
cesses and/or the promotion of social connections. Some of these appli-
cations are dependent upon or require a connection with social media.
When connected, social media APIs allow for the collection and produc-
tion of significant amounts of data by the social media participant. Taken
together, the proliferation of softwares that interconnect specifically with
social media APIs do not just indicate potential connections but the ‘suc-
cess’ of these social media to be an increasingly ubiquitous part of people’s
lives.

Identity Expressions and Social Media

The second facet that indicates how APIs connect with socio-technical
productions of identity is that these reiterate and produce certain types
of ‘relevant’ categories and classifications. The choices for a short-term
diversion in the form of a game or the ‘pinning’ of a recipe to what can
 A Social API for That 55
be seen as a virtual scrapbook or the indication of romantic interest in a
person based on a photo and profile, all of which are available through
applications, reinforce particular attributes of the application user. As
many of these either request or are based on connections with social
media data, they draw upon and produce particular forms of data con-
nected with the lifestyles of the social media participant. There is a dual
process here—social media participants choose what might be relevant
for themselves in the form of social connections, entertainment, activities,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and their lifestyle—all the things they ‘like’ (see Gerlitz and Helmond,
2013)—even as these are being reinforced by those they are connected
to by the algorithmic analyses of social media providers. They both refer
to current self-perceptions and aspirations as well as being placed upon
them from without.2
Social media provide reference points through which others, including
applications, are able to classify and categorize participants. APIs form the
basis by which this content of “circulating references” (see Latour, 1999;
see also Beer and Burrows, 2013) within social media—that is data about
the (selective) content of everyday life—moves from these proprietary sys-
tems to external applications. As such, APIs substantially contribute to and
reinforce forms of digital identity that applications are able to retrieve, but
they also allow the freedom of social media participants to shift and change.
Participants are free to change or emphasize particular associations and
relationships, to follow new people and like different status updates even
as these new choices are producing a ‘new and improved’ version of the
participant’s (altered/updated) identity. These decisions are then integrated
into the social media platforms’ algorithms, and different status updates
or tweets (in the cases of Facebook and Twitter) become more readily vis-
ible and promoted. APIs standardize access to certain social media practices
and categories of information that participants willingly choose to associ-
ate themselves but do not define these further—participants are encouraged
to ‘fill in the spaces’ (see Madrigal, 2013). The data they produce creates
selective biographies of participants in the first instance, but the permis-
sions given by social media participants further allow applications to access,
modify, and build upon these biographies in ways that are seen as relevant
to the participant, even as he or she may pursue new experiences, lifestyles,
relationships, or associations.

Market Devices That Make Changing Identities Accessible

It isn’t just that social media APIs reinforce the use of applications and provide
ways to express or classify the identities of participants—they also make a
contingent and relational notion of identity ‘workable’ on a continual basis.
Though the majority of social media are built on proprietary technologies,
the development of APIs allows interoperable access to information from
these media by different external systems. This is enabled in the first place at
 56 Jason Pridmore
the moment in which a social media participant accepts the conditions at a
singular point—the ‘clicking’ of the ‘log-in’ or ‘authorize’ button. This click
sets in motion the potential for information to be both obtained from and
produced for social media. In most normal circumstances, the authorization
obtained for the transfer of data remains in perpetuity, though this may be
requested at each update of the application.
Social media APIs provide a way for applications to measure and moni-
tor the everyday practices of social media participants by outside organiza-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tions. Though the data made available by these APIs is rarely requested in
full, the selection of categories of data made available by the likes of Face-
book and Twitter, amongst others, provide a selective and partial ‘under-
standing’ of ‘who’ a social media participant is on an on-going basis. In
terms of identity, this is not something simply voluntary or something sim-
ply classifiable and imposed, but rather an identity that is shaped in rela-
tion to both voluntary choices and categorical assumptions or analyses.
Shifts to different activities, locations, friendships, and other affiliations
are not only possible and anticipated, but these are easily made traceable
through the data provided. This not only serves to stabilize relationships
with identity as an ongoing production but allows these identity profiles
to move beyond the parameters of immediate interaction. As the permis-
sions noted above, a number of these application request the authority
to act on a user’s behalf specifically in relation to the social media. API
interactions (or calls) now number in the billions per day for larger social
media companies (such as Facebook, Twitter, and Google), and there are
others moving in this direction (Microsoft, LinkedIn).3 To be clear, these
API calls are not being initiated every time by social media participants.
Rather, applications are seeking and posting information in an automated
fashion. Arguably the sheer number of these calls can be seen to stabilize
conversations with consumers by making them more or less constant. This
strengthens the durability of consumer audiences, limiting the liabilities
inherent in the “irrational consumption behaviours of individual actors”
(Bauman, 2001, 17).
In the process, social media APIs can be seen to render the complexities
of changing digital identity accessible and stabilize access to this informa-
tion. The significance in the use of social media APIs in their various forms
is that these are distinct means through which observations and/or on-going
‘conversations’ are made embedded into technologies. This is of particular
interest to marketing companies (see Pridmore, 2013) as well as the social
media companies themselves, particularly as they ‘tweak’ their participant
interfaces, and should not be underestimated. Through the acceptance pro-
cess set out by APIs, applications connect social media participants with a
variety of resources and provide services on the basis of standardized sets
of dynamic information. They are a vital part of how a number of different
elements connect—they draw together identity, technology, marketing, and
consumption into a unified market device.
 A Social API for That 57
CONCLUSIONS

As social media play an ever more important role in people’s daily routines
and practices, they have become a significant aspect in the development of
personal understandings and presentations of identity. The consumption (or
perhaps ‘prosumption’) of social media has begun to change how identity is
‘made to happen’ in everyday life. The unseen role of APIs allowing for the
collection of rich personal data as well as the modification and alteration of
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

digital profiles suggests that identity remains active beyond the intermittent
interactions, categorizations, and profiles found within digital social media.
There are of course numerous concerns on the part of participants regarding
their own profiles (see Trottier, Chapter 3, this volume), but overall the use
of social media is largely seen as voluntary, as are the choices to use applica-
tions that connect with social media data.
However, there are some significant issues that arise in the ‘one-click’ process
presented by application programming interfaces. First, this raises concerns
about a continual ‘care for the self’ in a Foucauldian vein (Foucault, 1988;
see also Elliott, 2007). There are expectations of a continual engagement
built into social media and into applications that can access these media.
What is interesting is the degree to which applications are able to act on the
behalf of others in posting ‘status updates’ and ‘tweets’ in the social media
cases presented here. The acceptance process makes evident the blurred
boundaries among social media, its participants, softwares and hardwares,
and the expressions of personal identity.
Second, although there is flexibility in content, APIs dictate the types of
data that become accessible in terms of its own categorization. They rein-
force particular norms and limit the complexity of identity by dividing it into
certain attributes and associations its platform is dependent upon. This is of
course necessary to stabilize the production and consumption of ‘identity,’
yet in this way it seems that APIs can increasingly dominate the process by
minimizing opportunities for participants, particularly in decisions involv-
ing the connections of applications with social media.
Finally, the acceptance process for applications enacts a form of visibility
that is less than reciprocal, as the complicated protocols and processes of APIs
are not seen or, by and large, understood by participants on social media (see
Bodle, 2011). There is relatively little understanding of what data are both
available and not in relation to application access. What is clear is that both
marketers and social media platforms are increasingly leveraging these tools
even as consumers increasingly rely on their convenience. Social media may
facilitate participants in the production of personal identity—producing rel-
evant profiles about themselves—but the accessibility of these profiles through
APIs makes visible images and activities which individually may not be of
consequence or value, but in combination with others it, allows patterns to
become visible or associations (even non-obvious ones) to be connected. These
are critical ethical and social concerns given the ubiquity of social media.
 58 Jason Pridmore
This chapter has begun to show that although both social media and appli-
cations have become significant technologies in the contemporary world, it
is the connections between them and their users in the relatively obscure
and unseen APIs that are of significance. These have substantial implications
for how ‘identities’ are being both produced and consumed through digital
profiles, as APIs are now important parts of the socio-technical configura-
tion of identity in the contemporary world. This is particularly true in rela-
tion to the use of social media APIs. To begin to understand transitions in
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

identity, we must begin to understand and examine the underlying protocols

and processes that invisibly direct and are directed by the relations between
everyday practices and the technologies that are an increasing part of them.

NOTES

1. The work of Cooley, Mead, and Goffman, amongst others, focuses on con-
ceptions of the self. In large measure the distinctions between the notion of
‘self’ and that of ‘identity’ overlap significantly, but this discrepancy has raised
several critiques as to the usefulness of the concept of identity (see Brubaker
and Cooper, 2000).
2. This can be seen in terms of governmentality, which would require further
discussion and analysis (see Beckett, 2012).
3. Details and statistics on API usage rates are available at www.programmable
web.com/ or at www.slideshare.net/mashups/web-ap-is.

REFERENCES

3SCALE. (2011) “What Is an API? Your Guide to the Internet Business (R)evolu-
tion.” www.3scale.net/wp-content/uploads/2012/06/What-is-an-API-1.0.pdf.
AKRICH, M. (1992) The De-scription of Technical Objects. In Bijker, W. E. & Law,
J. (eds.) Shaping Technology/Building Society—Studies in Sociotechnical Change.
Cambridge (MA): The MIT Press, 205–224.
ANONYMOUS. (2014) “API Vs Protocol.” Accessed May 15. http://c2.com/cgi/
wiki?ApiVsProtocol.
ARAUJO, L., KJELLBERG, H., & SPENCER, R. (2008) Market Practices and
Forms: Introduction to the Special Issue. Marketing Theory, 8(1). p. 5–14.
BAJDE, D. (2013) Consumer Culture Theory (re)visits Actor-Network Theory: Flat-
tening Consumption Studies. Marketing Theory, 13(2). p. 227–242.
BAUMAN, Z. (2000) Liquid Modernity. Cambridge (UK): Polity Press.
———. (2001) Consuming Life. Journal of Consumer Culture, 1(1). p. 9–29.
———. (2005) Liquid Life. Cambridge (UK): Polity.
BECKETT, A. (2012). Governing the consumer: technologies of consumption. Con-
sumption Markets & Culture, 15(1), 1–18.
BEER, D., & BURROWS, R. (2010) Consumption, Prosumption and Participatory
Web Cultures: An Introduction. Journal of Consumer Culture, 10(1). p. 3–12.
———. (2013) Popular Culture, Digital Archives and the New Social Life of Data.
Theory, Culture & Society. (April) p. 47–71.
BODLE, R. (2011) Regimes of Sharing. Information, Communication & Society,
14(3). p. 320–337.
 A Social API for That 59
BRUBAKER, R., & COOPER, F. (2000). Beyond “identity”. Theory and society,
29(1), 1–47.
CALLON, M., MÉADEL, C. & RABEHARISOA, R. (2002) The Economy of Quali-
ties. Economy and Society, 31(2). p. 194–217.
DIX, A. (2011) Living in a World of Data. In Proceedings of the ACM SIGACCESS
Conference on Computers and Accessibility (ASSETS ’11). New York, NY.
ELLIOTT, A. (2007) Concepts of the Self: Second Edition. Cambridge (UK): Polity
Press.
FOUCAULT, M. (1988) The History of Sexuality, Vol. 3: The Care of the Self. New
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

York: Random House LLC.

GERLITZ, C., & HELMOND, A. (2013) The Like Economy: Social Buttons and the
Data-Intensive Web. New Media & Society, 15(8). p. 1348–1365.
GIDDENS, A. (1991) Modernity and Self-Identity: Self and Society in the Late Mod-
ern Age. Stanford (CA): Stanford University Press.
JENKINS, R. (2008) Social Identity. New York: Routledge.
JUNG, G. & LEE, B. (2010) Analysis on Social Network Adoption According to
the Change of Network Topology: The Impact of ‘Open API’ to Adoption of
Facebook. In Proceedings of the 12th International Conference on Electronic
Commerce: Roadmap for the Future of Electronic Business. p. 23–32.
LATOUR, B. (1991) We Have Never Been Modern. Cambridge (MA): Harvard Uni-
versity Press.
———. (1999) Pandora’s Hope: Essays on the Reality of Science Studies. Cambridge
(MA): Harvard University Press.
LURY, C. (2011) Consumer Culture: Consumer Culture (Revised and Updated).
Rutgers (NJ): Rutgers University Press.
LYON, D. (2009) Identifying Citizens: ID Cards as Surveillance. Cambridge (UK):
Polity.
MADRIGAL, A. (2013) How Facebook Designs the ‘Perfect Empty Vessel’ for Your
Mind. The Atlantic. [Online] Available at www.theatlantic.com/technology/
archive/2013/05/how-facebook-designs-the-perfect-empty-vessel-for-your-
mind/275426/. Accessed 12 July 2014.
MILLER, D. (1987) Material Culture and Mass Consumption. New York: Blackwell.
MUNIESA, F., MILLO, Y. & CALLON, M. (2007) An Introduction to Market
Devices. The Sociological Review, 55. p. 1–12.
PRIDMORE, J. (2013) Collaborative Surveillance: Configuring Contemporary Mar-
keting Practice. In Ball, K., & Snider, L. (eds) The Surveillance Industrial Com-
plex: A Political Economy of Surveillance. New York: Routledge. p. 107–121.
RITZER, G. & JURGENSON, N. (2010) Production, Consumption, Prosumption:
The Nature of Capitalism in the Age of the Digital ‘Prosumer.’ Journal of Con-
sumer Culture, 10(1). p. 13–36.
SHOVE, E., et. al. (2007) The Design of Everyday Life. Oxford: Berg.
SLATER, D. (1997) Consumer Culture and Modernity. Cambridge (UK): Polity
Press.
TURKLE, S. (1995) Life on the Screen. New York: Simon & Schuster.
VAN DIJCK, J. (2013) ‘You Have One Identity’: Performing the Self on Facebook
and LinkedIn. Media, Culture & Society, 35(2). p. 199–215.
WALKER, K. (2000) ‘It’s Difficult to Hide It’: The Presentation of Self on Internet
Home Pages. Qualitative Sociology, 23(1). p. 99–120.
ZHAO, S., GRASMUCK, S. & MARTIN, J. (2008) Identity Construction on Face-
book: Digital Empowerment in Anchored Relationships. Computers in Human
Behavior, 24(5). p. 1816–1836.
 3 Caring for the Virtual Self
on Social Media
Managing Visibility on Facebook
Daniel Trottier
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

This chapter considers the production, sharing, and gathering of personal

information in a culturally and historically situated context, namely, univer-
sity students who used Facebook during its expansion in the late noughties. In
adopting Facebook as a platform for interpersonal communication, relations
with peers are recast in terms of visibility and exposure. These users grew
comfortable sharing with their peers but at the same time were troubled by
the cumulative exposure to those peers and others. Interpersonal social media
usage—including submitting their own personal information and watching
over what other users publish—rendered respondents visible to one another
in a way that warranted a care of the virtual self (Whitson and Haggerty,
2008). This involved both self-scrutiny and watching over what peers upload,
as this may reflect poorly on them. Yet even this care was complicated by
social media’s growth at the time and especially Facebook’s cross-contextual
information flows that further publicized information beyond respondents’
expectations. For respondents, watching over their profiles and their peers
was a necessary—but not sufficient—condition of self-care on social media.
In the years following its launch, Facebook began to mediate social life
in a manner that some users recognized as neither fully private nor fully
public. Unanticipated visibility and exposure on social media was—and still
is—pervasive, insofar as it is a product of ubiquitous mobile technology and
a rapidly growing user base. These developments require scholarly attention
on the conditions surrounding peer visibility on sites like Facebook. This
research asks: in a context where early users were coping with an increas-
ingly publicized platform, what compelled these users to engage in sharing
and collecting personal information on Facebook, how did they perceive
these conditions of visibility, and how did they manage their online presence?

EARLY USERS TAKING CARE

Between October and December 2008, I conducted a series of semi-structured,

in-depth interviews to consider how Facebook users coped with peer-to-peer
 Caring for the Virtual Self on Social Media 61
surveillance. A convenience sample of thirty undergraduate students who
use Facebook was selected for study. Participants were enrolled at the same
mid-sized Canadian university. These students were selected from all facul-
ties and were recruited through a series of posters on campus as well as
notices sent by email to undergraduate mailing lists. Of the thirty students
interviewed, twenty-three were women and seven were men. As women
are both more likely to use social media in Canada (Dewing, 2010) and
participate in research (Sax et al., 2003), this distribution was expected.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Seven respondents were in the first year of their studies, eight were in their
second year, five were in their third year, seven were in their fourth year,
and three were in their fifth year. Twenty-six respondents were in the fac-
ulty of arts and sciences, three were in health sciences, and one student
was in the faculty of education. This distribution is explained by the fact
that the majority of the student population are enrolled in the faculty
of arts and sciences. Twenty-seven respondents checked their Facebook
account at least once a day. The three that did not perform daily checks
instead received email notifications from Facebook. Interviews focused
on a set of themes, including describing respondents’ Facebook usage
over time, the types of personal information made available to others
and reasons for doing so, the types of personal information acquired
from others through Facebook and reasons for doing so, and perceptions
of information exchange on Facebook. When quoted, respondents are
identified by their program, year of study, and gender. Interviews were
recorded as MP3 audio files, manually transcribed, and coded based on
interview questions as well as additional themes that emerged during the
interviews.
The interviews occurred at a noteworthy moment in Facebook’s growth.
While the platform was still closely linked to university student life in North
America, it had been available to non-students for approximately two years.
These respondents reported using Facebook to socialize with peers on cam-
pus and abroad. However, they were also coming to terms with unwanted
watchers, including parents and employers. Some considered a long-term
presence on Facebook to be a growing concern for when they would enter
the job market, pursue postgraduate studies, or raise children. They antici-
pated that their presence on Facebook could harm these plans.
In the years following these interviews, Facebook’s population grew to
more than one billion users, it experimented with several advertising schemes,
and it is now openly acknowledged to be a source of evidence for police and
other investigative agencies (Omand et al., 2012). Contemporary concerns
that are linked to these developments include how personal information may
be repurposed in ways that users neither anticipate nor endorse (cf. Trot-
tier, 2011). The focus of this chapter predates these developments, yet these
respondents were coming to terms with unanticipated forms of visibility on
Facebook. To be sure, Facebook has changed extensively in the years fol-
lowing these interviews. Its interface processes and displays personal details
differently, and a comparatively vast and heterogeneous user base engages in
 62 Daniel Trottier
a broader range of practices. However, a basic distinction remains between
information that is visible to all users and information that is meant for a
sub-set of users. Knowing and maintaining this distinction likely endures as
an element of care for the virtual self on Facebook.
Respondents described their presence on Facebook as a matter of being
both the subject and the agent of targeted scrutiny. In this context, Face-
book stood as an exemplar of social media. Respondents used it to maintain
profiles, upload photographs, and make personal information available to
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

their peers. Facebook shared some features with other social media, such
as the pervasive construction of an online presence populated by personal
information. Yet this presence transcended any specific context, and this
characteristic has become more evident in the years following this study.
Thus, sharing personal details with specific peers was not the extent of the
visibility it offers active users. The fact that invisible and unexpected audi-
ences (boyd, 2008) can also access this information is a risk of social media
use that has grown along with its user population.
Respondent adoption of social media was a form of visibility wherein
some everyday interactions more closely resembled surveillance. Surveil-
lance refers to the covert, sustained, and targeted collection of information
of an individual or group of individuals (Lyon, 2001). Respondents were
aware of this condition, with one respondent noting that through “examples
that happen all the time, you realize like how much some people could
be watching everything you say” (Artsci2F). Surveillance is a broader phe-
nomenon than data collection. It typically relies on asymmetrical relations
of visibility and power and is typically practiced in order to further these
relations. It is a prevailing organizational logic of late modernity. Facebook,
for example, organizes relations between individuals. Not only are inter-
personal social ties mediated on an organizational platform, but interper-
sonal activity becomes asynchronous. Users watch over each other’s profiles,
which may take precedence over directly watching each other. Other kinds
of surveillance take place through Facebook (Trottier, 2011), but peer rela-
tions also become more surveillant in nature.
Facebook can potentially be used as a surveillance tool, for example, by
an individual who stalks a targeted user by monitoring their activity on this
platform. However, a lot of interpersonal activity on this platform does not
strictly fit this description. User activity on Facebook ranges from willingly
publicizing details to attempts to obtain private details from another user
without their awareness or consent. However, respondents note that risk of
the latter shapes their relations for users and their practices on the site. One
respondent posted a link to politically inflammatory content on her profile.
Upon discovering that this reflected poorly on her, she stated, “there are
things that—it’s better just to sort of take care of it [by keeping it] private
overall. And it’s good to stay sort of apolitical on Facebook” (Artsci4W).
Thus, the consequences of visibility on Facebook may contribute to self-
monitoring and a dampening of political speech within this medium. This
 Caring for the Virtual Self on Social Media 63
particular response anticipated the risks of unwanted data collection and
exposure linked with the subsequent expansion of the Facebook platform,
as well as other platforms.
Respondents occasionally experienced interpersonal (or peer) surveil-
lance as a violation but also came to regard it as a pervasive condition of
social media, suggesting a further normalization of surveillance practices
(Murakami Wood and Webster, 2009). The intervisibility (Brighenti, 2010)
these individuals exhibit and exploit is a prevalent feature of contemporary
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

sociality. Peer surveillance is also mutual on social media, as users are able
to watch and be watched. Earlier research on user-led surveillance practices
noted that the possibility for users to watch others as well as make themselves
visible adds an empowering dimension to their use of digital media (Koskela,
2004; Albrecthslund, 2008). Yet on Facebook, intentional visibility cannot
be disassociated from unanticipated exposure. One function “creeps” into
another, and information “leaks” to new contexts (Lyon, 2001).
In the context of peer surveillance, Andrejevic describes a media culture
that compels individuals to watch over untrustworthy peers (2005). Para-
doxically, an increased reliance on platforms like Facebook for peer surveil-
lance furthers the risk of exposure and the need to be vigilant. As social
media are intimately tied to both identity and communication, they shape
how we are perceived and how respondents interact with others. The risk of
unwanted exposure and other social harm necessitate a care of the virtual
self, “whereby citizens are encouraged, enticed and occasionally compelled
into bringing components of their fractured and dispersed data double into
regular patterns of contact, scrutiny and management: (Whitson and Hag-
gerty, 2008, 574).
New risks emerge as a result of offloading social processes online. Whit-
son and Haggerty identify initiatives by the financial, law enforcement, and
government sectors to compel individuals to mitigate the risk of identity
fraud. Their prescriptions amount to a generalized project whereby “data
management practices” are located at “the forefront of personal projects of
responsibilized risk management” (ibid., 587). Whereas identity fraud pri-
marily concerns the individual’s financial well-being, personal reputation is
at stake on Facebook. That such self-care would emerge on social media is
not surprising, as Foucault notes that self-care constitutes “an entire activity
of speaking and writing in which the work of oneself on oneself and com-
munication with others were linked together” (1988, 51). Thus, self-care as
an individual project is both devoted to and expressed through social ties.
This care is manifest through the monitoring and more general control over
information flows, a feature that is congruent with the extent to which care
and control intersect in individual and institutional surveillance practices
(Lyon, 2001). Respondents are sorting out responsible use among them-
selves, yet this is an occasionally contradictory project.
Peer monitoring on social media is a product of the domestication of
media technology. Domestication literature (e.g. Silverstone and Haddon,
 64 Daniel Trottier
1996) recognizes technology as a lived experience and anticipates a quali-
tative focus on respondent reception of social media for peer surveillance.
Domestication addresses how information and communication technologies
are integrated into everyday life, with a focus on the tensions that emerge
in consequence. The tensions between familiar and unexpected conditions
(ibid.) are observable on a comparatively novel and dynamic platform where
interpersonal relations are mediated. As she was discovering how Facebook
would aggregate user activity—including her own—and display it on the
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

platform’s welcoming page, one respondent noted that she was “freaked
out” (Educ1W). Coming to terms with how Facebook displayed personal
information, she noted “all of a sudden, you can see what other people see.
It’s like you’re there and you know that when other people log in, that’s
what they see and it just brings it all full circle and it’s like ‘Wow, Facebook’s
creepy’” (ibid.). Social media reconfigure familiar relations through unfa-
miliar and often unwanted kinds of exposure as a result of its accumulation
of personal information and social ties. Also, following Whitson and Hag-
gerty, reputations on social and digital media are understood by users as an
individual responsibility, while in practice these exceed individual control.
The following sections are arranged thematically in order to chart respon-
dents’ exposure to, use of, and familiarity with Facebook. The first section
explores how as a result of peer pressure and convenience, joining Facebook
became a de facto decision. The next section addresses how respondents came
to realize that visibility was central to their interpersonal use of Facebook.
They described these practices through terms like “creeping” and “stalking,”
which are described in the third section. The fourth section considers how
these experiences lead to a reconsideration of privacy and publicity. The fifth
section considers how respondents developed a set of tactics to manage more
problematic forms of exposure. The sixth section addresses how respondents
regarded other users as being responsible for caring for their own virtual
selves but also acknowledged the difficulties involved in this task.

TIES THAT BIND: PEER PRESSURE AND CONVENIENCE

Many respondents felt compelled to join Facebook due to peer pressure,

either from new friends at university or high school classmates. On occa-
sion peer pressure went beyond recommendation or prescription. Some
respondents had friends who constructed profiles on their behalf and then
transferred control of this nascent profile to them. In more extreme cases,
non-users discovered Facebook contained photos and other content about
them. This perplexed respondents, who were not familiar with the kinds of
visibility in which they were implicated:

I was like, “How can pictures of me be on the Internet?” They would

show me and they’d show a picture of me with other people and
 Caring for the Virtual Self on Social Media 65
they’d show comments that other people had made and, like, “How
are people making comments on this photo of me and I don’t even
know it’s there? I should be involved in this. I should know what’s
going on.”
(Artsci2W)

This respondent realized they already had a presence on Facebook, even

if they never intended to join the site. This presence was only manageable by
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

becoming a Facebook user. Respondents joined Facebook as a result of their

peers’ influence, and they decided to maintain a presence because of its social
convenience. Facebook was often used for on-campus coordination, like
announcing one’s presence in the library or requesting class notes. In other
cases, it was used for a more spectacular kind of visibility, such as upload-
ing photographs from a recent vacation. Facebook was a mix of memorable
events coupled with mundane details, and respondents recognized it as a
resource for sociality. By allowing respondents to upload content for no
explicit audience, Facebook enabled a visibility between users that extended
beyond what they would typically know about each other. This provided
peers with “a general feeling of how your life is going” (Artsci3W). As sub-
mitting personal information was a more pervasive activity, Facebook pro-
duced a visibility akin to a ‘general feeling’ of individuals. This marked the
familiar side of visibility on domestic technologies (Silverstone and Haddon,
1996), as Facebook was mundane and consensual and could be integrated
into peer relations.
Some respondents actively disliked Facebook, going so far as to deactivate
their profiles. These respondents invariably returned to Facebook, though
their absence only strengthened their criticism of the site. Their return to
a service they disliked was fuelled by the perceived need to stay online.
Respondents equate being off of Facebook with being cut off from peers and
social events. Nobody directly prevented them from leaving, but they noted
clear costs associated with abstention:

It’s my damn friends, man. Like, there’s Homecoming and there’s

a bunch of people coming up to Homecoming and [for] a lot of my
friends, Facebook is the only reliable way to get a hold of them. Which
is annoying. It’s really annoying. Because I don’t really want to be on it
and I probably—like, after university’s done, I probably won’t be. But I
don’t know, right now it’s just—it’s almost too convenient not to have it.
(Artsci4M)

Staying on Facebook was seen as “too convenient,” and leaving it was too
costly in terms of the loss of connectivity with peers (cf. Chang and Chen,
2008). Thus, respondents chose to join and maintain a presence because
of social ties with friends, even if they have a negative opinion about the
platform itself. To be sure, this peer connectivity is presented as a desirable
 66 Daniel Trottier
feature. However, this connectivity—the manner in which Facebook diffuses
information—is also a source of unwanted exposure for respondents.

FACEBOOK PRESENCE AS EXPOSURE

Respondents presented Facebook as a vast and accessible resource for

personal information. In describing their visibility on this platform, they
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

reported feeling as though they were being watched by others. In addition

to former and current romantic partners, respondents were concerned with
parental scrutiny. Parental access to profiles led to a convergence between
social contexts respondents wanted to avoid (boyd, 2008). Facebook was
seen to provide a distorted representation of the person. One respondent
described how Facebook misrepresented him to his father:

My dad is on Facebook. He knows what I’m up to. He knows the she-

nanigans that happen. But at the same time, he doesn’t see that I put in
like twenty hours a week at the library along with fifteen hours a week
of classes. He doesn’t see that I’m working every weekend on essays and
stuff. He sees the guy holding the red plastic beer cup.
(Artsci1W)

The above user makes reference to several social contexts in which he is

engaged and notes that Facebook provides a selective account of these con-
texts. We may presume that the photographs of “shenanigans” are shared
among his university peers and that this is a less problematic exchange for
him. However, this becomes unwanted exposure when this content—and
only this content—is made accessible to his father.
Peer monitoring on Facebook is also mutual, as watching and being
watched both feature prominently. Respondents turned to Facebook to dis-
cover more about people who were of interest to their romantic lives. This
is justified by the availability of relevant information on profiles, including
sexual orientation, relationship status, the kind of relation the user wishes
to pursue, and conversation topics. Respondents described their own expo-
sure as a distinct concern from watching their peers. They held different
standards between what they wanted to expose about themselves and what
they wanted to find out about others. Yet when considering their own vis-
ibility, their ability to access their peers’ personal information was the most
accessible yardstick:

I find it kind of weird that I have such insight into her life and I can,
like, basically judge her, like I don’t know, it’s really weird to me. That’s
why I don’t really use it that much, because I feel like all these people
can do the same thing.
(Artsci4W)
 Caring for the Virtual Self on Social Media 67
Peer visibility led respondents to guard their own information. Knowing
they were being watched and knowing the extent to which they could watch
others compelled respondents to monitor their online presence for content
they believed others would find objectionable. This kind of self-scrutiny trig-
gered the strategies for impression management detailed later in this chapter.

“CREEPING” AND “STALKING” ON FACEBOOK

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Two terms associated with peer surveillance on Facebook warrant explora-

tion: ‘stalking’ and ‘creeping.’ Both describe problematic ways of getting
information on Facebook. In the interviews, creeping was seen as a milder
version of stalking, which in turn was meant to reflect more negatively on
the respondent. Creeping was a more involved and targeted way of using
Facebook, though respondents treated this as a matter of circumstance: “It’s
all a matter of degree. I mean, if you were looking to assign either Facebook
stalking or Facebook creeping to one person’s activities, I’d think you would
have to do it on a case-by-case basis” (Health3M). Creeping involves perus-
ing content: a few pages of wall posts or a photo album. It is also presented
as function using the site in the way it was intended, which in consequence
amounts to a prolonged scrutiny of other users’ information. Creeping can
also be brought on due to boredom or simply because content on a user’s
news feed—the content that first greets people when they sign on—caught
their attention.

I guess most people would define the stalking one as more, like, actively
searching, but I guess creeping would be if you’re just bored and you’re
just looking at people’s profiles and you don’t have an active interest or
motive to look at it.
(Artsci4W)

The fact that creeping is fuelled by a lack of purposive motivation (i.e.

boredom) and consists of viewing readily available information suggests
that this behaviour is anticipated and reinforced by the way Facebook is
configured. By “looking at people’s profiles” in the absence of an “active
interest,” the respondent who engages in creeping is making use of the way
Facebook renders personal details visible to its user base.
Stalking resembles creeping, although it is more purposive and “a little bit
more aggressive” (Artsci4W). If a user consistently returned to a particular
user’s profile, this would be framed in terms of stalking. As Facebook stalk-
ing is not restricted to those who have been apprehended and prosecuted
as stalkers, respondents suggested the site facilitates this kind of behaviour
from its users. Much as Shirky (2008) claims emergent media lower the
threshold on group activity, social media facilitate peer monitoring by virtue
of how they organize interpersonal exchanges.
 68 Daniel Trottier
Terms like ‘creeping’ and ‘stalking’ imply some forms of exposure on
Facebook are more troubling than others. Yet in general, respondents grew
accustomed to this range of visibility, illustrating a tension between accepted
and unwanted kinds of exposure on social media. One respondent stated
having his personal information accessible to a network of peers had conse-
quences he was only beginning to realize and accept:

I’ve had like a few people phone me and I’d say, “How’d you get my
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

phone number?” “Oh, yeah I got it from Facebook, sorry.” “Okay,

yeah, no that’s totally fine.” Because I put it up there, right? I put it up
there for a reason.
(Artsci2M)

After he had uploaded personal details, including phone numbers, to his

profile, other people were able to use his information in ways that he did not
anticipate initially. Such experiences provided respondents with a general
impression of their visibility on Facebook. These incidents lead to greater
self-scrutiny and management of their presence but also a level of comfort
with their exposure. Visibility between users ranges from casually discover-
ing what a close friend did over the weekend to the targeted and prolonged
monitoring of strangers. This range of practices is unique in comparison
to earlier forms of online digital media, as they emerge seamlessly from a
localized everyday context. For this reason, many respondents suspected
Facebook was designed specifically for activities like ‘stalking’ and ‘creep-
ing.’ As a domesticated media technology, Facebook itself creeps from the
familiar to the objectionable.

MAKING SENSE OF A PRIVATE/PUBLIC DISTINCTION

In its simplest terms, privacy is a value that is endangered by this expo-

sure on social media. Scholars and respondents both approach social media
surveillance by way of privacy, yet these sites underscore the complexity
of understanding and maintaining privacy (Nippert-Eng, 2010). Other
research on digital youth suggests youth reconsider while actively managing
their privacy on social media (boyd, 2008; West, Lewis and Currie, 2009).
These users were experienced in maintaining a degree of privacy from par-
ents prior to social media, and these experiences inform their use of sites like
Facebook. Yet they do not simply hide information from some peers; they
actively share it with others. In doing so, they maintain a trade-off between
ensuring privacy and achieving public exposure when placing their personal
content on Facebook (Tufekci, 2008; boyd and Hargittai, 2010).
Respondents made reference to the terms ‘private’ and ‘public’ when
discussing Facebook. Not only was a public/private binary insufficient, but
they framed Facebook as a catalyst to reconsider these values. Facebook
 Caring for the Virtual Self on Social Media 69
represented a kind of blurring of private and public, at least as generally
understood in North America prior to the advent of social media. In addi-
tion, the term ‘personal’ crosscut these discussions, such that personal infor-
mation appeared in public spaces, private spaces, or both. Facebook was a
public presence where respondents are ostensibly comfortable sharing per-
sonal details with friends they have wilfully chosen. Some respondents, espe-
cially those in their first year of studies, drew parallels between their exposure
on Facebook and the kind of publicity sought by reality TV stars. Clearly one
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

of the key motivations for using Facebook was to share specific information
with a somewhat amorphous (though occasionally specific) audience.

It’s pretty much all about attention, as far as pictures; taking pictures,
things like that—because everyone can see that. (. . .) It’s almost as if
we’re in a world where we watch a lot of reality TV and there’s [reality]
TV shows like The Hills and things like that, Facebook is our own little
form of entertainment where we can get a glimpse of everyone’s life.
(Artsci3M)

Other respondents approached Facebook publicity with trepidation,

making comparisons to other types of exposure to underscore the difficulty
of adjusting to the service. One respondent likened the circulation of pho-
tos on Facebook as if they “had a photo album at my house and some-
body came and copied it and then put it in their photo album” (Artsci4M).
Another claimed the wall was modelled after the whiteboard on students’
doors in residence halls, making it a kind of public space. Respondents also
drew comparisons between putting content on Facebook and being visible
outdoors. However, the implications of this imagery were not unanimous.
While some believed being outside legitimated scrutiny, others still held
expectations against unwanted forms of exposure.
Privacy violations mattered to respondents, notably when their personal
information leaked in unanticipated or undesirable ways. One respondent
who described Facebook as private referred to its extensive privacy controls
yet also acknowledged a public element insofar as the site is open to all:

I’d say it’s private in the sense that like your own profile is like, your
own profile and you can control what’s on it because even if somebody
writes on your wall, you can make it invisible kind of thing. So, it’s
totally private because you can have as much information as you want
or as little information as you want, but then it’s public in the sense that
anyone can use it.
(Health1W)

Most respondents contemplated their privacy and publicity simultane-

ously. One respondent referred to Facebook as “a completely public expres-
sion of private and personal matters” (Artsci4W). This statement suggests a
 70 Daniel Trottier
paradox of values that governed respondents’ use of Facebook. Following
default settings, respondents placed their personal information in a context
that was more public than desired. The fact that this was a deliberate and
wilful act perplexed respondents:

It’s supposed to be personal information but Facebook makes it very

public and you’re supposed to be, I guess, with every post you make,
keeping in mind the fact that everyone can see this and that this is a
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

public sphere for, I guess, private communication.

(Artsci4W)

In suggesting that a typical user ought to “keep in mind” the extent to

which personal details on Facebook are public, this respondent was coming
to terms with potential exposure on this platform.

MANAGING ONLINE PRESENCES

Although social media like Facebook offer new opportunities for public
exposure, respondents had a range of tactics at their disposal to manage
their presence. Caring for the virtual self (Whitson and Haggerty, 2008)
involved restricting information flows and exceeded the range of privacy
features offered by Facebook. In general, respondents made extensive use of
privacy settings. They did this to restrict information from a general public
but also for more targeted purposes like hiding a particular photo album
from specific individuals. The majority of respondents were familiar with
privacy settings. Others found them confusing but maintained a commit-
ment to mastering these settings. Based on prior experiences, respondents
periodically returned to their settings to ensure they were still in order. Upon
joining his school’s academic network, one respondent noticed his personal
information had leaked to a degree he did not anticipate. Following this
incident, he appraised his privacy settings “every so often (. . .) to make sure
that only my friends can see my profile” (Artsci3M). One respondent said
she set her privacy such that other users could not seek her out. In effect, she
had to initiate contact with others.
Another tactic employed by respondents was to maintain dynamic and
contextual privacy settings. Many respondents stated their intention to revise
their settings, with one respondent doing this mid-interview. Respondents
augmented their privacy during job searches, such that potential employers
could not locate their personal content. One person went so far as to cancel
his Facebook account temporarily in recognition of how his position as a
political leader could be compromised by his online presence:

I cancelled Facebook for a little while; it was for quite a long period of
time. It was because I was vice-president of the [youth political group].
 Caring for the Virtual Self on Social Media 71
(. . .) I had that leadership position too and, that’s when I realized that
you know, this is my personal life but, it’s publicly . . . like my public
personal life, that could reflect negatively on the [youth political group].
(Artsci1M)

Other respondents logged onto their friends’ accounts to see the extent
to which they were visible. By looking at their presence from another user’s
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

perspective, they had a better sense of how their presence was perceived.
Facebook later integrated this feature into their privacy settings, an acknowl-
edgment that peer visibility matters to respondents.
Respondents also cared for their virtual selves by choosing not to upload
certain information. This self-censorship is described along common-sense
lines, as respondents simply did not share content that they believed would
harm their reputation:

Anything I put out there on the Internet—if I’m afraid of it ever coming
back to haunt me, I won’t put it out there. And I consciously make that
decision for every piece of information that I put out. So I’m not really
worried about stalking or creeping because anything out there on me is
kind of what I’ve already put out.
(Artsci3W)
Anything that you wouldn’t want your parents to know isn’t something
that should be on the Internet.
(Artsci4W)

Respondents cared for their virtual selves by omitting or removing dam-

aging content. Parents and stalkers were invoked as potential watchers and
thus as justification for self-care through self-censorship. The latter included
not uploading information but also not behaving in a way that could be
photographed or otherwise documented:

For me to be caught on photo doing something stupid, I had to be doing

something stupid in the first place. And if I avoid that, which I have been
hit or miss about in the past, then it’s a non-issue. They can’t post photos
of me that didn’t happen.
(Health3M)

Many respondents monitored the content their friends posted about

them. These respondents objected to some wall posts and photos authored
by friends, deleted these posts, and scrutinized their friends. In extreme cases
respondents also pruned their social network by removing users from their
friend lists who posted problematic content on their profiles. Following
Foucault (1988), care for the virtual self on social media was manifest as a
combination of reflection on risk and visibility, as well as routine activities
 72 Daniel Trottier
performed in order to manage that exposure. Yet some respondents also
suggested they could not know the full extent of their virtual self on social
media and that their efforts to manage their reputation were insufficient.

RESPONSIBILITY AND FUTILITY

Respondents felt responsible for their exposure on social media and extended
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

this sense of responsibility to others. In the event of privacy violations or other

unwanted consequences, respondents believed other users only had them-
selves to blame. This suggests a perceived locus of control by respondents
when users upload information to Facebook: “you can exercise so much con-
trol over what’s visible, it’s boggling to me, honestly. If I didn’t want people
to see my profile, I’d make it private, and that would be that” (Health3M).
Because of this perceived sense of agency, respondents had little sympathy for
users who encountered trouble with exposure. Respondents claimed that by
uploading information onto their profiles, users were “inviting people to look
into their life” (Artsci3M). Such users were seen as deliberately uploading
information about themselves in a public setting and then coping with con-
sequences they presumably should have anticipated. Respondents did treat
their own creeping and stalking of others as problematic, but self-judgement
was tempered by the perception that other users are responsible for the care
of their virtual selves. As a result, they justified their lateral surveillance of
others by citing the other user’s decision to upload this information or their
failure to use more stringent privacy settings: “If I’m looking at it, I feel like
if she has it public, then I can just look at it and I don’t feel bad” (Artsci4W).
Respondents felt responsible for managing their online visibility, all while
acknowledging the challenges associated with this duty. Attempts to man-
age privacy are often case based; that is, to stop specific information from
circulating in a particular context. Many respondents believed information
would still leak beyond an intended audience. The sheer volume of infor-
mation as well as social ties complicates an online presence. Respondents
claimed simply having information on a profile, whether public or private,
left users open to considerable risk because of the growing number of indi-
viduals who would have access to that information: “I guess we all tend
to forget is that what we put on Facebook isn’t personal or private in any
means because of the hundreds and sometimes thousands of people that you
allow to see your profile” (Artsci4W). This respondent compared this unex-
pected exposure to the ‘reply all’ button on email interfaces, where users
accidentally send a message meant for one person to an entire community.
Respondents also described the non-friend friend (a Facebook friend who is
a stranger in other social contexts) as a potential vulnerability:

I think just the people who you kind of add, maybe two or three years
ago, who you’ve kind of forgotten about. They’re still on your friends
 Caring for the Virtual Self on Social Media 73
list but you may not actually be friends with them in real life and you
may not see them ever or talk to them ever, but you still have access to
their profile and they still have access to yours.
(Artsci4W)

The quantity of seemingly trusted friends prevented respondents from

managing their entire audience. Even when respondents placed the locus
of control—and blame—on individual users, they acknowledged Facebook
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

was primarily a public domain, and attempts to limit exposure were futile.
Social media are thus presented as domestic technologies through which
personal information leaked in unanticipated directions. Respondents were
aware that proper conduct on Facebook is based on a contradiction: Indi-
vidual users are expected to be vigilant, but this vigilance will not offset all
potential risk. For these respondents, safe use of social media was necessary
but not sufficient to prevent unwanted exposure to peers and strangers.
The above suggests an acknowledgement of partial futility: If information
is uploaded, it may leak into unanticipated contexts. Indeed, this condition
has likely become more pronounced as Facebook’s user population contin-
ued to grow and as more institutions began to treat the site as a source of
information. Respondents viewed the consequences of Facebook exposure
as a concern, because the platform took on a greater presence in various
social contexts:

Like, Facebook is so new, like, we don’t know what kind of social impli-
cations it’s going to have. And it’s becoming such a momentous force
that I don’t think, like, it’s like people jumped on board before they
knew where it was going.
(Artsci4M)

As this exposure was rooted in the everyday use of social media, the cor-
responding self-care was embedded in mundane practices like un-tagging
photos and choosing not to upload damaging content. These risks were tied
to information flows between previously distinct contexts, and respondents
used privacy settings extensively to ensure these leaks were kept to a mini-
mum. However, some respondents were overwhelmed by Facebook’s oppor-
tunities for public exposure and cited this as a reason to not be diligent:

I’m pretty sure all my private information is already long gone. There’s
no sense of privacy in this modern world and because everything I do
is basically online these days, I feel like there’s little or no safety and,
therefore, I don’t need to curb what I’m doing on Facebook.
(Artsci4W)

This respondent was ambivalent about her comfort with Facebook. She
grew accustomed to being visible, but this was in response to the overwhelming
 74 Daniel Trottier
exposure on the site. She actively participated in her visibility, but this was
not necessarily the kind of social media she wanted. From her perspective,
care for the virtual self was understood though—and overshadowed by—a
perception that her control of her personal information had already been
compromised.

CONCLUSION
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

At the time of data collection, respondents were developing a growing famil-

iarity with interpersonal visibility on Facebook. However, this familiarity
was mixed with uncertainties and tensions. Respondents joined at the behest
of their friends and provided personal information to Facebook as a way
to communicate with these friends. The kind of visibility required from
Facebook was thus borne out of everyday practice, with respondents main-
taining relations with their colleagues. In order to maintain social ties with
peers, they willingly supplied information about themselves to this platform,
making it suitable for interpersonal scrutiny. An ever-growing friendship
network led to unanticipated risks, but a sense of agency and responsibility
was placed on users themselves despite this complexity. Tensions between
individual control and greater complexity on social media, much like ten-
sions between private life and public exposure, are indicative of the further
growth and domestication of social media (Silverstone and Haddon, 1996).
As a central feature of their everyday life, respondents framed social media
exposure as less of a violation and more of a condition that they need to
manage. Visibility and exposure on Facebook were normalized and were
not limited to any specific instance but were instead framed as a pervasive
condition of social life on social media.
Respondents expressed some ambivalence in their responses. They grew
accustomed to Facebook visibility, yet its effects remained chilling. These
developments complicated efforts to manage their online presence. Respon-
dents felt responsible for their presence but were aware that managing this
presence was beyond their control. They perceived online reputation as
being a personal responsibility, even while acknowledging that caring for the
virtual self by way of vigilant self-scrutiny is not enough. They were aware
different audiences and social contexts intersect on the site, but they still
constructed a visible profile for close friends. This is emblematic of interper-
sonal visibility on Facebook: An online presence is built from mundane con-
tributions by users and their friends. Respondents join because their peers
are online and build a presence to remain visible to these friends. No single
act seems risky or malicious. However, when taken together over time, they
contribute to exposure that may lead to social harm. The risk of unwanted
visibility and social harm becomes especially salient at a time when Face-
book’s user population is sharply increasing, with a consequential increase
of contributors to any user’s visibility, audiences of that visibility, and social
 Caring for the Virtual Self on Social Media 75
contexts in which that visibility will have consequences. A photograph that
was uploaded when there were a million Facebook users becomes all the
more important when its potential audience exceeds one billion users.
This chapter has focused exclusively on interpersonal visibility and moni-
toring on social media. Respondents expressed their concerns on this topic
in terms of their relations with friends, family, and other known individuals.
They were generally concerned with situated and immediate forms of sur-
veillance. Respondents wanted to maintain boundaries that separated dif-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ferent social contexts. From their perspective, these concerns eclipsed other
kinds of scrutiny conducted through Facebook. Subsequent research should
consider these other kinds of social media monitoring. As law enforcement
branches, marketers, employers, and governments take a continued interest
in sites like Facebook, the visibility produced by interpersonal social media
surveillance will undoubtedly augment the scope and capacity of other kinds
of social media surveillance. This research is also limited to digital youth
in a university environment. While this provides specific context for user
concerns, subsequent research should focus on how other populations are
coping with social media platforms and their conditions of intervisibility.

REFERENCES

ALBRECHTSLUND, A. (2008) Online Social Networking as Participatory Surveil-

lance. First Monday, 13(3). Accessed August 16, 2011. http://firstmonday.org/
htbin/cgiwrap/bin/ojs/index.php/fm/article/view/2142/1949
ANDREJEVIC, M. (2005) The Work of Watching One Another: Lateral Surveil-
lance, Risk, and Governance. Surveillance & Society, 2(4). p. 479–497.
BOYD, D. (2008) Facebook’s Privacy Trainwreck: Exposure, Invasion, and Social
Convergence. Convergence: The International Journal of Research into New
Media Technologies, 14(1). p. 13–20.
BOYD, D. & HARGITTAI, E. (2010) Facebook Privacy Settings: Who Cares? First
Monday, 15(8). Accessed August 16, 2011. http://firstmonday.org/htbin/cgiwrap/
bin/ojs/index.php/fm/article/view/3086/2589
BRIGHENTI, A. M. (2010) Visibility in Social Theory and Social Research. Hamp-
shire: Palgrave Macmillan.
CHANG, H. & CHEN, S. W. (2008) The Impact of Customer Interface Quality, Sat-
isfaction and Switching Costs on E-Loyalty: Internet Experience as a Moderator.
Computers in Human Behavior, 24. p. 2927–2944.
DEWING, M. (2010) Social Media: 2. Who Uses Them? Parliamentary Information
and Research Service. Accessed March 4, 2012. www.parl.gc.ca/Content/LOP/
ResearchPublications/2010–05-e.htm
FOUCAULT, M. (1988) The History of Sexuality, Vol. 3: The Care of the Self. New
York: Vintage.
KOSKELA, H. (2004) Webcams, TV Shows and Mobile Phones: Empowering Exhi-
bitionism. Surveillance & Society, 2(2/3). p. 199–215.
LYON, D. (2001) Surveillance Society: Monitoring Everyday Life. Buckingham
(UK): Open University Press.
MURAKAMI WOOD, D. & WEBSTER, C. W. (2009) Living in Surveillance Societ-
ies: The Normalisation of Surveillance in Europe and the Threat of Britain’s Bad
Example. Journal of Contemporary European Research, 5(2). p. 259–273.
 76 Daniel Trottier
NIPPERT-ENG, C. (2010) Islands of Privacy. Chicago (IL): University of Chicago
Press.
OMAND, D., BARTLETT, J. & MILLER, C. (2012) Introducing Social Media Intel-
ligence (SOCMINT). Intelligence and National Security, 27(6). p. 801–823.
SAX, L. J., GILMARTIN, S. K. & BRYANT, A. N. (2003) Assessing Response Rates
and Nonresponse Bias in Web and Paper Surveys. Research in Higher Education,
44(4). p. 409–432.
SHIRKY, C. (2008) Here Comes Everybody: The Power of Organizing Without
Organizations. New York (NY): The Penguin Press.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

SILVERSTONE, R. & HADDON, L. (1996) Design and the Domestication of Infor-

mation and Communication Technologies: Technical Change and Everyday Life.
In Silverstone, R. & Mansell, M. (eds.) Communication by Design: The Politics
of Information and, Communication Technologies. Oxford: Oxford University
Press, p. 44–74.
TROTTIER, D. (2011) A Research Agenda for Social Media Surveillance. Fast Capital-
ism, 8(1). Accessed March 4, 2012. www.uta.edu/huma/agger/fastcapitalism/8_1/
trottier8_1.html
TUFEKCI, Z. (2008) Grooming, Gossip, Facebook and MySpace: What Can We
Learn About These Sites From Those Who Won’t Assimilate? Information, Com-
munication, and Society, 11(4). p. 544–564.
WEST, A., LEWIS, J. & CURRIE, J. (2009) Students’ Facebook ‘Friends’: Public and
Private Spheres. Journal of Youth Studies, 12(6). p. 615–627.
WHITSON, J. & HAGGERTY, K. (2008) Identity Theft and the Care of the Virtual
Self. Economy and Society, 37(4). p. 572–594.
 4 Shaping Children’s Consumer
Identity Within Contemporary
Dutch Market(ing) Practices
Isolde Sprenkels and Irma van der Ploeg
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

Contemporary marketing communication and research practices are increas-

ingly designed with children’s digital technology use and online activities
in mind. Often as part of a cross-media approach, market practitioners1
employ such digital means in order to connect with children. The online
presence they establish for corporations and brands includes banner adver-
tisements, different forms of gamevertising, and branded social networking
profiles, viral video advertisements, brand presence within online virtual
worlds, and websites aimed at children, related for instance to food, bever-
age, and toy brands or television programmes. Such presence offers ways
to meet corporate goals such as building brand awareness and stimulating
product purchase. It also enables corporations to monitor children and col-
lect valuable information about them, turning children’s online playgrounds
into sites for detailed surveillance (Chung and Grimes, 2005; Steeves, 2007).
As Kathryn C. Montgomery and Jeff Chester claim,

“the interactive nature of digital technologies makes it possible for mar-

ket research to be woven in the content of new media, offering mar-
keters the opportunity to remain in constant contact with teens and
creating a feedback system for the refinement of marketing techniques”
(Montgomery and Chester, 2009, S19).

This co-shaping of contemporary marketing communication and research

is central to this chapter. We will unfold ways in which the ‘Dutch kids’ mar-
ket’ and children’s identities as consumers are constituted within a configu-
ration of such practices by re-describing them in order to disclose some of
the socio-material mechanisms at work. For this re-description, a practice-
based approach to markets and marketing is used (Araujo, 2007; Kjellberg
and Helgesson, 2007), which draws from the works of Bruno Latour and
Michel Callon. Inspired by Latour’s first rule of method, that we should
“study science in action and not ready-made science or technology” (Latour,
1987), Hans Kjellberg and Claes-Frederik Helgesson claim that it is impossi-
ble to define a list of properties that are typical of markets and that therefore
 78 Isolde Sprenkels and Irma van der Ploeg
markets should be studied in the making (Kjellberg and Helgesson, 2007).
Markets are not fixed entities to be found or discovered, and as ‘social-
theory-inspired’ perspectives on identity claim, neither are identities2 (Van
der Ploeg, 2007; Van Zoonen, 2013). A practice-based approach suggests
that market practices contribute to the continuous constitution of markets
and consumer identities; they have a “performative3 role in helping to create
the phenomena they purportedly describe” (Araujo, 2007, 211).
In this chapter, we adopt the model Kjellberg and Helgesson introduce,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

which presents markets as continuously realized, as the ongoing results

of three strongly interlinked types of practices: representational practices,
which serve to portray markets and how they work; exchange practices,
which serve to realize economic exchange of goods and services or transac-
tions; and normalizing practices, which serve to establish rules and norms
concerning market behaviour, how a market should work (2007). Interest-
ingly, they use the concept of translation, the way something spreads and
transforms across time and space (Callon, 1986; Latour, 1987), as a way
to address how the three practices are entangled as they are linked through
chains of translations, which themselves are brought about by those prac-
tices. It is these links that are considered to be central to the process through
which market practices constitute markets and consumer identities, and we
closely follow these links within the two sections of this chapter.
In the first section, I describe two representational market practices using
a selection of quotes from semi-structured interviews with market practi-
tioners and presentations held at the annual Dutch conference on kids and
youth marketing.4 Market practitioners constantly seek representational
conceptions of children in order to tune their marketing communication
strategies with these conceptions and optimize certain results. By employing
such practices serving to depict markets (Kjellberg and Helgesson, 2007),
they seek what they call ‘insights’ in matters such as children’s development,
behaviour, interests, preferences, consumption habits, brand experiences,
media use, or leisure activities. Or, as Daniel Thomas Cook explains, they
seek to ‘know’ children (2004, 2011). In order to ‘accurately’ conceptualize
children, market practitioners appear to build upon academic theories on
children’s (cognitive) (consumer) development and conduct their own mar-
ket research in which children’s life worlds are studied and translated into
‘insights’ and models representing children’s abilities, interests, ‘needs, and
wants.’ These two practices help market practitioners characterize and cat-
egorize children, dividing them up in to smaller and more specific ‘manage-
able’ groups or segments. We use Geoffrey Bowker and Susan Lee Star’s work
on categories and knowledge production (1999), together with John Law’s
work on the performativity of (social) research methods (2009; Law et al.,
2011), to describe ways in which these two representational practices depict
Dutch children and the Dutch kids’ market. Such an approach challenges
the basic assumption that there are child consumers ‘out there’ in an exist-
ing and clearly defined ‘kids’ market’ which can be described with specific
 Shaping Children’s Consumer Identity 79
characteristics (Harrison and Kjellberg, 2010) and helps illustrate instead
that they are constituted within these very same practices employed to ‘get
to know them.’
In the second section, we provide a case study of market exchange prac-
tices. We describe how two commercial online environments for children are
designed in such a way that they attempt to meet particular corporate goals,
ranging from stimulating brand awareness and building a ‘personal relation-
ship’ between brand and child to the stimulation of product purchase and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

consumption and collecting personally identifiable information. We demon-

strate that their design is informed by both representational practices. Build-
ing on what we have seen market practitioners do as they characterize and
categorize children, we follow the ways in which that obtained ‘knowledge’
or those ‘insights’ are materialized, translated, and inscribed (Akrich, 1992;
Bowker and Star, 1999; Oudshoorn et al., 2004) into these two contempo-
rary (digital) marketing communication strategies. Inspired by Madeleine
Akrich’s concept of script (1992), we argue that representational practices
are translated into age-specific scripts, which are attributing and delegating
specific competencies, actions, and responsibilities to children and are pres-
ent within or translated into exchange and normalising practices, the latter
establishing particular norms, rules, and codes concerning the Dutch kids’
market. We demonstrate that developmental psychology–inspired insights
and the ones retrieved from market research practices help design exchange
practices in such a way that they enable marketers to reach particular corpo-
rate goals. We further reveal that both insights are used to legitimize exchange
practices as well as representational practices themselves, which, as we argue,
holds some serious misapprehensions.

DEPICTING ‘CHILD CONSUMERS’

Market practitioners claim that the ‘kids’ market’ is a difficult and even a vol-
atile one to fully understand (Buckingham, 2007). These practitioners con-
tinuously seek representational conceptions of children, ‘knowledge’ about
their presumed abilities, interests, ‘needs, and wants,’ in order to tune their
marketing communication with these ‘insights.’ These are based on ways in
which they translate (cognitive) developmental theory into various age-based
classification models and on particular market research strategies, employed
to study and translate children’s life worlds into valuable ‘insights.’

Age-Based Categories
Gender, and many other variables such as ethnicity or socio-economic status,
can be used to divide children up into ‘target’ groups, but market practitio-
ners first and foremost divide children up by age. Many of the practitioners
we interviewed and who presented at the conference appear to base their
 80 Isolde Sprenkels and Irma van der Ploeg
categorizations on (cognitive) developmental psychology.5 Their practices
rely heavily on developmental stages, relating to children’s presumed cogni-
tive abilities. Such a take can be traced back to the seminal work of devel-
opmental psychologist Jean Piaget (1954), who understood childhood as
a patterned, predictable sequence of stages (Cook, 2004). Various practi-
tioners presenting at the conference on kids and youth marketing talked
about the need to focus on child development from a variety of academic
disciplines. Even though they seemed to lump them all together, their work
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

reveals similarities with developmental-stage models. The following para-

graphs demonstrate ways in which they have translated and transformed
such models into their own age-based classification models.
Vic,6 the commercial director of a Belgian (television) entertainment
company popular in the Netherlands, explains in the following quote that
at his company, they classify children based on developmental stages and
subsequently consider how they need to act upon them. Their classification
helps answer questions about how to design their marketing communica-
tion strategies and what types of products and merchandising they need to
manufacture for these age-group categories:

The target group between 0 and 15 does not exist . . . There are clear
developmental stages, we conducted research together with [a research
agency] . . . 2–3 year olds focus on themselves . . . copy others . . . develop
eye hand coordination, 4–5 year olds are more explorative . . . 6–8 year
olds fantasize and understand more complex stories . . . 9–12 year olds
peer pressure is important . . . how do you aim for that as a brand,
show individual children or groups in our commercials? 13–15 is more
analytical . . .
What types of merchandising do you have to make for which age-
category, what types of toys do you have to make for which age-category?
(Presentation 4, ECX, 2012)

This quote illustrates that for Vic, ‘the kids’ market’ is not a homoge-
neous group of children but consists of separated age-group categories.
These categories are created together by market actors such as Vic, the Bel-
gian (television) entertainment company, and the marketing communication
research agency.
Although Vic talks about ‘clear’ developmental stages, other presenters at
the conference spoke of slightly different age groups and mentioned slightly
different abilities and interests to go with these stages. A couple of weeks
after the conference, we conducted an interview with Cherie, the owner of a
kids and youth communication research agency. Cherie responded to Vic’s
conference presentation, and claimed, “We later went up to [(television)
Entertainment Company] and said that they were wrong, that they had
added two years to every category” (Interview 8, YCRA-3, 2012).
Cherie is working on a model “for the whole development of children”
herself. During her presentation at the conference, she explained this model.
 Shaping Children’s Consumer Identity 81
She suggests that between 4 and 6, children are in a magical phase in which
the difference between fantasy and reality is still blurry. Between 7 and 9
they are in a more realistic phase in which they develop friendships, and peer
influence is increased. And between 10 and 12, which she calls ‘tweens,’ a
tension is experienced between becoming an adult and being a child7 (Pre-
sentation 7, YCRA-3, 2012). Cherie claims that the entertainment company
added 2 years to every category, but after doing the maths, her claim doesn’t
seem to match the differences that can be distinguished between the two
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

age-based classification models of Vic and Cherie. Her remark, being precise
in demarcating age ranges, illustrates a literal and very strict application of
stages, while Piaget, for instance, actually stressed the succession of stages
and not exact age-ranges (Ville and Tartas, 2010).
There were also market practitioners who expressed experiencing difficul-
ties relating to children’s abilities and interests within a particular category.
As Deniz, owner of a kids and online digital production agency, mentions:

Some kids’ marketers use the phrase ‘tweens’ to refer to the category of
10–14 year olds. But there are problems with this category: children up
to 12 years of age stick to what they know. But when they are 12 years
and over, they go and look for new things.
(Interview 10, DPA, 2011)

Deniz explains that there are problems with the ‘tween’ category itself, as
according to her own experience, children between 10 and 12 years of age
and children between 12 and 14 years of age differ strongly. This, according
to Deniz’s understanding, would lead to ‘problems’ when market practi-
tioners want to target this tween category, as they will fail to speak to all
children within this category due to the strong differences between them.
The existence of another tween age-range—Cherie, for instance, referred to
10- to 12-year-olds as ‘tweens’—confirms the varieties in age range, indicat-
ing a lack of consensus between practitioners.
Other practitioners point at differences between age-category concepts
among distinct disciplines and argue that classification models and their
categories change over time. Mischa, a senior researcher with a kids and
youth communication research agency, explains:

Tweens is a marketing concept I think because in psychology we call the

10–14 year olds early adolescence . . . When I started working in the
kids’ market back in the nineties, we approached the 6–12 year olds as
one group. . . . Then, there was a switch, children getting older younger.
(Interview 7, YCRA-2, 2012)

Before working as a researcher with agencies in this field, Mischa con-

ducted scientific research at a university. Both her knowledge of develop-
mental psychology and her experiences as a market practitioner inform her
comments. According to Mischa, in the 1990s, ‘tween’ was a non-existent
 82 Isolde Sprenkels and Irma van der Ploeg
category. Such a category is constituted into being, as Natalie Coulter found
in her work on the discursive construction of the tween girl market (2009).
Mischa connects the rise of the tween category with the phenomenon
of “children getting older younger.” For market practitioners this is bet-
ter known as KAGOY: Kids Are Getting Older Younger. Sociologist Juliet
Schor refers to this phenomenon as ‘age compression’ and explains that
market practitioners offer products originally designed for older children
and target them to younger ones (2005). David Buckingham states that
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

children are “constantly calibrating themselves in terms of age: other

social institutions, most obviously the school, help to define what it means
to be a child of a particular age” (2007. 20). But in practice, they often
perceive or act out on it in far more complicated ways (ibid.). Indeed, as
mentioned, sometimes they aspire to consume things that appear to be
targeted to older children, but on the other hand, for some older children,
products and services targeted at younger children are appealing, too. This
is illustrated by the following quote from Joss and Esra, brand managers
at a toy company:

Our target group is girls between 6–10, though we see younger girls get-
ting interested in brands for older girls, and children over 10 still like the
brand as well, perhaps not the dolls, but other merchandise.
(Presentation 9, TC, 2012)

This longing to be slightly older is what market practitioners employ

by inscribing it in their practices, their design of products, merchandising,
shops,8 and marketing communication. They serve their commercial inter-
ests when they speak of KAGOY, as they will make more money when their
target market for products and services used by adolescents, which are often
more expensive, becomes larger. Making younger children aspire to those
products and services as well increases profit (Schor, 2005).
Interestingly, in one way, market practitioners are inspired by (cogni-
tive) developmental theory, conceptualising children’s development in a
linear way with predetermined stable sets of age stages. But in another
way, they seem to let go of such rigid, structured, universalistic ideas on
child development, leaving room for change and variation, as for instance
the quote from Mischa shows. As demonstrated, market practitioners have
slightly different ideas on the matter. Some of them come up with their
own categories, different realities so to say, which is a continuous effort
to construct and stabilize their ‘stages.’ This indicates that categories are
contestable constructions, they are contingent, could have been, or are
different according to numerous actors and within diverse settings; they
travel from theory to various practices or models or systems. Just as Geof-
frey Bowker and Susan Lee Star have shown with their work, classification
systems organize and are organized in work practice (1999). They are “a
set of boxes (metaphorical or literal) into which things can be put to then
 Shaping Children’s Consumer Identity 83
do some kind of work—bureaucratic or knowledge production” (Bowker
and Star, 1999, 10).

Empowering ‘Child Consumers’

Market practitioners’ ‘insights’ on children come not only from (self-made)
models of child development. These ‘insights,’ this ‘knowledge’ about chil-
dren’s abilities and interests, also come from other representational prac-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tices, such as their market research and communication concept-developing

practices. There is a large variety of research methods employed in order to
depict children’s life worlds. Schor mentions organizing consumer panels,
test panels, the use of surveys, questionnaires, and focus groups, as well
as conducting interviews and ethnographically oriented observations—
ranging from playgrounds to bedrooms—hiring children as company con-
sultants, and using eye movement tracking and neuro-imaging technology
(2005). With the rise of the Internet and its World Wide Web, new methods
emerged; Internet market research and data mining techniques are increas-
ingly employed within children’s online playgrounds, like the online environ-
ments described in below, covertly collecting various types of information
about children (Chung and Grimes, 2005).
Cofounder of kids and youth communication research agency 4, Jules,
explains how she envisions traditional and new research methods to co-exist
and hints at involving youth directly in research:

Traditional research methods will stay, but I think it will be both. I think
it would be great to say to a particular focus group: download this app
and keep track of the following things this week and you will receive
a new push message every day when we have a new assignment . . . I
think youth would enjoy it very much . . . they like to cooperate with
our research because we focus on them, on what they think . . . we often
say to them that they are really helping us and that we are very happy
with that.
(Interview 9, YCRA-4, 2012)

In line with Cook’s descriptions of a rise in ‘taking the child’s perspective’

in the work of marketing scholars and market practitioners (2004) and the
increased attention in interdisciplinary childhood studies since the 1980s
for children’s subjectivity and agency and involving them as participants in
research (Qvorttrup et al., 2011), calls for the ‘right’ of children to ‘have
a voice’ have also increased in the field of market research (Todd, 2010).
Children are engaged more directly in market research and communica-
tion concept-developing practices, using notions such as ‘giving children
a voice,’ ‘give them influence,’ ‘involve them,’ and ‘start a conversation.’
This empowerment language contributes to market practitioners’ discursive
legitimization of their practices.
 84 Isolde Sprenkels and Irma van der Ploeg
One of the ways in which market practitioners tend to ‘give children a
voice’ is through what is referred to as ‘co-creation.’ The website of kids and
youth communication research agency 4 addresses potential clients of their
agency and explains their ambition of involving youth in their practices:

Do you want to involve youth in your organization, then ask them what
you want to know from them! Do you have a vision, an idea, a plan;
present it to youth. Start a conversation with them, organize a creative
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

session, have them talk to each other, brainstorm. What form you use
does not matter, as long as you involve them. Because what could be
more fun than creating value together, and while talking, exploring and
brainstorming, going from ‘them’ to ‘us’!
(YCRA-4)

This invitation or advice illustrates situating children discursively as the

‘other,’ or ‘them,’ about whom market practitioners want to find out as
much as possible. Natalie Coulter might describe this in terms of the colo-
nization of young people as markets (2012),9 but it also demonstrates the
assumption of ‘joining forces’; of turning ‘them’ into an ‘us,’ a ‘co,’ creating
marketing communication, products, and services together. Who is helping
whom here remains unclear.
Pascal, senior advisor at kids and youth communication research agency 1,
stresses the value of co-creation:

I think your relationship with your target group is very important for
making really good choices . . . co-creation is the core of our think-
ing . . . it allows us to have children design in ways that really ‘fit’ them,
and by which implicitly you can analyse a lot of research data . . . hav-
ing a child [involved in] design is a way to find out what the core behind
that topic is for a child . . . co-creation is helpful to get to the real per-
ception of things.
(Interview 6, YCRA-1, 2012)

According to Pascal, co-creation is a good way to get to ‘the real percep-

tion of things,’ to ‘discover’ the core of what children are like and what they
consider to be important, interesting, or desirable. It also offers important
research data along the way. However, following John Law and his work
on the performativity of methods (2009), there is no such thing as ‘dis-
covering’ and gaining a ‘real perception’ of something ‘out there,’ some
sort of essence of children’s abilities, interests, ‘needs, and wants.’ Instead,
these are produced or enacted within these very practices employed to ‘get
to know them.’10 While conceptualising and categorising children, market
practitioners construct various identities of their target group. These are
constituted into being by re-describing them in their own terms and acting
upon them.
 Shaping Children’s Consumer Identity 85
“Good Work and Good Business Go Hand in Glove”
Market practitioners research the abilities, interests, ‘needs, and wants’
of youth through representational practices in order to take these into
account in marketing communication strategy design. At the same time,
it is a way to legitimize these practices: Practitioners argue to do ‘good
work’ because they meet children’s ‘needs and desires.’ Amy Henry
explained having a similar experience. According to her, US industry lead-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ers claim that “good work [meaning ethical11] and good business go hand
in glove”; they argue “that the same process that allows them to create
effective advertising also ensures that the advertising is ethical and that
age-appropriate advertising is both clearer and more appealing to chil-
dren” (2005, 11).
Meeting children’s ‘needs and desires’ for market practitioners means that
their marketing communication, products, and services ‘add value’ to chil-
dren’s lives. As Pascal, working at kids and youth communication research
agency 1, clearly explained:

Though it is not that I would say that selling candy doesn’t have any
value . . . value is not a judgement about a product, about content, nor
about the services of a client, it is about the desires and needs of youth.
(Interview 6, YCRA-1, 2012)

‘Having value’ is important for market practitioners’ professional ethics,

which is based on normalizing practices, on establishing rules and norms on
how they should act. Jules mentioned:

When it comes to commercial clients . . . This means that if I help them,

the world for youth becomes a little bit better, as it means less annoy-
ing advertising . . . more fit to their needs . . . Of course, it is still selling
to youth, but we can’t change that any more . . . and youth want those
products too, they want those phones.
(Interview 9, YCRA-4, 2012)

Jules explains that she can make the world for youth a bit better when
she succeeds in developing marketing communication that fits the ‘needs’ of
youth. At the same time, she distances herself from her industry when she
claims to be able to develop less annoying advertising, implying that there
are others who do not succeed in meeting youth’s needs.12 She expresses her
ambivalence with her industry also in another way, saying it is still selling
to youth, while condoning it at the same time by claiming that it can’t be
changed anymore and that it is what youth want.
In the next section, we use a case study analysis of two exchange practices,
of two commercial online environments for children, to demonstrate that
both depicting representational practices are translated and inscribed into
 86 Isolde Sprenkels and Irma van der Ploeg
the design of these online environments in order to serve economic as well
as moral purposes.

TRANSLATING AND INSCRIBING ‘INSIGHTS’

In this section, we will first describe how two advergame environments for
children are designed in such a way that they meet particular corporate goals,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ranging from stimulating brand awareness and building a ‘personal relation-

ship’ between brand and child to the stimulation of product purchase and
consumption and collecting personally identifiable information. This is fol-
lowed by a description of ways in which market practitioners deal with ethical
criticism by employing representational market practices in a particular way.

Achieving Corporate Goals

Many children play online casual or mini-games. These short, ‘free,’ and
easy-to-learn games have friendly designs with bright colours and fun tasks
to perform and are developed to entertain, to educate, or to deliver a par-
ticular commercial message. This case focuses on the latter, ‘advertisement
as game,’ often referred to as ‘advergame.’ Within this chapter, advergames,
or more broadly, advergame environments, are considered to be online envi-
ronments owned by brands, consisting of elements such as casual or mini-
games specifically developed around a particular product or brand.13

OLA Ice Age and Max Adventures

In 2007, OLA and Nickelodeon organized a multi-channel campaign
together, which involved branded competition days in swimming pools in
order to enlarge OLA’s market share and to generate a healthy image for
a product which normally isn’t associated with a healthy lifestyle. Among
other things, OLA provided their ice cream wrappers with stamps called
‘heartbeats,’ which could be saved for the inflatable water fun products used
at the competition days. Nickelodeon reported on these days during one of
their best-viewed shows named SuperNick, and a special OLA Water Games
website was built with information about the competition.
For their successor campaign, OLA and Nickelodeon decided to build
on the positive experiences they had with the popular gaming section on
this website (Goeij and Kwantes, 2009). In the process of coming up with a
concept for their new campaign in 2009 to enlarge OLA’s market share and
to familiarize children with OLA’s assortment, they decided to invite chil-
dren over to brainstorm with them about their associations with ice cream
(Houben, 2009). Generating ‘insights’ through such sessions with children
can be considered a form of ‘co-creation,’ a representational market prac-
tice. These ‘insights,’ in turn, were translated into an exchange practice,
 Shaping Children’s Consumer Identity 87
the specifically designed advergame environment OLA Ice Age,14 which was
placed at the centre of their multichannel campaign. The leading role for the
OLA Ice Age campaign was taken by a cartoon character: OLA ice cream
inventor ‘Professor Freeze,’15 who lives in an ice cave, where he turns chil-
dren’s dreams into ice cream by using his ice cream machine.
The campaign was followed by the Max Adventures campaign in 2011,
which is developed in order to create brand awareness for the Dutch OLA
sub-brand for children called Max Adventures, to generate a healthy image
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

by promoting an active lifestyle (Colak and Helmons, 2013). Central to

this campaign are an animation series starring lion king Max and his fellow
adventurers,16 programmed at the Nickelodeon television channel, and a Max
Adventures advergame environment in which reruns of this animation series
can be watched and where Max related casual or mini-games can be played.17

Building a Relationship
The parents’ section on OLA Ice Age claims that Professor Freeze is “taking
children on an adventure and guides them through the games and activities
on the website” (OLAIJstijd, 2010). In their brochure informing potential
advertisers, they explain that Nickelodeon is like a crazy mentor, taking chil-
dren on an adventure within a secure and trusted environment and bringing
them back home safely afterwards (MTV Networks, 2009). Both Professor
Freeze and lion Max, characters representing the OLA brand, are presented
as mentor or role models, someone to trust. Inviting children to engage in a
trust relationship, building a personal relationship with them, is a strategy
used in marketing targeted at children. Or, as Valerie Steeves states, children
are encouraged to consider brands “as role models for the child to emulate,
in effect embedding the product right into a child’s identity” (Steeves, 2006).
When he’s not on vacation, Professor Freeze drives around in his ice cream
van to copy children’s dreams at night. Back at his ice cave, he turns these
dreams into ice cream by using his ice cream machine. When one enters the
website, an animation is showing Professor Freeze packing his bags, announc-
ing that he is going on a vacation and needs an assistant to operate the ice
machine. Professor Freeze promises that when a child has earned all badges in
the ice cave, he or she can become the assistant that Professor Freeze is looking
for. On the Max Adventures website, children are not ascribed such a role in
order to connect with them, but they are invited to choose either lion Max or
his friend lioness Leena as their avatar when playing the games.18 Pretending
to be either Max or Leena offers children the opportunity to identify with
these characters, their characteristics, and their place in the story line.

Raising Brand Awareness

Within the ice cave, there are several games to play, wallpapers and ring-
tones to download, quizzes to enter, pictures to upload, and videos to watch.
 88 Isolde Sprenkels and Irma van der Ploeg
These videos contain the OLA Ice Age commercial aired on Nickelodeon
and a realistic video about how ice cream is actually made, wrapped, and
transported to stores. Children are also encouraged to play with actual OLA
products in the ice cave, and most games and quizzes are related to existing
ice creams in OLA’s assortment. These features enable children at a later
point in time, for instance in a store, to identify the brand and its products
(Grimes and Shade, 2005). On the Max Adventures website, the connec-
tion with ice cream or OLA is not made that explicit. Only after watching
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

an episode of the animation series Max Adventures is it mentioned that the

series is offered by OLA. But Max Adventures is a sub-brand for OLA, and
children recognize the character and his name in many places, they see him
on television, on the Nickelodeon website, and the Max Adventures website,
and recognize him in stores on things like ice cream wrappers and boxes.
OLA Ice Age contains multiple game levels and public displays of high
scores, and new games are added on a regular basis. This prolongs con-
tact time with the brand continuously (Moore, 2004) and happens on the
Max Adventures website as well, contributing to this brand awareness. High
scores are shown, and multiple game levels are built into the games. New
games are displayed but inaccessible, which triggers children to go back to
the website regularly to see whether the new games are made available. As
they progress within these games, children can earn coins and buy accesso-
ries, collect dinosaurs, and develop their status from ‘adventurer’ to ‘master’
to ‘hero’ and finally to ‘legend,’ which stimulates their return to the adver-
game environment as well.

Spreading the Brand ‘ Virally’

Receiving many unique website visitors and possible new customers is also
important to the brand, and there are several ways to make that happen.
One of the many features of OLA Ice Age is decorating ‘one’s own room’ by
choosing particular colours for the elements in it. It is one of the first things
a child is invited to do. A picture of this ‘room’ can be sent to friends, which
allows OLA, represented by Professor Freeze, to invite other children to the
OLA Ice Age website. This is a form of viral marketing in which children are
encouraged to share their experience and communicate with others about
OLA. Friends receive a message such as “Do you want to decorate your
own room in the OLA Ice Age cave and play cool games? Go to . . .” Shar-
ing one’s gaming experience with others also offers the possibility to collect
personally identifiable information such as email addresses (Gurău, 2008).
The first competition related to Max Adventures was a dance competi-
tion. Market practitioners from OLA and Nickelodeon based the idea of
this competition on the claims or ‘insights’ that children like to dance, want
to become famous, and want ‘priceless prizes’ (Colak and Helmons, 2013).
They inscribed these insights into the Max Adventures campaign. Children
were invited to mimic a pre-choreographed Max or Leena dance and to
 Shaping Children’s Consumer Identity 89
upload it to the Max Adventures website and to Twitter, Facebook, its Dutch
equivalent Hyves, and a special YouTube channel. The ‘priceless prize’ was
dancing a part in the video clip of a Dutch junior song contest idol. Sharing
this video with friends and family created a viral effect for OLA and its sub-
brand Max Adventures, which triggered many more children to participate
in the competition and to visit the Max Adventures website (ibid.).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Encourage Product Purchase

Other competitions related to OLA Ice Age were designing one’s ultimate
fantasy ice cream using the ice cream machine and an arts and crafts com-
petition for which children are asked to draw or glue together their own ice
cream, take a picture of it, and upload it to the website. Both competitions
offer ‘insights’ into children’s ice cream preferences. Also, a competition was
organized for which children were encouraged to buy a popsicle containing
a code on its stick, required to enter the competition. Such direct encour-
agement to purchase products is also a common strategy in advergames
(Moore, 2004). Currently there is no active competition on the Dutch Max
Adventures website. But the Australian website, for instance, named Paddle
Pop adventures,19 shows a competition called ‘lick a prize’; the popsicles
contain codes which can be collected and are rewarded with prizes ranging
from bikes to family vacations to the US.
As the above illustrates, advergame environments such as OLA Ice Age
and Max Adventures are designed in such a way that they offer a brand
the opportunity to reach a range of corporate goals. Advergame research
also shows how some of these games include product-related polls or quiz-
zes, offering valuable information for market research on children’s hab-
its and preferences (Moore, 2004; Grimes, 2008). As in OLA Ice Age and
Max Adventures, these environments may also encourage players to reg-
ister and share their gaming experience with friends or family as we have
shown, thus enabling the collection of personally identifiable information
(Gurău, 2008). Combined with possibilities these digital technologies offer
for the analysis of in-game-behaviour and activities, marketers are able to
construct detailed consumer profiles based on the aggregation of behav-
ioural with demographic data (Chung and Grimes, 2005; Grimes, 2008).
Therefore, advergame environments can be described as Schull’s “electronic
surveillance devices” (2005), as they enable tracking children’s activities and
whereabouts both within the game as well as their physical location through
IP addresses. This illustrates that they are not only market exchange prac-
tices informed by representational practices, they can also be considered rep-
resentational practices themselves, crossing market practice boundaries.20
Notably, the OLA website contains no clear statement of its marketing
purposes to either children or their parents. The ‘parents’ section’ on the
website states that the goal is to “entertain children, stimulate their imagina-
tion and discover the world behind OLA ice cream” and “Ice cream is not
 90 Isolde Sprenkels and Irma van der Ploeg
to be sold on the website, as it is only a fun world for children” (OLAIJstijd,
2010). However, beyond a fun world for children or one that inspires chil-
dren to be adventurous and exercise more, as claimed on the Max Adven-
tures website, commercial ends are served with advergame environments
designed this way.
Agnes Nairn and Haiming Hang criticize disguising the purpose of adver-
games as a marketing communication strategy (2012). For instance, they
found that Miniclip—an internationally oriented game site for children—
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

advises companies who want to place an advergame or another form of

gamevertising on their website as follows: “Don’t use your brand name too
much. Subtle brand messages can get the message across better than in your
face logos. The key point is that the player enjoys the game” (ibid., 10).
Similarly, game advertising expert David Edery states in a presentation on
the general field of gamevertising that “the whole point is to eliminate rec-
ognition” (2009) of the advertisement. With advergame environments such
as OLA Ice Age and Max Adventures, play and fun tend to be put all up
front21 while remaining silent on processes, activities, and intentions in the
background, rendering any market practice described in this chapter invis-
ible. To children it seems indeed “just a fun world,” as play and fun are
probably the first and only associations they have with these environments.
According to the current critical academic and societal debate22 regarding
advergame environments, children are most vulnerable because they are still
developing the competencies to understand techniques of persuasion and are
not yet capable of defending themselves and thinking critically. Market prac-
titioners tend to deal with this critique regarding the ethics of digital forms
of marketing communication such as advergame environments in different
ways. Their defence has its roots in the representational market practices
discussed in this chapter.

Deflecting Moral Criticism

Academic, societal, and business attempts to evaluate digital forms of
marketing communication with children such as OLA Ice Age and Max
Adventures from an ethical or moral point of view are rooted in a history of
studying and questioning the ‘fairness’ of more traditional forms of advertis-
ing. It dealt with children’s cognitive abilities to recognize advertising and
develop a critical understanding of its mechanisms. The main points made
in this body of work, as Sonia Livingstone explains, merge “a philosophical
question about ethics (is it fair to persuade those who are unaware of such
efforts?) with an empirical question about influence (who is particularly
susceptible to persuasive messages?)” (Livingstone, 2009, 170). It is usually
based on certain presumed cognitive abilities of children, or lack thereof.
Underlying this and frequently referred to is the age-stage evolution model
of children’s cognitive development from developmental cognitive psychol-
ogy (Nairn and Fine, 2008), which is translated into models of consumer
 Shaping Children’s Consumer Identity 91
socialization (Roedder-John, 1999). This posits several developmental ‘mile-
stones’ (Lunt and Livingstone, 2012, 147), the first of which is reached when
a child is around 3 years old. This is the moment when a child is considered
to be somewhat capable of distinguishing advertising from programming.
The moment a child can understand the selling and persuasive intent behind
an advertisement and can distinguish it from information, usually at around
8 years old, is considered to be the second milestone.23 A third and final
‘milestone’ is reached somewhere around age 12, when a child is presumed
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

to have acquired the ability to critically reflect on, weigh, and refuse an
advertisement. It is not until children have reached this latter stage that tar-
geting advertising to them is considered to be ‘fair’ (Nairn and Hang, 2012).
Given this, the primary way in which market practitioners deflect moral
criticism, claiming their market practices to be ethical, is to argue that they
take children’s (cognitive) developmental abilities into account. They base
their market exchange practices and normalising practices such as their
codes of conduct on children’s cognitive psychological development and
associated interests. But, as we claim in what follows, they appear to do so
selectively.
For instance, the academic discourse on children and advertising and the
age-stage model of consumer socialization has evolved. In this academic
discourse, it is argued that several difficulties exist when using the model
as an ethical yardstick for assessing the ‘fairness’ of advertising to children.
Following the argument, it is generally taken for granted that “those whose
literacy is lower are assumed to be more susceptible to effects”24 and that
“an increase in media or advertising literacy is assumed to reduce suscep-
tibility to media effects” (Livingstone and Helsper, 2006, 564). But Sonia
Livingstone and Ellen Helsper found little empirical evidence for this claim.
They argue that although teenagers are presumed to be more ‘media liter-
ate,’ that does not mean that they are less influenced by advertising than
younger children are. They conclude that “different processes of persua-
sion are effective at different ages, precisely because literacy levels vary by
age” (ibid.). Which means that younger children, being less media literate,
are merely persuaded differently than teenagers are. Younger children tend
to be persuaded by superficial or peripheral features of advertising such
as jingles and colourful and funny images, whereas teenagers tend to be
persuaded by, for instance, strong arguments and references to peer-group
approval (ibid.).25
Another problem with this claim on literacy levels and susceptibility, which
has already led certain academics to reject such a model, has to do with
contemporary marketing communication formats such as OLA Ice Age
and Max Adventures. Livingstone concludes that the age-stage evolution
model of children’s cognitive abilities no longer fits the diversity of the 21st-
century media environment (2009). She claims that the idea of ‘milestones’
related to age is not convincing, “both because there is no universal relation
between understanding and age, and because persuasion occurs, in one way
 92 Isolde Sprenkels and Irma van der Ploeg
or another, across the range” (Livingstone, 2009, 172). In saying this, she
follows Agnes Nairn and psychologist Cordelia Fine, who doubt the pos-
sibility of establishing any ‘magical age’ at which children are supposed to
understand and resist persuasion, as contemporary formats often intention-
ally bypass children’s explicit persuasion knowledge and instead persuade
implicitly (2008). They tend to be more covert, as the brand and its corpo-
rate goals are integrated into non-commercial contexts. This challenges not
only younger and older children’s cognitive defences but those of teenag-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ers and even adults as well (Moore, 2004; Fielder et al., 2007; Nairn and
Fine, 2008). It indicates that vulnerability depends on both an individual’s
cognitive social resources and the (social or mediated) environment (Living-
stone, 2009). Focusing exclusively on cognitive abilities implies a specific
view of technology, considering it to be inherently neutral. We have argued
elsewhere that the ‘fairness issue’ can be considered as something that is a
function of the set of socio-material relations in which children’s options,
choices, and chances to resist are also shaped by the marketing communica-
tion tactics and formats themselves (Sprenkels and van der Ploeg, 2014).
The specific characteristics of contemporary marketing communication
touch on some of the issues with market practitioners’ codes of conduct.
They argue that they have translated the abovementioned developmental
milestones into normalising practices, into their codes of conduct based on
self-regulation. One example is the Dutch Advertising Code26 and its spe-
cial section on children and advertising,27 which proposes to help children
differentiate content as it prescribes that the distinction between editorial
content and advertising targeting children—defined as a person being 12 or
younger—should always be made recognizable. For instance, television com-
mercials are preferably clustered, and banner and pop-up advertisements on
the internet should be labelled with the word ‘advertisement.’ However, as
mentioned, the age-stage model of consumer socialization is problematic,
and when it comes to contemporary formats such as advergame environ-
ments, the code is not followed because the distinction is not made explicit
in any way at all (Fielder et al., 2007). Not only do such ‘seamless environ-
ments’ (Moore, 2004) fail to follow the code, this form of self-regulation
advocates a strategy that runs exactly counter to what advergames tend to
be about, namely presenting advertisements, and all related market practices
running in the background, as game.
As suggested earlier, market practitioners translate academic theories
on children’s cognitive development into various age-based classification
models. They come up with a variety of age categories in practice, refining
and reducing them to smaller and smaller age groups. It is a continuous
effort to construct and stabilize these ‘stages,’ which they constitute into
being by re-describing them in their own terms. We have demonstrated that
such categories or market segments are contestable constructions, they are
contingent, different according to different actors and within different set-
tings. Market practitioners consider their work to be ethical, as they claim to
 Shaping Children’s Consumer Identity 93
follow children’s developmental needs, which they presume to be clear cut,
out there, which as shown, they are not.
The secondary way in which market practitioners consider their mar-
ket practices to be ethical relates to their claim that they take children’s
‘needs and wants’ into account by giving them a voice, empowering them,
including them, using representational practices such as co-creation. But
we have illustrated in the previous sections that those practices tend to be
performative, too, producing what they seek to describe. There appears to
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

be an assumption that it is perfectly legitimate to target children as long as

this is based on and provides ‘what children want and need,’ as if the latter
can be considered to be a moral right. Thus, the more market practitioners
attempt to know what children think, want, and need, the better they will
be able to provide this, and hence, the less reprehensible it is to use sophis-
ticated marketing strategies aimed at children. But as illustrated earlier in
this chapter, the ‘knowledge’ about children required for this, the ‘listening’
to their ‘needs and wants,’ comes from their own marketing communication
and research focussed on children from the very market practices in need of
legitimization. We argue that contemporary marketers’ constitution of chil-
dren’s preferences and identities as consumers are actually socio-technical
enactments enabled by the market practices and research techniques under
discussion and described in this chapter. Using children’s identities as com-
petent and desiring consumers as legitimization of these same practices is
circular.

CONCLUSION

This chapter portrayed a configuration of contemporary market practices,

characterizing and categorizing, monitoring and targeting Dutch children in a
digitising society. It illustrated the performativity of such socio-material prac-
tices and discussed market professionals’ self-produced legitimizations based
on their presumed ‘insights’ or ‘knowledge’ of children. It demonstrated ways
in which the Dutch ‘kids’ market’ and children’s identities as consumers are
constituted within this configuration, that they are not pre-existing entities,
but that they are continuously enacted into being. What became clear was that
the ‘insights’ gained from the representational market practices are not only
translated and inscribed into market exchange practices, such as the commer-
cial online environments or advergames discussed, in order to meet corporate
goals. They also contribute to the legitimization of both exchange practices
and representational practices themselves. But the outcomes of market prac-
titioners’ representational practices are contingent and performative; they
partly constitute these abilities, interests, and ‘needs and wants’ themselves.
In other words, within the configuration of market practices, children are
enacted as full-fledged competent consumers, a notion that is subsequently
invoked to legitimize the very practices involved in its production.
 94 Isolde Sprenkels and Irma van der Ploeg
The practice-based approach as well as the semiotic approach to user–
technology relations used in this chapter, with notions such as performa-
tivity, enactment, translation, inscription, representational, exchange, and
normalizing practices, can contribute to the current social and academic
debate concerning contemporary (digital) marketing communication tar-
geting children. Contrary to most views on the matter, it allows us to take
a broad perspective on the various market practices involved, how they
are entangled and translated, and what it is that these practices and their
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

connections do. It also allows us to take the various digital means them-
selves, used by both children and market practitioners in a digitising soci-
ety, into account. These digital means are not to be considered neutral; for
instance, advergames play an active role in shaping these market practices,
children’s consumer identities, and children’s commercial literacy (see also
the chapter by Sprenkels and Wyatt in this volume), and conversely, they
are actively shaped by market practices and children’s new media use
themselves.

NOTES

1. Daniel Thomas Cook defines the term ‘market practitioners’ as “marketers,

researchers, designers, manufacturers, and other market actors” (2011, 258).
2. Following social and constructivist studies of science and technology (STS),
more particularly actor-network theory (ANT) (Callon, 1986; Latour, 1987,
1991; Akrich, 1992), identity can be considered to be enacted within socio-
material arrangements, by and through relations between both human and
non-human actors, shaping and defining each other and constituting networks.
3. As explained by MacKenzie and colleagues (2007), the notion of performativ-
ity has its origins in the work of philosopher J. L. Austin on speech acts, utter-
ances that bring a state of affairs into being such as ‘I apologize.’ Later on, the
notion has been taken up and broadened by the social sciences, humanities,
and science studies.
4. ‘Trends in kids en jongerenmarketing’ in Dutch, currently the only major
Dutch conference in this field.
5. Sociologist Juliet Schor explained having a similar experience when talking to
market practitioners about their research methods; they appeared to rely on a
common psychological model of the child, seeing children steadily developing
into adulthood (2005).
6. The names of interviewees and conference speakers in this chapter are fictitious.
7. Interestingly, Cherie refers to this phase as ‘no longer being a child.’ This prob-
ably stems from children’s often-heard statement to consider certain things
to be ‘childish.’ But it is re-used and translated in such a way that it benefits
market practitioners’ ‘kids getting older younger’ claim discussed later.
8. For instance, in his work on the rise of the children’s clothing industry, Cook
explained that younger children needed to pass through the older children’s
area in a shop (2004).
9. Coulter applies Homi Bhabha’s reading of colonialism to the colonizing of
young people as markets in need of ‘discovery.’ She uses Bhabha’s “regime of
truth” (Bhabha, 2001; in Coulter, 2012), in “which the colonizer controls the
colonized through finding out everything about the colonized and using this
 Shaping Children’s Consumer Identity 95
knowledge to construct an identity of the colonized in a unified and coherent
way” (Coulter, 2012, 151).
10. Similarly, Eva Heiskanen has conducted research on the performative nature
of consumer research depicting ‘the green consumer’ (2005), and Catherine
Grandclément and Gérard Gaglio have studied the performance aspect of
focus group–based market research (2011).
11. By most industry leaders defined as “advertising that is age-appropriate,
socially sensitive, and honest in its portrayal of the advertised product” (Henry,
2005, 11).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

12. Amy Henry, who wrote a report on marketers targeting children and their
perspectives on ‘good work’ and US industry leaders’ discomfort with the
label ‘kids’ marketer,’ also found that when it comes to professional ethics,
marketers stressed that they did not engage in unethical work themselves but
expressed their concern that others might do so (2005).
13. Advergames are considered to be a particular form of gamevertising, with
gamevertising in general being the promotional or advertising possibilities
before, within, or after an often already existing console, PC, or Internet game
(Hufen, 2010).
14. In Dutch: OLA IJstijd. The website www.olaijstijd.nl is no longer in use.
15. ‘Professor De Vries’ in Dutch, which is a very common name in the Netherlands.
16. In many other countries such as South Africa, Indonesia, India, New Zealand,
and Australia, Max is known as the Paddle Pop Lion. Similar campaigns run
in those countries. Paddle Pop is, just like Max Adventures is for OLA, a
sub-brand for Streets, the Australian equivalent of OLA. It originates from
a particular Australian ice cream type developed in the 1950s. The ice cream
wrapper contained a cartoon character which was known as the Paddle Pop
lion.
17. www.max-adventures.com/nl-nl/
18. Lion Max has many fellow adventurers besides lioness Leena, as well as sev-
eral enemies, all playing active roles in the series. On the website all the char-
acters of the latest story ‘Dino Terra’ are described. The website shows two
previous series as well: Kombatei and Begins. They are all divided up in 8 to
10 episodes with a duration of 10 to 24 minutes.
19. See endnote 16.
20. Marketing professor Agnes Nairn and market research company director Bar-
bie Clarke stress a line between market research, which according to them
does not seek to alter opinion, and marketing, which according to them does
seek to alter opinion (2012). This chapter clearly argues that such clear distinc-
tions are difficult to uphold.
21. According to Cook, market practitioners tailor messages and design products,
packages, websites, campaigns, and advertisements in such a way that they
appeal to children, designing it to be fun, playful, entertaining (2011).
22. For instance, there is some debate regarding marketing communication target-
ing children in the Netherlands. In reports from consumer organizations (Food-
watch, 2013), governmental institutions (Autoriteit Consument en Markt,
2013), and child and media knowledge centres (Mijn Kind Online, 2008), vari-
ous practices are being criticized.
23. Although, as David Buckingham explains, some studies suggest that this under-
standing is not necessarily used. He claims differences in these estimations are
a consequence of research method (Buckingham, 2009).
24. Advertising effects can be intended by advertisers, such as brand awareness
and buying intent, and non-intended, such as materialism and family conflicts
(Valkenburg, 2002, 140). In this chapter we call intended effects by advertisers
‘goals.’
 96 Isolde Sprenkels and Irma van der Ploeg
25. Here Livingstone and Helsper are inspired by Petty and Cacioppo’s elabora-
tion likelihood model of persuasion, which distinguishes two routes of persua-
sion, a central one and the peripheral one.
26. Nederlandse Reclame Code
27. For another example, the Dutch Advertising Code for Foods—Reclamecode
voor Voedingsmiddelen—forbidding targeting of children under 7 years of age
with (commercial) food advertisements and stressing the abstention in target-
ing children between 7 and 12 years of age, view (Sprenkels, forthcoming). It
offers a close analysis of Unilever’s principles for responsible food and bever-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

age marketing and their motivation of their practices related to OLA Ice Age
and Max Adventures.

REFERENCES

AKRICH, M. (1992) The De-scription of Technical Objects. In Bijker, W. E. & Law,

J. (eds.) Shaping Technology/Building Society—Studies in Sociotechnical Change.
Cambridge (MA): MIT Press, 205–224.
ARAUJO, L. (2007) Markets, Market-Making and Marketing. Marketing Theory,
7(3). p. 211–226.
AUTORITEIT CONSUMENT EN MARKT. (2013). Marktscan naar online games.
Den Haag: Mediawijzer.
BHABHA, H. (2001) The Other Question: Stereotype, Discrimination and the Dis-
course of Colonialism. In Seidman, S. & Alexander J. C. (Eds.), The New Social
Theory Reader: Contemporary Debates. New York: Routledge, 388–402.
BOWKER, G. C. & STAR, S. L. (1999) Sorting Things Out. Classification and Its
Consequences. Cambridge (MA): MIT Press.
BUCKINGHAM, D. (2007) Selling Childhood? Children and Consumer Culture.
Journal of Children and Media, 1(1). p. 15–24.
BUCKINGHAM, D. (2009) Beyond the Competent Consumer: The Role of Media
Literacy in the Making of Regulatory Policy on Children and Food Advertising in
the UK. International Journal of Cultural Policy, 15(2). p. 217–230.
CALLON, M. (1986) Some Elements of a Sociology of Translation: Domestica-
tion of the Scallops and the Fishermen of St Brieuc Bay. In Law, J. (ed.), Power,
Action and Belief: A New Sociology of Knowledge? London: Routledge,
196–223.
CHUNG, G. & GRIMES, S. M. (2005) Data Mining the Kids: Surveillance and Mar-
ket Research Strategies in Children’s Online Games. Canadian Journal of Com-
munication, 30, p. 527–548.
COLAK, E. & HELMONS H. (2013) OLA Max Adventures. Presentation at Branded
Content Event, January 31, 2013. Retrieved from http://brandedcontentevent.nl/,
accessed July 23, 2013.
COOK, D. T. (2004) The Commodification of Childhood. The Children’s Clothing
Industry and the Rise of the Child Consumer. Durham and London: Duke Uni-
versity Press.
COOK, D. T. (2011) Commercial Epistemologies of Childhood. “Fun” and the
Leveraging of Children’s Subjectivities and Desires. In Zwick D. & Cayla, J.
(eds.), Inside Marketing: Practices, Ideologies, Devices. Oxford: Oxford Univer-
sity Press, 257–268.
COULTER, N. (2009) Tweening the Girl: The Crystallization of the Tween Market
1980–1996. PhD dissertation, Simon Fraser University.
COULTER, N. (2012) From Toddlers to Teens: The Colonization of Childhood the
Disney Way. Jeunesse: Young People, Texts, Cultures, 4(1). p. 146–158.
 Shaping Children’s Consumer Identity 97
EDERY, D. (2009) Rethinking the Impact and the Potential of Brands in Games.
Presentation at Brands and Games 2009. The game advertising network event.
March 10, 2009.
FIELDER, A., GARDNER, W., NAIRN, A. & PITT, J. (2007) Fair Game? Assessing
Commercial Activity on Children’s Favourite Websites and Online Environments.
London: Report National Consumer Council and Childnet International.
FOODWATCH. (2013). Kindermarketing: Onverantwoord en Ongereguleerd. Retrieved
from www.foodwatch.nl, accessed December 3, 2013.
GOEIJ, M. DE and KWANTES, E. (2009) Nickelodeon and OLA. Hoe betrek je kids
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

écht bij je merk? Presentation at Brands and Games 2009. The game advertis-
ing network event. March 10, 2009. Retrieved from www.nlgd.nl/fog/pdf/bag09/
BG8.pdf, accessed June 10, 2010.
GRANDCLÉMENT, C. & GAGLIO, G. (2011) Convoking the Consumer in Person:
The Focus Group Effect. In Zwick, D and Cayla, J. (eds.), Inside Marketing: Prac-
tices, Ideologies, Devices. Oxford: Oxford University Press, 87–115.
GRIMES, S. M. (2008) Kid’s Ad Play: Regulating Children’s Advergames in the
Converging Media Context. International Journal of Communications Law and
Policy, (12). p. 161–178.
GRIMES, S. M. & SHADE, L.R. (2005). Neopian Economics of Play: Children’s
Cyberpets and Online Communities as Immersive Advertising in NeoPets.com.
International Journal of Media and Cultural Politics, 1(2). p. 181–198.
GURĂU, C. (2008 The Influence of Advergames on Players’ Behaviour : An Experi-
mental Study. Electronic Markets, 18(2). p. 106–116.
HARRISON, D. & KJELLBERG, H. (2010) Segmenting a Market in the Making:
Industrial Market Segmentation as Construction. Industrial Marketing Manage-
ment, 39(5). p. 784–792.
HEISKANEN, E. (2005) The Performative Nature of Consumer Research: Consum-
ers’ Environmental Awareness as an Example. Journal of Consumer Policy, 28,
179–201.
HENRY, A. (2005) Marketing to Children: Industry Insiders’ Perspectives on
Good Work (p. 1–33). Harvard University, Good Work Project Report Series,
Number 40.
HOUBEN, J. (2009) OLA IJstijd Kids campagne 2009. Presentation at Kids Insights
Day, October 21, 2009. Retrieved from: http://media.mtvnetworks.nl/kidsday/
kidsdayjoosthouben.pdf, accessed June 1, 2010.
HUFEN, B. (2010) Laat met je merk spelen. Games als marketingmiddel. Amster-
dam: Kluwer.
KJELLBERG, H. & HELGESSON, C. F. (2007) On the Nature of Markets and Their
Practices. Marketing Theory, 7(2). p. 137–162.
LATOUR, B. (1987) Science in Action. How to Follow Scientists and Engineers
Through Society. Cambridge (MA): Harvard University Press.
LATOUR, B. (1991) Technology Is Society Made Durable. In Law, J. (ed.), A Soci-
ology of Monsters: Essays on Power, Technology and Domination. New York:
Routledge, 103–131.
LAW, J. (2009) Seeing Like a Survey. Cultural Sociology, 3(2). p. 239–256.
LAW, J., RUPPERT, E. & SAVAGE, M. (2011) The Double Social Life of Methods.
CRESC Working Paper Series. Working Paper No. 95.
LIVINGSTONE, S. (2009) Debating Children’s Susceptibility to Persuasion—Where
Does Fairness Come In? A Commentary on the Nairn and Fine Versus Ambler
Debate. International Journal of Advertising, 28(1). p. 170–174.
LIVINGSTONE, S. & HELSPER, E. J. (2006) Does Advertising Literacy Mediate
the Effects of Advertising on Children? A Critical Examination of Two Linked
Research Literatures in Relation to Obesity and Food Choice. Journal of Com-
munication, 56(3). p. 560–584.
 98 Isolde Sprenkels and Irma van der Ploeg
LUNT, P. & LIVINGSTONE, S. (2012) Media Regulation. Governance and the
Interests of Citizens and Consumers. London: Sage Publications.
MACKENZIE, D., MUNIESA, F. & SIU, L. (2007) Introduction. In MacKenzie, D.,
Muniesa, F. & Siu, L. (eds.), Do Economists Make Markets? On the Performativ-
ity of Economics. Princeton: Princeton University Press, 1–19.
MIJNKINDONLINE. (2008) Gratis! (maar niet heus)—dossier over digitale reclame
voor kinderen.
MONTGOMERY, K. C. & CHESTER, J. (2009) Interactive Food and Beverage
Marketing: Targeting Adolescents in the Digital Age. Journal of Adolescent
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Health, 45. p. S18–S29.

MOORE, E. S. (2004) Children and the Changing World of Advertising. Journal of
Business Ethics, 52(2). p. 161–167.
MTV NETWORKS. (2009). Adverteren bij MTV Networks in 2010. Available at:
www.mtvnetworks.nl/index.php?list/5, accessed 12.06.2010.
NAIRN, A. & CLARKE, B. (2012) Researching Children: Are We Getting It Right? A
Discussion of Ethics. International Journal of Market Research, 54(2): 177–198.
NAIRN, A. & FINE, C. (2008). Who’s Messing With My Mind? The Implications
of Dual-Process Models for the Ethics of Advertising to Children. International
Journal of Advertising, 27(3). p. 447–470.
NAIRN, A. & HANG, H. (2012) Advergames: “It’s not an advert—it says play!”: A
review of research. Retrieved from Family & Parenting Institute, p. 1–22 http://
www.bath.ac.uk/management/news_events/pdf/advergames-report-december
2012.pdf
OLAIJstijd (2010) Informatie voor ouders. Online at: www.olaijstijd.nl/ijstijd/info_
ouders.html, accessed 15.06.2010.
OUDSHOORN, N., ROMMES, E. & STIENSTRA, M. (2004) Configuring the User
as Everybody: Gender and Design Cultures in Information and Communication
Technologies. Science, Technology, and Human Values, 29(1). p. 30–63.
PIAGET, J. (1954) The Construction of Reality in the Child. (M. Cook, Trans.). New
York: Routledge.
PLOEG, I. VAN DER. (2007) Social and Ethical Aspects of Digital Identities, Towards
a Value Sensitive Identity Management. European Research Council Starting Grant
research proposal.
QVORTRUP, J., CORSARO, W. A. & HONIG, M. S. (2011) Introduction: Why
Social Studies of Childhood? In Qvortrup, J., Corsaro, W. A. & Honig, M.-S.
(eds.), The Palgrave Handbook of Childhood Studies. Basingstoke (UK): Palgrave
Macmillan, 21–33.
ROEDDER-JOHN, D. (1999) Consumer Socialization of Children: A Retrospective
Look at Twenty-Five Years of Research. Journal of Consumer Research, 26(3).
p. 183–213.
SCHOR, J. B. (2005) Born to Buy. New York: Scribner.
SCHULL, N. D. (2005) Digital Gambling: The Coincidence of Desire and Design.
The ANNALS of the American Academy of Political and Social Science, 597(1).
p. 65–81.
SPRENKELS, I. & PLOEG, I. VAN DER (2014) Follow the Children! Advergames
and the Enactment of Children’s Consumer Identity. In van der Hof, S., van den
Berg, B. & Schermer, B. (eds.), Minding Minors Wandering the Web—Regulating
Online Child Safety. The Hague: TMC Asser Press/Springer Press. p. 173–188.
STEEVES, V. (2006) It’s Not Child’s Play: The Online Invasion of Children’s Privacy.
University of Ottawa Law and Technology Journal, 3(1). p. 169–188.
STEEVES, V. (2007) The Watched Child: Surveillance in Three Online Playgrounds.
In International Conference on the Rights of the Child. Montreal: Wilson Lafleur,
119–140.
 Shaping Children’s Consumer Identity 99
TODD, S. (2010) The Ethics of Marketing to Children. In Marshall, D. (ed.), Under-
standing Children as Consumers. London: Sage, 221–238.
VALKENBURG, P. (2002) Beeldschermkinderen. Theorieën over kind en media.
Amsterdam: Boom.
VILLE, V.-I. DE LA & TARTAS, V. (2010) Developing as Consumers. In Marshall, D.
(ed.), Understanding Children as Consumers. London: Sage, 23–40.
ZOONEN, L. VAN. (2013) From Identity to Identification: Fixating the Fragmented
Self. Media, Culture & Society, 35(1). p. 44–51.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Part II

Growing Up
Children and Guardians
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 5 Risk Identities
Constructing Actionable Problems
in Dutch Youth
Karolina La Fors-Owczynik and Govert
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Valkenburg

INTRODUCTION

Dutch child-care policy has become increasingly focused on making the lives
of children ‘transparent.’ This is accomplished by connecting multiple digi-
tal databases and aggregating data about children and the persons to which
they are connected. This is to give the most ‘complete picture’ possible of
children who may be at some sort of risk. This is believed to increase the
success rates of interventions made in the lives of these children (Keymolen
and Prins, 2011, 21). The (potential) problems these systems are to address
include abuse and negligence within families, educational problems includ-
ing dropping out of school, and petty criminal behaviour such as shoplifting,
nuisance, and vandalism.
In this chapter, we explore developments in ‘completing the picture’ of
children through the use and development of youth-care–related databases.
First, we argue, these systems are not just a matter of installing the right
software and operating it appropriately. Rather, it requires a lot of work
to make the systems work and literally make them match a rather complex
and ambiguous world. Second, while this work is intended to make the risks
visible, it also inevitably obscures some of the realities it tries to represent,
which in turn entails that some of the risks are ‘constructed’ rather than
merely ‘represented.’ Finally, we observe that the whole logic of the system is
such that it tends to expand: If it fails in some sense, the response is typically
to install more risk assessments, to implement more risk indicators, and to
start the risk identification processes earlier in a child’s life.
It is perhaps not that surprising that a comprehensive approach to youth
risks includes the establishment of the most complete possible set of infor-
mation from a variety of professionals, as this increases the likelihood that a
risk or future wrong is identified in time. One of the explicit aims of national
youth-care policies is that ‘no child should go unseen’ (Inspectie Jeugdzorg,
2008).1 However, we will show that the ever-increasing demand for infor-
mation on a child does not always work out the way it was intended, for
example when stigmatization arguably occurs (Dutch Youth Institute,
reported by NOS, 2010).
 104 K. La Fors-Owczynik and G. Valkenburg
Risk and prevention are not simple concepts. On the contrary, in this
context they refer to complex arrangements of children, professionals, insti-
tutions, personal records, data technologies, legal statuses, communication
protocols, and professional routines. Within the Netherlands, youth care
is organized across medical, educational, and law-enforcement institutions.
In practice, this means risks are identified and shared among various youth
healthcare organisations, police departments, schools, judicial institutions
such as the Child Protection Council and the Youth Care Bureau, municipal-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ities, emergency departments of hospitals, social workers, housing compa-

nies, sport clubs, and a variety of other organizations involved with youth.
We show that some problems occurring in child care are in fact owing to the
heterogeneity across parties.
We will discuss three database systems currently used in the Netherlands.
First, we discuss the Digital Youth Healthcare Registry (DYHR, or Digitaal
Jeugdgezondheidszorgdossier in Dutch). It is designed for the registration
of healthcare information on children between 0 and 19 years. Second, we
discuss the Reference Index for High Risk Youth (RI, or Verwijsindex Risi-
cojongeren in Dutch). The RI is a large-scale risk-signalling system that con-
nects various digital systems, youth-care organizations, and professionals
by providing digital exchange of risk signals about children and youngsters
aged between 0 and 23 years in the Netherlands. Finally, we discuss ProKid
SI 12- (pronounced ‘ProKid SI twelve-minus,’ henceforth ProKid). This is
a tool for risk assessment on children aged between 0 and 12 years, used
nationwide by the police. In ProKid, colour codes are assigned to children,
reflecting various levels of estimated risk.
At face value, it appears as if ‘risks’ are simply identified and then com-
municated through particular networks and relationships. On the one hand,
‘risks’ indeed appear in the form of simple indicators such as classifications,
signals, and colour codes. These forms make risks ‘actionable.’ On the
other hand, the indicators consolidate extensive assessments. When indica-
tors arrive at a new location, they are in turn interpreted and re-imbued
with a potentially different meaning, instigating different interventions. This
occurs because of divergent backgrounds of professional knowledge, rou-
tines, codes of conduct, and anything else on which professional practices
may just differ. As we continue to argue, the work done to deal with risks is
successful in some ways and perilous in others.

RISK ASSESSMENT SYSTEMS FOR YOUTH

IN THE NETHERLANDS

Following Dutch youth policy, both youth care and law enforcement have
increasingly become geared towards prevention: problems are to be solved
proactively before they become real. ‘Risks’ and their identification serve to
make problems actionable before they actually occur. Indeed, wordings such
 Risk Identities 105
as “making risks visible” or “making the child transparent”2 often appear
in youth policy. This eagerness for making visible the potential problems
of children is reflected by mottos from professional reports: “to establish
a comprehensive approach to children” (Berkeley and Van Uden, 2009),
“to create a complete view of children,” and “no child should go unseen”
(Inspectie Jeugdzorg, 2008). Identifying those who might be responsible for
future crimes provides the rationale of many prevention practices. As one
professional explains:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The goal initiating ProKid was to ensure that police officers in the field
work more efficiently. What you see with the police is that normally they
respond to an incident. Only when something happens, the police are
called, and then they arrive to start an investigation. [. . .] The idea was
that you want to know the 5% of people that are responsible for 60%
of all crime incidents. If you know them, you can focus on the 5% of
children who may actually come in contact with the police.
(ProKid professional, city A)

This ‘making visible’ of children’s problems serves a two-tier purpose:

First, it facilitates the identification and classification of the ‘abnormal,’ and
second, by consequence, it makes these abnormalities actionable or at least
presents the indispensable object for action. This practice is justified by the
assumption that intervention will be more effective if it takes place earlier.
This entails a need for early identification of risks. Thus, early risk assess-
ments on youth are considered essential resources for professionals.
Emphasis on prevention is also evident in other areas of youth care,
such as for example youth healthcare (Gorissen, 2002; Mathar and Jansen,
2010). The availability of interconnected risk identification systems also
facilitated police corps and social workers to increasingly carry out pro-
active actions and prevent harms not suffered and crimes not committed yet
by youth (Min. van Veiligheid en Justitie, 2012). At the same time, youth
healthcare professionals increasingly take up tasks that shift from coun-
selling and support towards the prevention and control of youth and the
crimes they potentially commit (Frissen, Karré and Van der Steen, 2011).
This preventive approach to youth care and the according arrangements of
organizations have received wide acclaim in professional policy discourse.
The consensus seems to be that the results are good and investments justified
(ACTIZ, 2012).

Digital Youth Healthcare Registry

In the Netherlands, children aged between 0 and 19 are screened periodi-
cally through consultations with healthcare professionals. These occur at
specialized youth healthcare offices for children from 0 to 4 years and by
school doctors for children from 5 to 19. The Digital Youth Healthcare
 106 K. La Fors-Owczynik and G. Valkenburg
Registry (DYHR) is a healthcare database used for the registration of chil-
dren’s psychological, social, and cognitive data. Use of the DYHR by school
doctors and advice centres became mandatory as of 2010.
In the DYHR, a child’s record is kept as a basic data set (BDS). In addi-
tion to generic medical data (weight, height, visual and auditive capacity,
etc.), the BDS contains indicators that could be regarded as risk indicators.
Most prominent among these is the list of “invasive events” that was part of
the 2011 version of the BDS (Nederlands Centrum Jeugdgezondheid, 2011).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This includes suspected physical abuse, serious illness of the parents, parental
divorce, alcohol abuse by the parents, and teenage pregnancy. Notably, as this
list was found to be too much of a straightjacket, it was replaced by a ‘free
text field’ in the subsequent version (Nederlands Centrum Jeugdgezondheid,
2011, 2013). Although the text field now allows for less specified (‘free’)
descriptions of risk, it remains a significant indicator for risk. Also, while the
‘freedom’ to report increases, it is still only a freedom to report risks, which
entails that the possibilities for a child to become risk profiled grow.

Reference Index for High-Risk Youth

The Reference Index for High-risk Youth (RI) is a comprehensive risk-
signalling system that connects a variety of digital systems. It enables youth-
care organizations and professionals to digitally exchange risk signals on
children and young adults aged 0 to 23 years. It is an infrastructure rather
than a single system. Each time a risk is identified and entered into a local
system, this identification is automatically forwarded to the national RI.
The aforementioned DYHR as well as other youth healthcare practices in
general are connected to the RI (Rijksoverheid, 2005; Rouvoet, 2007), and
it has been operational since 2010.
The RI facilitates the communication of alerts regarding children among
professionals. The transmitted risk signals represent potential harms. The
risk signals only consist of risk flags and are not automatically supplied
with explanatory information or any other content of the children’s files.
The only information accompanying the risk flag is the name of the profes-
sional who issues it, the name of the organization the professional works for,
and the name of the child the signal is related to (Nouwt and Hogendorp,
2010). The concealment of more substantial information was demanded by
the Dutch Medical Association (Dutch: KNMG, Koninklijke Nederlandse
Maatschappij tot Bevordering der Geneeskunst), as the RI would otherwise
have posed too much of a privacy and confidentiality encroachment.

ProKid SI 12-
ProKid SI 12- is a risk-categorization system for children between 0 and
12 years. The child’s address serves as the primary key to the database.
ProKid employs four colour codes to represent risks: white, yellow, orange,
 Risk Identities 107
and red, respectively signifying a growing gradation of concern. White sig-
nifies that no risks have been identified as yet, yellow represents a possible
development of risks, orange indicates that problems and even criminal
behaviour have already occurred, and red symbolizes children who have
repeatedly shown criminal behaviour or of whom multiple cases of suspi-
cion have been recorded (Abraham, Buysse, Loef and Van Dijk, 2011). Fol-
lowing a pilot phase, the system has been in operation nation-wide by the
Dutch police as of 2013 (Tweede Kamer, 2012).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The risk categories and their attributions are based on behavioural-

scientific models and explicitly aimed at early detection of potential criminal
behaviour of children. The attribution of colour codes is done automatically,
by means of algorithms, on the basis of information in other large police
databases. These include the Basic Facility for Law Enforcement (BFLE,
or Basisvoorziening Handhaving in Dutch) as well as the Basic Facility for
Forensic Investigation (BFFI, or Basisvoorziening Opsporing in Dutch).
Information from ProKid is shared among youth-care partners if a ProKid
manager decides that there is reason to do so.

DOING ‘RISK’ IN PRACTICE

Identifying and eliminating youth risks is complex work. Numerous profes-

sionals are involved, ranging from educational and social practitioners to
healthcare workers and legal and police professionals. By definition, the
knowledge one professional has of a child is always limited: not so much
because they do their work poorly or because they have limited resources
but because any profession has things on which it focuses and things on
which it does not. A medical professional ‘sees’ different things than a police
officer or a schoolteacher does.
In recognition of this partiality of knowledge, the information systems
introduced are explicitly set up to meet the challenge of crossing boundaries
between professions. The systems are intended to yield a more comprehen-
sive view of the risks children are exposed to than any single profession
could have by itself. This sharing of information is increasingly seen as an
essential resource for a range of youth-care professionals in the execution of
their daily preventive tasks (Radaradvies, 2011). The emphasis is adopted
by many regulations that legally stimulate the deployment of digital risk
profiles. The Public Health Act (Rijksoverheid, 2008) requires profession-
als to digitally register data including risks about children in DYHR. The
Law on the Reference Index (Rijksoverheid, 2010) stimulates the aggrega-
tion of risk profiles on children by a variety of professionals involved in
youth care.
However, it turns out that the sharing of information is not just a pana-
cea to eliminate risks. Managing the information requires work of its own,
and while often fruitful, this work is also at times difficult and frustrating.
 108 K. La Fors-Owczynik and G. Valkenburg
One corollary of the fact that professions frame problems differently is
that seemingly the same information might have different meanings and
implications in different situations. In the doctor’s office, a mother with a
medical problem might be just that, but from a social-work perspective,
more structural problems might be suspected. In choosing between these
options, a healthcare professional may need more comprehensive informa-
tion from other professionals, which may not be available through mere
risk indicators.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Risk indicators are compact representations of risks. They may be shaped

as, amongst others, categories, colour codes, or signals. While facilitating
communication, indicators also consolidate comprehensive assessments,
complex sets of information, and tacit professional knowledge. An indica-
tor (at least partly) obscures the process of risk identification. This (at least
partly) precludes their discussion and revision, and indicators become more
fixed de facto. A perfect example is the use of colour codes in the ProKid
system, which turn out to be both elegantly compact and sources of confu-
sion (Willems and Van der Heijden, 2011, 3).
The practices in which individual professionals operate are more unruly
than the ideal of comprehensive risk identification suggests. Among differ-
ent institutions and professions, consensus is often lacking on what a risk
exactly is. A youth healthcare professional, for instance, explained her frus-
tration over this problem and how it troubles her ability to assess risks:

You see the categories with a red checkmark, with certain families the
screen is almost only red. You still need to read the record [underlying
the checkmarks], and you can relatively quickly figure out what the
problems are. But it is often difficult to fill in the risk assessment form,
or to risk-score children, because there is no clear definition of risks.
For instance, a ‘sleeping problem’ of a child can be just that, but it can
also be a symptom of issues between the parents; or it can be caused
by cultural elements that are just different between one family and
another.
(CB-nurse, city D)

While the meaning of the symbolic representation of the risk may remain
unclear, its presence in a graphic material form seems to express a very clear
urgency for action. Yet if professionals find themselves confronted with this
urgency, they obviously want to be sure that they can tell a sleeping prob-
lem from parental tensions before proceeding to action. While an indicator
is aimed to deliver ‘actionability,’ it also carries ambiguity. This ambigu-
ity entails that different professionals may conclude to different ‘optimal’
solutions.
Within the RI, similar ambiguities occur in the presentation and interpre-
tation of risks. One professional using the RI reports that it often remains
unclear what others mean by particular risk signals. This is owing to the
 Risk Identities 109
fact that the contents of children’s records are not shared for privacy and
confidentiality reasons.

Professionals have a web-based system, which can send signals auto-

matically. In addition, we have also some ‘soap connections’ [automated
connections]. Risk signals can be sent directly to the RI, if professionals
register a risk in their own files. But they only send signals, no content.
We do not know what the risk signal is about.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

(RI manager, city B)

Finally, the very definitions of risks within the context of ProKid are
themselves ambiguous. This is remarkable, as ProKid is based on standards
that explicitly describe when a child needs to be scored with a white, yel-
low, orange, or red flag. For example, the orange code is to be attributed to
children

“who either have at least once been registered as suspects of a serious

crime (such as animal abuse, arson, sexual offences, public violence or
robbery) or more than once have been reported as missing. Furthermore,
a child can also be profiled with an orange flag, if he/she has been reported
as a suspect five times or more, or he/she it has been registered five times
in relation to various incidents, or if he/she has been reported ten times or
more as a victim or a witness” (Willems and Van der Heijden, 2011, 3).

This orange colour code thus represents a rather heterogeneous set of prob-
lems, the complexity and heterogeneity of which remain implicit. Within
the orange class, the distinctions among perpetrators, victims, and witnesses
vanish. Whereas the colour codes were intended to speed up the assessment
of reports by offering a quick summary, the resulting ambiguity compels
police officers to yet investigate the whole background of an issue.
On top of these direct identifications of risks—in the sense that profession-
als explicitly regard something as a risk and enter it into the system—risks
are also constructed indirectly, through what could be called ‘circumstantial
evidence.’ Such construction is particularly facilitated by systems that are
connected to ProKid, such as the BFFI. The following quote shows how a
risk identification emerges in or ‘between’ these systems. It is the result of
observations made by a police officer at a crime scene which are not directly
related to the crime but may be related to a child:

[W]hen our officers are at an address, and find drugs, [. . .] this infor-
mation becomes registered in BFLE. But if they find a baby bottle in
the kitchen, and ask whether there is a baby; and the inhabitants of
the house answer: ‘No, the bottle belongs to my sister who comes here
occasionally.’ If the officer cannot find hard information about whether
we need to seek care for the baby, or who exactly lives there, to what
 110 K. La Fors-Owczynik and G. Valkenburg
extent the person makes something up . . . and the officer has a bad feel-
ing [about it], and wants to use the information [about the bottle], he
can register it in BFFI. We, ProKid managers, can see that.
(ProKid manager, city C)

This ‘soft’ information will eventually become relevant in risk assess-

ments done by means of ProKid. It thus adds to the whole assemblage of
early warnings and early risk identification, so as to further help preven-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tion. This emergence of risk identification is exemplary for what happens

when multiple institutions become involved: risk identifications propagate
through systems, through which their genesis fades out of scope, and their
content becomes more solidified.
Early detection and a comprehensive view hinge importantly on commu-
nication and the distribution of information.

[The DYHR] helps you to create a better problem overview. You can
provide better advice and guidance based on the information it contains.
[. . .] Previously, we had a hand-written ‘integral paper record’ in which
you needed to search for information manually. Now, you click on a
section and it appears immediately before your eyes in a digital format.
Extracting, sharing and analysing data from the digital dossier is much
faster.
(CB-nurse, city D)

Yet communication and distribution of risk assessments are not just that.
They are mechanisms, mobilized to bring a variety of professionals together
and promote communication among them. In the view of one RI manager,
such promotion of communication and the matching of risk signals are ben-
eficial to rendering multiple problems with children visible. The infrastruc-
ture brings those aggregated problems under the attention of professionals.
As the RI manager explains:

The number of [risk signal] matches increases over time. For us, this
indicates that professionals increasingly contact each other over a
child’s case. In 2008, we had 30,000 signals and 2,277 matches. In
2009, we had 30,313 signals and 3,100 matches. Up until now [July
2010], we have had 30,081 signals and 3,500 matches. [. . .] The idea
behind the system was to arrange a child’s case as a point of connection
between professionals. Thus, children with multiple problems would
become visible earlier, [. . .] receive better help, in case they have mul-
tiple problems within a family. Such complexity needs to be addressed
comprehensively, so that a child will not be ‘ping-pong-ed’ between
professionals, each of them needing to start the child’s assessment from
scratch.
(RI manager, city A)
 Risk Identities 111
However, this leaves untouched the fact that risk-related data potentially
represent different issues for different professionals. For example, the police
officer managing ProKid reported her difficulty in interpreting what the
colour-coded risks symbolize. For example, concerning one incident, the
police officer explained her hesitation in assessing the gravity of the incident
and reducing it to a single risk indicator:

The problem we encounter with ProKid is that we get our colour lists
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in Excel each week, but we also need to read all the reports and what
a yellow or a red colour symbolizes. If there is a report about a child,
who has set something on fire, we only know that it has set something
on fire. We do not know whether it has set a newspaper on fire, or a
whole school building, for instance. So, you have to read the reports to
learn the severity of the crime. This is important for your risk analysis.
(ProKid manager, city B)

It is not straightforward what kind of risk information is to be shared

between the police and youth-care institutions. This is partly the consequence
of the fact that policies to stimulate co-operation between law enforcement
and youth care organizations are still developing. The following manager
doubts the validity of the instruments in practice, even though they are ‘sci-
entifically founded.’ He explains that it would be better to exchange infor-
mation already if files of children are tagged with orange, yellow, and even
white flags:

If I see a child who has just a yellow or white coloured registration—

and I know this does not yet qualify it for sending to the Youth Care
Bureau—but the child’s friend is ‘red’ in the system; then I bring this
extra information into the case discussion within the Youth Care Bureau.
(ProKid manager, city B)

The professional thus seeks his own way in dealing with colour codes and
looks for ways to include information that would otherwise not fit into the
system. Particularly, information from the social environment of the child is
sought to be incorporated.
Numerous examples exist of when the smooth, digital sharing of infor-
mation obscures rather than improves the view of professionals on children.
For example, the distribution of erroneous information shows a perilous
irreversibility:

Everything I register goes instantly into the computer and can be moni-
tored. [. . .] If someone has filled in something wrong, then I need to
receive an email query whether I would take out the wrong information.
Until then, [the wrong information] remains in the file.
(CB-doctor, city A)
 112 K. La Fors-Owczynik and G. Valkenburg
This quote indicates how risks are foregrounded and consolidated by
digital mediation. On the one hand, the registration of data is distributed
in such a way that it becomes instantly visible for professionals in differ-
ent locations. This is supposed to be helpful for the early identification of
risks. On the other hand, erasing these data from files becomes a highly
work-intensive process. Erasure is not automated but involves extensive
e-mail correspondence. This means that, in case of erroneous informa-
tion, the rapid sharing adds confusion rather than clarity, as corrections
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

are always delayed. As the information is predominantly about risks and

the identification of those risks travels much faster than their possible
correction, the system de facto favours the emergence of risks over their
correction.
This is where the systems clearly show a non-neutrality between the nega-
tive realities of risks and potential positive realities that may be articulated
to oppose them. They are designed to store not just any information but a
particular selection of information. This entails that some information can-
not be stored, even if a professional assesses the information as relevant and
even if the data is seemingly neutral or innocent.

RISK PREVENTION BITING BACK

As we have discussed, the DYHR, RI, and ProKid systems all in some way
manage risk identities of children. Also, each system connects multiple
practices, providing channels among different professions. In this section,
we discuss this identity management across professional boundaries as an
‘ontological practice of selection’ (Hacking, 1986, 1991; Bowker and Leigh
Star, 1999; Law, 2008). Whether something counts as ‘normal’ or ‘abnor-
mal,’ ‘deviant,’ ‘at risk,’ or ‘as risk’ always has consequences for the very
world those attributed classes describe, which makes this ‘description’ also
a performative affair. One intended consequence is that these classes offer
a ground for preventive action. Yet not all consequences are intended or
justified, and we will articulate some of the structural issues that carry unin-
tended consequences.
The dominant idea informing the installation of the risk-assessment
systems is that transparency, visibility, and comprehensiveness are key to
prevention of youth problems (Van der Hof, Leenes and Fennell-van Esch,
2009; Keymolen and Prins, 2011). We approach this ‘making visible’ as an
ontological affair, as risk identification also transforms the problems that are
identified. For sure, there are serious and real problems among youth that
merit intervention by social workers and other professionals. Clearly, these
problems must first be identified, if possible before they actually emerge,
before any intervention can be made. However, the shape in which these
problems are articulated does not stand in a one-to-one, unequivocal rela-
tion to what actually happens in the life of the child. We show that the
 Risk Identities 113
DYHR, the RI, and ProKid do indeed identify real problems, but that they
also transform these problems, represent them in non-neutral ways, and
introduce ambiguities of their own.
To identify risks in children and make them actionable, risks are consoli-
dated into indicators: items on which children are scored and which offer
justification for a particular intervention to be made. On the one hand, such
indicators are compact and practical: They reflect the severity of a situation
and instigate interventions. On the other hand, these indicators conceal a
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

lot. A percentage, a colour code or a ‘yes/no’ item hides the work through
which a professional arrives at the indicator. This work includes profes-
sional interpretation and judgment, professional discretion, conversation
with a range of other professionals, intuition, tacit knowledge, contextual
knowledge, and so forth. The necessity of these non-formalized ingredients
is recognized by official reports and guidelines (Inspectie Jeugdzorg, 2008),
yet they are not reflected by the indicator itself. Even if formal standards
are in place that specify how an indicator is to be scored, as with the colour
codes, this formalized part of the assessment is never the whole story (for
otherwise we would not need professionals but only administrative clerks or
even computers to do the assessment).
In addition to identification, the systems are about communication and
distribution of risk assessments. More specifically, communication is to take
place between different practices of youth care, and these practices employ
potentially incompatible approaches. This entails that the identified risks
need to be presented in multiple ways, that they are mobilized for multiple
programmes, that they are to fulfil multiple tasks, and that they will be
imbued with different meanings across situations. Communication in this
sense is not the transparent transport of information but complex work of
translation across boundaries.
The use of indicators as tokens of complex judgments makes such com-
munication possible in the first place. Indicators are ‘mobile’: they are
compact and can be conveniently passed on to another professional. They
are also ‘immutable’: as they are deeply rooted in the practices that estab-
lish them, it becomes hard or even impossible to deny or contest them. As
‘immutable mobiles’ (Latour, 1987), indicators are able to mobilize a wide
range of actors by their alerting characteristic. At the same time, the com-
pactness that allows their mobility entails a continuous need for reinterpre-
tation and adjustment to local knowledge, traditions, codes of conducts,
and objectives.
While crossing disciplinary boundaries, the aforementioned concealment
enables indicators to help keep responsibilities in place. For example, even
though intensive sharing of data between medical and non-medical youth-
care professionals was pursued by connecting the DYHR and the RI, the
Dutch Medical Association KNMG pled for strict limitations on the sharing
of medical information. Obviously, the KNMG held paramount the privacy
and confidentiality of patient data. The ‘best interest’ of the child should be
 114 K. La Fors-Owczynik and G. Valkenburg
the touchstone for whether medical data would be shared (KNMG, 2009).
By reducing such complex information to single indicators, exchange of that
information became possible in the first place.
Offering both ‘transport’ and ‘seclusion’ of information, risk indicators
are perfect examples of what have been conceptualized as boundary objects:
entities that are ambiguous and flexible enough to work in different contexts
but also stable enough to meaningfully connect between those contexts (Star
and Griesemer, 1989; Star, 2010). They deliver some of the promises for
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

which the information systems were installed, namely the connection of dif-
ferent professional practices. At the same time, what they actually convey
is not a simple representation of a pre-existing reality but a particular con-
struction of that reality.
Once these indicators are established, they become harder to modify. To
some extent, they start to live a life of their own, not least because the exact
conditions under which they were established become invisible. While they
are supposed to be representations of chances and possibilities for deviance,
deviance also becomes more ‘real’: It ‘exists’ as representations presented by
the systems. The reality of the issues in part consists of a certain irrevers-
ibility once an indicator has been assigned to a child. As one RI manager
explains:

The problem is that the Reference Index at the time of its introduction
was presented as a ‘Concern reporting system,’ similar to the ‘concern
reports’ police write. You know, it created among professionals an idea
of quite a ‘serious and heavy’ system. Hence, professionals are often
afraid of signalling anything in there, although in practice it is just an
alerting system and only contains data from the municipality register.
This fear to signal is a pity.
(RI manager, city D)

We observed that professionals sometimes become hesitant to enter issues

into the system, as they cannot oversee the effects their data input might
bring about elsewhere. The same professional expresses her own fear of the
potential consequences of risk profiles. Quite delicately, a few minutes later,
she reports:

[O]h while I am showing you this, I have almost put a risk flag on my
own child, that is something I would absolutely not want as a mother.
(RI manager, city D)

The ‘making real’ of risks through the establishment of indicators is not

just in making a solid representation of what is already there. This ‘solidi-
fication’ is not neutral but prioritizes risks: Indeed, they are risk indicators,
and a risk constitutes a negative frame and calls for action. The risk assess-
ment system does not accommodate positive realities or anything else that
 Risk Identities 115
could mitigate or oppose an observed risk. One professional explained her
frustration at not being able to register that a child is doing well:

Many professionals who have already been working for years within
youth healthcare, notice for instance, that they cannot register signals
in the system that reflect that things are going well for the child. I find
it to be frustrating that I can only register negative things, only risks,
and not that the kid is doing well. The system only allows you to record
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

risks. For example, that the child is hyperactive. It is as if children can

only have problems. If things are going well for the child, then I would
like to register that, but I cannot.
(CB nurse, city D)

The system is geared towards prevention, but it de facto disallows profes-

sionals to retract concerns if they think the measures would be too much or
unnecessary.
Although solidifying risks is intended to ease professional communica-
tion across practices, certain translations of risks into technological software
or scripts (Akrich, 1992) can be sources of frustration for professionals. A
quote from an interview with a ProKid managing professional exemplifies
this. For ProKid, changes of address are key. However, the system is not able
to register two addresses with one child. This is potentially problematic:

Address changes in the [ProKid] system are not connected to changes

in the municipal registry.3 When a police officer reports on an address,
[the data] are checked automatically against the municipal registry.
However, if parents divorce and move to two different addresses, and
they arrange co-parenthood, the child is usually registered only at the
mother’s address. So, if a problem occurs at the father’s address, ProKid
unfortunately misses out on that.
(ProKid manager, city B)

The inability to associate two or more addresses with a child is seen to

hamper risk assessments. This is more than just a design flaw. It is a techno-
logical solidification of the social norm that children live with their married
parents. Oddly, it opposes the tendency of risk assessments to extend into
the broader social environment of children beyond their immediate family
(see what follows).

PROLIFERATION OF RISK ASSESSMENT

The preoccupation with risks as actionable, negative realities parallels a pro-

liferation of risk indicators. It is one thing that risk indicators are arranged
in ways that they can be shared easily among different professionals. It is
 116 K. La Fors-Owczynik and G. Valkenburg
another thing that risk assessments and the communication of those risks
seems to engender an ongoing expansion of the number of risk indicators.
This connects to what Van der Ploeg (2001) observes in the context of
pre-natal care. She shows that pre-natal interventions are always susceptible
to the critique that the intervention would have been more effective if it had
been made earlier, when the problem was supposedly still smaller and easier
to encapsulate. This instigates ever more intensive and earlier screening. Van
der Ploeg describes this as the “logic of infinite regression.” In the present
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

case, a similar logic suggests that risks linked to children can be dealt with
better if their registration is more comprehensive and indeed earlier.
We discuss three dimensions of this multiplication of risk indicators.
First, practices of risk assessment increasingly include the family of the child
explicitly in the registration of risks. Second, a similar development is visible
regarding the broader social environment of the child. Third, as suggested,
there is an assumption that earlier identification of risks leads to their being
better manageable. In conclusion, these three developments culminate in the
growth of the number of indicators.

Extending the View Onto the Family

The attention paid to a child’s family in youth-care practices continues to
increase and to be integrated formally in the system. A police officer work-
ing with ProKid, for instance, emphasizes the influence of parents on their
children. As changes of the parents’ relationships may have serious conse-
quences for a child, according to him, these should be part of a child’s risk
assessment. The following example illustrates how a risk profile of a new
family member may affect a child’s colour code in ProKid:4

Many times you see, the mother is divorced and has a child, and the child
gets a report. The child is reported as hyperactive. Then, later the mother
starts a new relationship with a man, who moves in at the same address.
If the mother’s new boyfriend, for instance, already has a police record,
or has been involved in domestic violence, this creates a high risk factor
for the child. The child lives at one address with this man. Therefore, the
yellow colour of the address [in the child’s record] will turn into orange
[. . .] Another example is when the mother has been caught with shop-
lifting, and then she explains that she did the shop lifting because of her
severe circumstances, e.g.: ‘otherwise I am not able to feed and take care
of my child.’ This also constitutes a risk factor.
(ProKid manager, city B)

Beyond criminal concerns, the profiling of parents also plays a central

part during risk assessments in youth healthcare practices. Screening the
health conditions of parents constitutes a significant part of health screening
 Risk Identities 117
processes for children. In the use of the DYHR, the social and financial
statuses of parents increasingly receive attention from professionals, and
they are taken into account when assessing the health risks children may be
exposed to:

You need to know of possible healthcare risks; whether the parents, for
instance, have a [social or financial] problem. Moreover, if the parents
have problems, the broader picture needs to be looked at, and not only
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

the child. It is possible that you have already noted down issues in the
DMO protocol.5 For example, the parents might be unemployed, have
high debts, or the child has been taken out of the family home, etc.
These issues [. . .] importantly influence the child’s development.
(CB nurse, city D)

In some youth healthcare practices, even the extended-familial relations

are considered essential for an adequate risk assessment. A professional
using the DYHR shares her frustration at not being able to properly assess
a child’s risk profile, as assessment of the broader social environment of
the child was not possible with the digital system at the time. This compli-
cated her work, as here knowledge remained incomplete. She could not see
a child’s relation to its stepfather, stepmother or to its grandparents. She
argues that familial relations should be made more visible:

You actually need a structural description of the family. I mean all those
relatives with which the child interacts daily. So far, only the biologi-
cal father and mother are registered in the [DYHR] record, but not
the stepfather, stepmother, stepbrothers and stepsisters. Or, if the child
often stays with grandpa or grandma, then the grandparents need to
be registered in the dossier, too. But the system does not allow for such
registration, unfortunately. Nowadays, there are a lot of divorces. We
should also make those family relations visible, I think.
(CB nurse, city D)

In contrast to this, the RI contains a novel ‘family functionality.’ This

functionality increases the number of family members that can be made rel-
evant in a risk profile of a child. It also increases the number of youth-care
professionals that are connected to the RI. In particular, it ropes in profes-
sionals who are not focusing on the child but on older (even adult) brothers
or sisters and on parents.

Expansion Into the Broader Social Environment

Another dimension of proliferation is the increasing attention paid
to the social environment of the child within practices connected to
 118 K. La Fors-Owczynik and G. Valkenburg
ProKid. Even the behaviour of neighbours can be relevant to a child’s
assessment:

Suppose there was a fight between neighbours. [. . .] Then, I know that

the children of the neighbours have probably been involved in that fight,
too. The police come and make a report. Then, in ProKid, I see that
a child from one address fought with a child from the neighbouring
address. I can then check back in time, and see that these children’s
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

fathers also fought, two years ago. I can then connect these things, and
that is the analysis I do.
(ProKid manager, city B)

This shows how multiple profiles, including the profile of the neighbour
or the father, are drawn together in a ProKid-based analysis. From these
combined risk profiles, risk factors to which the child is exposed are identi-
fied and registered more sophisticatedly in the child’s ProKid record.
In addition to familial and social relations, ProKid even brings forward
a child’s relations to animals. ProKid incorporates behavioural-scientific
research findings holding that the way children interact with animals is
indicative of the risks a child might pose to society in the future:

We also have another, somewhat strange risk factor in our list: animal
abuse. If a child abuses an animal at an early age, then, according to the
scientific findings of Radboud University of Nijmegen, that child has a
higher inclination for sexual abuse at a later age. We relied on scientific
explanation to include such a category into the system.
(Police officer responsible for ProKid project, city B)

Also the child’s friends’ risk profiles are made part of risk assessments.
Links between friends’ risk profiles increase the weight attributed to those
profiles:

If I see a child who has a ‘yellow’ or ‘white’ colour code, and although
officially the case is ‘not yet worth’ being sent to the Youth Care Bureau
(YCB),6 but if a friend of that child is ‘red’ in the system, then I share
this extra information [about the ‘yellow’ or ‘white’ colour-coded child]
during the case discussion within YCB.
(ProKid manager, city B)

These examples demonstrate that the proliferation in social space is

becoming visible as a redefinition of social roles. Children who are diag-
nosed as being ‘at risk’ are increasingly also treated themselves ‘as risk’:
Behavioural-scientific evidence suggests that children who are exposed
to risks run a greater chance of becoming perpetrators themselves. While
this is taken as a reason for additional care, and not, say, for pre-emptive
 Risk Identities 119
punishment, it is indicative of the logic of ever-earlier and more comprehen-
sive detection. What is more, indicators attributed to persons surrounding
the child not only offer more risk indicators against which a child is to be
assessed but also intensify the preventative monitoring of persons who are
seen to fall within the social space of a child.

Extension in Time
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The third dimension of this proliferation of risk assessment concerns time.

One example is provided by the installation of so-called ‘pre-signals’ into the
Reference Index. These are flags that can be added to a child’s record in the
local Reference Index. If such a flag is set, on any new information added
about the child by a different youth care agency, an automatic email is sent to
the institution that originally set the flag. Differently, a ‘real’ match leads to
an email being sent to all those institutions that have signalled a risk about
the same child in the reference index. These pre-signals are not forwarded to
the National Reference Index; neither are they ‘matched’ by any automatic
process. Pre-signals provide opportunities for early warning, legitimated by
the consequences of the pre-signal being comparably limited. However, they
remain part of socio-technical assessments and stimulate earlier evaluations
and consequently interventions. Likewise, the aforementioned advancement
of colour codes based on indirect risks can be seen as a proliferation into
time: issues surface earlier, and this is thought to improve prevention.
Proliferation in time concerns not only the past but also the future. The
RI distinguishes between ‘active’ and ‘passive’ risk signals. The active state
of a signal refers to the period a risk signal can be automatically matched:

Each signal has an expiration date. [. . .] For instance a police signal

remains for 3 months, a signal from an educational institution for 6,
signals of other organizations generally remain for 12 months, and we
keep BYC signals for 24 months in RI. Signals can be matched only
during these periods. [. . .] According to the new Youth Care Act, after
a signal becomes passive, it remains visible for another 5 years on the
web-site. Afterwards, it must be deleted.
(RI manager, city C)

Despite technical restrictions, such as the absence of automated match-

ing, the ‘passive state’ in practice produces an extension of the time a signal
remains actionable for professionals. Within this period, a ‘passive’ signal
keeps presenting a negative reality that should be mended:

This passive signalling allows for another professional to use the source
of the passive signal, in order to acquire information from the profes-
sional about the child.
(RI manager, city D)
 120 K. La Fors-Owczynik and G. Valkenburg
The constructions of pre-signals and ‘passive’ signals in the RI and the
colour-code classifications of risks in ProKid epitomize the proliferation of
risks in time. Risk indicators accompany a child on a longer time span in prac-
tice, thus offering a longer window of opportunity for preventive intervention.

Ever More Indicators

Taken together, these developments lead to a growth in the number of risk
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

indicators. In addition, difficulties of communication across various profes-

sions are often attributed to a certain incompleteness in the available infor-
mation on a child (Inspectie Jeugdzorg, 2008). This incompleteness instigates
the addition of ever more risk indicators, which are thought to be more usable
in different contexts if these indicators are more diverse and comprehensive.
If risk assessments work imperfectly in one context or another, the response
has been to add more indicators so as to increase the likelihood that they
contain information that is meaningful in a particular context. Certainly,
professionals working with these systems are not necessarily uncritical or
naïve about such additions. Quite the contrary, the removal of the fixed list
of invasive events from the BDS described earlier reflects a critical stance,
as does the recognition that building systems are a matter of trial and error.
The mere growth of the number of indicators perpetuates and amplifies
problems we have observed so far. Opacity and ambiguity are more likely to
cause problems if there are more indicators to materialize them. Similarly, the
increasing number of indicators combined with their ‘permanence’ increases
the chances for a child to become and remain qualified as being either ‘at risk’
or ‘as a risk.’ Also, the increasing number of connections among risk factors,
colour codes, risk signals, and risk profiles and their surrounding practices
increases the chance that combinations of those indicators produce a ‘hit.’ As
Van der Hof (2010) argues, while the DYHR eases the circulation of informa-
tion, the resulting escalation of information itself poses new difficulties for
professionals, as the most problematic cases become more troublesome to
sort out. What is more, ‘less serious’ issues increasingly raise alerts, because
the large number of indicators makes the system in a way too sensitive.
All the issues described are the consequence of a sincere pursuit of good
care for children. New risk indicators often result from new scientific evi-
dence and the recognition of real flaws in care practices. However, practices
of youth care and the digital systems through which they operate entail a
proliferation that brings severe problems of their own.

CONCLUSION: IT’S HARD NOT TO BE AT RISK

Modes for identifying and preventing risks in youth care in the Netherlands
are being reconfigured in unprecedented manners. Prevention is performed
by a growing range of youth-care and law enforcement organizations,
 Risk Identities 121
systems, and professionals. The problems they face include various forms
of child abuse, anti-social behaviour, and medical problems. Attempts to
solve these problems occur through the use of comprehensive digital means
and the enrolment of ever more socio-technical practices of youth care,
as well as an extension of what is thought to be relevant. Involved prac-
tices currently include the police, healthcare professionals, social workers,
municipality administrators, educational institutions, and debt-counselling
services.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Underlying all these developments is the sincere aim of improving the

protection of children. First, this aim is translated in a need for ever more
information and improved sharing of information and translated into digital
technology. Second, this need for information is transformed into a pursuit
of ‘making children transparent,’ or to create a ‘comprehensive picture’ of
the child. Third, as we have shown, this picture is subject to the dynamics
that spawns risk assessments in various ways described.
In this study, the practices of youth care seem to be built on the presump-
tion that risks are pre-existing and real and can be made more actionable
by making them visible. While we do not deny the existence per se of such
risks, our analysis shows that in practice a very specific reality of risks is
constructed. Risk indicators proliferate, propagating ambiguities and inter-
pretational problems. Making risks ‘real’ and existent in the form of tech-
nologically mediated indicators also transforms chances of deviance into
real expectations. Particularly, the gradual transformation from a child ‘at
risk’ towards a child ‘as risk’ is to be understood as a consequence of this
proliferation.
This study suggests that the mentioned practices of youth care ‘produce’
risks in two ways. The first refers to the construction of risks as risk indica-
tors. This construction is key to making risks actionable and for initiating
prevention. The second sense raises more concern. The dynamics of con-
solidating risks into indicators and of the proliferation of those indicators
simply raises chances for a child to become subject to a regime of care or
raised attention. In many cases, this will be justified; however, in other cases
it is problematic.
These ‘false positives’ are recognized, but accepted as a ‘cost’ of the sys-
tem. The spokesman of an umbrella organization7 for Dutch youth-care
agencies responds to recent findings that 10% of all suspicions of child abuse
reported to the Advice and Reporting Centre Child Abuse (ARCC, in Dutch:
Advies en Meldpunt Kindermishandeling, AMK)—a sub-organization
within Bureau Youth Care—were false.

JOURNALIST: The Dutch Youth Institute argued today in the daily

newspaper ‘Trouw’8 that families feel stigmatized, if they are falsely
accused [of child abuse]. Do you recognize this?
SPOKESMAN: Yes. If for a small circle of professionals risks remain
known about a child, and if parents return to the GP, they know
 122 K. La Fors-Owczynik and G. Valkenburg
that a suspicion had been registered and investigations took place.
We argue, however, that this price we must pay for the remaining
90%. That is more than 14000 children per year in The Nether-
lands where investigations happen correctly. In these cases, it proves
to be crucial that investigations happen. So this [10% false suspi-
cion] is a consequence that we cannot entirely prevent.
(NOS, 2010)
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Despite this acceptance, our analysis has shown that such negative side
effects are contingent upon particular technological choices, such as the
reduction of complex observations to simple indicators. Side effects include
false positives, the impossibility of indicating that a child is doing well,
and emerging ambiguities of risk indicators that entail misunderstandings
between care practices. Solutions for these effects may not be in finding
better technological solutions, nor would we argue for an abolition of risk
assessments as if they were only bad. Rather, improvement could be sought
in reconsidering how risks are made present and whether technological, pro-
fessional, and organizational elements could be arranged differently, such
that they present those risks in ways that invite reflexivity and awareness of
the side effects showcased in this chapter.

NOTES

1. This chapter contains numerous phrases that we translated from Dutch to

English. We take full responsibility for the translations and choose not to point
this out throughout the entire text. Only with proper names, we add the Dutch
original when possible.
2. The paradoxical point that transparency strictly refers to a particular kind of
‘invisibility’ will not be elaborated further here.
3. In the Netherlands, municipalities maintain central registries of addresses of
all their inhabitants.
4. One assumption underlying ProKid is that incidents with persons living at the
child’s address constitute a risk factor for the child. Therefore, they are to be
reported.
5. The DMO protocol is a standard, early signaling/warning instrument used
within Dutch youth healthcare practices in order to assess children aged 0 to
4 years (Hielkema, de Winter, de Meer and Reijneveld, 2008).
6. According to the ProKid instructions, each case with a ‘red’ or ‘orange’ colour
code needs to be forwarded directly to Bureau Youth Care.
7. www.mogroep.nl/
8. www.trouw.nl/tr/nl/4492/Nederland/article/detail/1087513/2010/02/09/Een-
op-tien-meldingen-kindermishandeling-vals.dhtml

REFERENCES

ABRAHAM, M. et al. (2011) Pilot ProKid Signaleeringsinstrument 12- Geeval-

ueerd. Den Haag: WODC.
ACTIZ. (2012) Flyer: Preventie door Jeugdgezondheidszorg Loont. Utrecht: ACTIZ.
 Risk Identities 123
AKRICH, M. (1992) The De-Scription of Technical Objects. In Bijker, W. & Law,
J. (eds.), Shaping Technology/Building Society: Studies in Sociotechnical Change.
Cambridge (MA): MIT Press, 205–224.
BERKELEY, E. & VAN UDEN, A. (2009) Risicojongeren: Een Bundeling van
Inzichten uit Onderzoek, Beleid en Praktijk over een Effectieve Aanpak. Den
Haag: Nicis Instituut.
BOWKER, G. C. & STAR, S. L. (1999) Sorting Things Out: Classification and Its
Consequences. Cambridge (US), London (UK): MIT Press.
FRISSEN, M., KARRÉ, P. M. & VAN DER STEEN, M. (2011) Zorg door de Staat:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Gevolgen van Gemeentelijke Keuzes in de JGZ. Utrecht: ACTIZ.

GORISSEN, W. (2002) Kennis als Hulpbron: Het Gebruik van Wetenschappelijk
Kennis bij Beleidsvorming in de Jeugdgezondheidszorg voor 4–19 Jarigen. PhD
Thesis. Utrecht: University of Utrecht.
HACKING, I. (1986) Making-up People. In Heller, T. & Wellbery, E. D. (eds.),
Reconstructing Individualism, Autonomy, Individuality and the Self in Western
Thought. Stanford CA: Stanford University Press, 222–236.
HACKING, I. (1991) The Making and Molding of Child Abuse. Critical Inquiry,
17(2). p. 253–288.
HIELKEMA, M. et al. (2008) Effectiviteit van Vroegsignalering binnen het Pro-
gramma Samen Starten. ZonMw Project Website. Available from: www.zonmw.
nl/nl/projecten/project-detail/effectiviteit-van-vroegsignalering-binnen-het-
programma-samen-starten/samenvatting [Accessed 21–09–2013]
INSPECTIE JEUGDZORG. (2008) Jaarverslag 2007. Utrecht: Inspectie Jeugdzorg.
KEYMOLEN, E. & PRINS, C. (2011) Jeugdzorg via Systemen. De Verwijsindex Ris-
icojongeren als Spin in een Digitaal Vangnet. In Broeders, D., Cuijpers, C. M. &
Prins, J.E.J. (eds.), De Staat van Informatie. Amsterdam/Den Haag: Amsterdam
University Press/WRR, 293–348.
KNMG. (2009) EKD Niet Te Breed Toegangelijk Maken. Web News Item. Available
from: http://knmg.artsennet.nl/Nieuws/Nieuwsarchief/Nieuwsbericht-1/KNMG-
Elektronisch-Kinddossier-niet-te-breed-toegankelijk-maken.htm. [Accessed
23–03–2014]
LATOUR, B. (1987) Science in Action: How To Follow Scientists and Engineers
through Society. Cambridge (MA): Harvard University Press.
LAW, J. (2008) On Sociology and STS. The Sociological Review, 56(4). p. 623–649.
MATHAR, T. & JANSEN, Y. (2010) (eds.), Health Promotion and Prevention Pro-
grammes in Practice: How Patients’ Health Practices are Rationalised, Reconcep-
tualised and Reorganised. Bielefeld: Transcript Verlag.
MINISTER VAN VEILIGHEID EN JUSTITIE. (2012) Rechtstaat en Rechtsorde.
Brief van de Minister van Veiligheid en Justitie aan de Voorzitter van de Tweede
Kamer. Den Haag, 29 juni 2012, kst-29279–147.
NEDERLANDS CENTRUM JEUGDGEZONDHEID. (2011) Inhoud voorstel 127.
Available from: www.ncj.nl. [Accessed 18–12–2012]
NEDERLANDS CENTRUM JEUGDGEZONDHEID. (2013) Basisdataset Jeug-
dgezondheidszorg Versie .3.2.1. Available from: www.ncj.nl/informatisering-jgz/
basisdataset. [Accessed 03–10–2013]
NOS. (2010) Veel Onterechte Meldingen over Kindermishandeling. Web News
Item. Available from: http://nos.nl/audio/135462-veel-onterechte-meldingen-over-
kindermishandeling.html. [Accessed 21–03–2013]
NOUWT, S. & HOGENDORP, J. (2010) Landelijke Verwijsindex Risicojongeren
Ingevoerd. Medisch Contact, 2010(31–32). p. 1560.
RADARADVIES. (2011) Preventief Risicosignaleringsinstrument Jongeren en Over-
last . Brochure. Available from: www.radaradvies.nl/producten/preventief_
risicosignaleringsinstrument_jongeren_en_overlast/521. [Accessed 21–03–2013]
RIJKSOVERHEID. (2005) Wet op de Jeugdzorg. Available from: http://wetten.over-
heid.nl/BWBR0016637. [Accessed 25–04–2012]
 124 K. La Fors-Owczynik and G. Valkenburg
RIJKSOVERHEID. (2008) Wet Publieke Gezondheid. Available from: http://wetten.
overheid.nl/BWBR0024705. [Accessed 12–04–2012]
RIJKSOVERHEID. (2010) Wet Verwijsindex risicojongeren van kracht op 1 augus-
tus 2010. National Government Press Release 09–07–2010. Available from:
www.rijksoverheid.nl/documenten-en-publicaties/persberichten/2010/07/09/
wet-verwijsindex-risicojongeren-van-kracht-op-1-augustus-2010.html. [Accessed
15–03–2012]
ROUVOET, A. (2007) Alle kansen voor alle kinderen. Beleidsprogramma Jeugd en
Gezin 2007–2011. Policy Document. Den Haag: Ministry for Youth and Family.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

STAR, S. L. (2010) This Is Not a Boundary Object: Reflections on the Origin of a

Concept. Science, Technology & Human Values, 35(5). p. 601–617.
STAR, S. L. & GRIESEMER, R. J. (1989) Institutional Ecology, ‘Translations’ and
Boundary Objects: Amateurs and Professionals in Berkeley’s Museum of Verte-
brate Zoology, 1907–39. Social Studies of Science, 19(3). p. 387–420.
TWEEDE KAMER DER STATEN GENERAAL. (2012) Vaststelling van de Begrot-
ingsstaten van het Ministerie van Veiligheid en Justitie (VI) voor het Jaar 2013.
Den Haag: Tweede Kamer der Staten General.
VAN DER HOF, S. (2010) Het Elektronisch Kinddossier: Kansen en Kantekeningen.
In Munnichs, G., Schuijff, M. & M. Besters (eds.), Databases: Over ICT-beloftes,
Informatiehonger en Digitale Autonomie. Den Haag: Rathenau Istituut, 54–62.
VAN DER HOF, S., LEENES, R. & FENNELL-VAN ESCH, S. (2009) Framing
Citizens’ Identities: The Construction of Personal Identities in New Modes of
Government in the Netherlands. Research on Personal Identification and Iden-
tity Management in New Modes of Government. NWO, Network of Networks
Programme, Tilburg: TILT, Tilburg University.
VAN DER PLOEG, I. (2001) Prosthetic Bodies: The Construction of the Fetus and
the Couple as Patients in Reproductive Technologies. Dordrecht: Kluwer Aca-
demic Publishers.
WILLEMS, M. & VAN DER HEIJDEN, M. (2011) Provinciale Pilot Vroegsignalering
risicojeugd. Available from: www.nogveiligerhuis.nl/UserFiles/File/thema_jeugd/.
pdf. [Accessed 21–03–2013]
 6 Swimming in the Fishbowl
Young People, Identity, and Surveillance
in Networked Spaces
Valerie Steeves
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This chapter draws on the findings of Media Smart’s Young Canadians in

a Wired World research project in order to examine young people’s experi-
ences with identity and surveillance in networked spaces.1 Canada is an
interesting example to examine, because the Canadian government was the
first in the world to connect all its schools to the Internet in 1999 and,
within a decade, home access to networked technologies approached uni-
versality (CRTC, 2010). Since that time, Canadian youth have continued
to be early adopters of various networked communications technologies,
the most recent example being social media. The statistics on Facebook use
are illustrative of this trend; Canada has the largest per capita participation
on Facebook in the world (Oliveira, 2012) and 95% of 17-year-old Cana-
dians have a Facebook account (Steeves, 2014). As such, Canadian youth
are among the most wired in the world and have fully integrated networked
technologies into their schooling and social lives.
From as early as 2000, the young Canadians we spoke with were using
the technologies of the day—primarily chat rooms and instant messaging—
to try on new identities and connect with friends. Our research participants
also celebrated the privacy they found online; they believed themselves to
be outside the gaze of their parents and teachers and accordingly embraced
the Internet as a safe place to explore and experiment (Media Awareness
Network, 2001). As new technologies emerged over the next 10 years, other
researchers reported similar findings. Chat rooms (Mendoza, 2007), instant
messaging (Steeves, 2005), personal home pages and blogs (Stern, 2004),
cell phones (Ito, 2005), and social networking sites (boyd, 2007; Living-
stone, 2008; Shade, 2008) were each, in turn, appropriated by children and
reconstituted as socio-technical spaces; children used these spaces to connect
with peers and engage in a reflexive project of constructing the self, away
from the watchful eyes of adults.
When we returned to the field in 2012 to 2013, the children we inter-
viewed reported that they continue to use the latest networked technolo-
gies, including social media, smartphones, networked MP3 players, tablets,
and gaming platforms to experiment with their identities and connect with
their friends. However, the online spaces they now frequent are structured
 126 Valerie Steeves
by pervasive monitoring, particularly from peers, parents, family members,
teachers, and school administrators, and the various forms of surveillance
they encounter online pose challenges for them when it comes to using the
technology for identity play.
This chapter discusses the ramifications of our findings, in three sections.
First, I explore the types of visibility and lateral surveillance our participants
experienced with peers and the kinds of strategies they relied on to manage
their online personas. Second, I examine the dual face of parental surveil-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

lance as care and control and the complex negotiations that the children
and parents we talked to undertook with respect to both access to and use
of online media. Third, I turn to the panoptic surveillance our participants
experienced at school and explore their perceptions of the impact of this
surveillance on their ability to use networked technologies to enhance their
learning.

PEERS, LATERAL SURVEILLANCE, AND

IDENTITY-MANAGEMENT STRATEGIES

As noted, our participants in 2012 to 20132 enjoyed the ways in which net-
worked communications enable a playful engagement and experimentation
with different ways of being online. For example, our 11- and 12-year-old
participants searched for information to learn more about things they would
encounter in the future, like high school and jobs, and saw this as a safe way
to ‘rehearse’ the roles they would play as teens and adults. Our teenaged
participants relied heavily on various media to express themselves, commu-
nicate with friends, explore their interests, keep in touch with family, and
generally figure out who they wanted to be when they grew up.
A large part of the appeal of networked technologies for all age groups
was the visibility they provide. A number of researchers (Peter, Valkenburg
and Fluckiger, 2009; Phillips, 2009; Shade, 2011; Draper, 2012) suggest that,
by monitoring how others respond to their online personas, young people
are able to evaluate their various identity performances. This allows them to
co-produce their subjectivity through their interaction with others who mir-
ror back their performances for them to see (Mead, 1934; Goffman, 1959).
From this perspective, their online identities are fluid (Giddens, 1991) and
tied to an ongoing project of “writing the self into being” (boyd, 2007).
The concept of lateral surveillance (Andrejevic, 2005), in which peers
monitor peers and are monitored in turn, is a particularly important ele-
ment of this kind of reflexive identity construction and performance, espe-
cially for our teenaged participants. Ninety-five percent of the 17-year-olds
we surveyed had a Facebook account, and 72% read or posted on friends’
social network sites at least once a day or once a week. Thirty-nine per-
cent of all survey respondents indicated that they slept with their cell phone
next to them in case they received messages during the night. The teenagers
 Swimming in the Fishbowl 127
we interviewed saw this as a way to monitor the ‘drama’ that unfolded
when they were not at school, and indicated that they used their networked
devices to keep an eye on peers throughout the day and night.
Much of this lateral surveillance was playful in nature. Our participants
knowingly slipped in and out of the online gaze to play jokes on each other
and enjoy themselves. Pranking—setting online traps for someone or misdi-
recting them to a joke site—was a common activity,3 especially among boys.
Although pranking was “just for fun” (Steeves, 2012a, 28), it was also a
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

way of acquiring and demonstrating the skills they needed to distinguish

between ‘real’ and ‘fake’ online representations. Another form of playing
with the gaze was to impersonate a friend by using his or her login and then
posting comments under his or her name.4 Our interview participants indi-
cated that this typically involved in-jokes or poking fun at each other and
was a regular—though sometimes trying—part of their online interactions
with their friends.
However, lateral surveillance also exacerbated the consequences of a
failed performance (e.g. a picture in which they appeared foolish or did
not look their best, or a sext—a sexually explicit text message—that was
forwarded to others within the teen’s peer network) because failed perfor-
mances could be seen, copied, and forwarded across a range of techno-
logical platforms used by their peer group. In other words, although being
seen online was part of the fun, being seen badly was a significant risk.
Our participants accordingly developed a number of strategies in order to
minimize this risk.
In this regard, decisions about what photographs of themselves to post
were particularly important. Our qualitative participants told us that they
put a great deal of care and attention into selecting these images—as one
15- to 17-year-old girl put it,

I just don’t take stupid pictures that I know could ruin my reputation,
or something.
(ibid., 33)

Of our survey respondents, 91% reported that they used privacy settings
to block someone from seeing the photos they post, and most of the persons
blocked were known to the respondents. Of the people being blocked, 31%
were friends, 20% were people they had stopped being friends with, and
20% were people they knew but with whom they were not friends.
They also closely monitored their peers’ sites to see how they were being
portrayed by them. Forty-five percent of our survey respondents indicated
that they had asked someone to delete something that person had posted
about them because they did not want someone else—especially friends
(21%)—to see it. The young people we interviewed told us they routinely
‘de-tagged’ photos of themselves so they could retain control over the dis-
tribution of their image. Accordingly, this lateral surveillance enabled them
 128 Valerie Steeves
to see how they were being portrayed and to take proactive steps to protect
their online image. As one 15- to 17-year-old boy said,

I don’t want . . . myself to be in someone else’s phone or computer . . . Or,

like, other people showing other people, being like, “look at this!’”
(ibid., 32)

Our interview participants also relied on friends to help them manage

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

their online personas. It was generally understood that friends do not

post embarrassing or compromising photos of friends online. Moreover,
if someone did post an embarrassing photo or say something mean about
someone, that person’s friends were expected to go online and restore his
or her reputation. For example, 11- to 12-year-old Emma told us that an
acquaintance had posted an unflattering picture of her on Facebook, and
people were making mean comments about her appearance. She texted
her friends, who immediately went to the site and posted comments like,
“‘no, Emma looks cool, she’s awesome, she’s so brave’ and stuff, and
[Emma] was like, ‘I love you guys’” (ibid., 32). Although lateral surveil-
lance exacerbated the impact of the mean comments because they were
seen by so many people, it also alerted Emma to the attack and enabled
Emma’s friends to repair her online reputation by visibly coming to her
defence.
Interestingly, our participants proactively purged particularly candid
or bad photographs of themselves before they could be posted online in
order to keep these images away from the lateral surveillance of their peers.
Although they routinely allowed friends to take photos of them when they
were goofing around, they would later go into their friends’ phones or cam-
eras to delete them. As Emma and Taylor explain:

EMMA: Cause . . . if there’s a picture of my goofing off, like making a

funny face, you don’t want everyone to see that, it’s between you
and your friends.
TAYLOR: Yeah, other people, other people probably all make fun of you,
and then that’ll stay around for a while because that’s happened before.
EMMA: Yeah, only your friends understand why you’re doing it . . .
TAYLOR: Yeah, and then everyone else, like, sees it and then they’re kind
of like, “oh, why are you doing this?”
(ibid., 32)

Failure to allow someone access to a device so he or she could delete a

photo placed a strain on the friendship. A number of our participants told
us that, in those circumstances, it was acceptable to break into the person’s
phone or social media account without his or her permission to delete the
photo. If they were unsuccessful and the person posted the photo or made
negative comments about them online, the friendship was at an end.
 Swimming in the Fishbowl 129
In this sense, lateral surveillance was a form of self-protection; by rou-
tinely monitoring what photos of them were held by others, they were
able to intervene when necessary to prevent the distribution of unwanted
images. On the other hand, possession and non-distribution of a potentially
embarrassing photo was seen as a sign of close friendship. For example, dur-
ing one 15- to 17-year-old qualitative group session, Bridget started teasing
her best friend Maddy about a particularly embarrassing photo of Maddy
she had on her phone:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

BRIDGET: [Giggling] Look at the picture that I have of you. [Laughter

from everyone even though Bridget only showed it to Maddy]
ALICIA: But it’s not like something I wouldn’t send to somebody, I
wouldn’t post pictures that I have of people . . .
MADDY: Oh, it’s nothing dirty. [Laughter] It’s just a lot of makeup
with . . . facial hair . . .
BRIDGET: I wouldn’t post it on Facebook, I’m not like that.
FACILITATOR: By keeping it on your phone you’ve got it, but it’s not as
public as Facebook?
MADDY: Yeah.
(ibid., 33)

Keeping something off social network sites and away from the lateral
surveillance of peers was therefore a way of signalling closeness and
trustworthiness.
Lateral surveillance among peers accordingly played a variety of func-
tions. It provided access to an audience for the purposes of identity construc-
tion, in the Meadian sense, and opened up spaces for fun and playful teasing.
Although it magnified the potential consequences of a failed performance,
it also enabled our participants to monitor how they were represented by
others and respond in ways that allowed them to publicly repair their repu-
tations. It further served to demarcate intimacy when images shared with
trusted friends were kept out of the surveillant gaze and stored ‘privately’ on
media such as mobile phones.

PARENTS AND THE COMPLICATIONS OF SURVEILLANCE

AS CARE AND CONTROL

Parental monitoring was one of the most common forms of surveillance

our participants discussed. However, both the parents and young people we
talked to were ambivalent about it and entered into complex negotiations
around access to and use of online services, social media accounts, and cell
phones.
From the parents’ perspective, online monitoring was a form of care.
Almost all of them expressed concerns over the possibility that their child
 130 Valerie Steeves
could be hurt online,5 and most required full password access to their child’s
various accounts so they could keep an eye on him or her to keep him or her
safe. Although the nature of online risks was ill defined and parents could
not point to specific harms, they felt that they were forced to monitor their
children to protect them from stalkers and other ill-intentioned strangers.
However, the parents who did this also worried that it would negatively
affect their relationship with their child, because monitoring could poten-
tially signal that they did not trust their child to act appropriately. As Rooney
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

notes, a child learns how to enter into trusting relationships with others by
relying on “the good will of others . . . such as parents, caregivers, friends
and strangers in a variety of ways to care for and protect them” (Rooney,
2010, 346). Having opportunities “to trust and be trusted” are an integral
part of learning “how to be with others in a way that supports [children’s]
capacity to live and live in a meaningful way.” Accordingly, children need to
learn to “trust with good judgment,” actively negotiating situations in which
they can rely on others and situations in which distrust may be an appropri-
ate response to some potential harm (Steeves, 2012a, 347). She concludes:

Where there is a climate of fear about public spaces, it is possible to

see how parental fears might lead to a tendency to use tighter mecha-
nisms of control . . . However, such an approach, particularly where it
is an overreaction to the risks involved, makes it difficult for children
to negotiate an appropriate, realistic and constructive balance between
trust and risk . . . and as a result the opportunities for a child to negoti-
ate terms of freedom or to subvert the controls that are placed on them
rapidly diminish.
(ibid., 350)

Interestingly, our 11- to 12-year-old survey respondents were much more

comfortable with parental monitoring than were their older counterparts
and accepted it as a form of care. However our qualitative participants told
us that much of this comfort was rooted in the fact that they found posting
personal information on social media sites or talking with strangers ‘boring.’
As one girl summarized,

Wilhelmina: Ben normalement nous, à notre âge, on a pas vraiment

besoin de . . . ben y’a des filles au secondaire qui ils ont besoin mais
comme moi admettons j’ai comme cinq amis—Well normally, us, at our
age, we don’t really need to [use social media] . . . there are high school
girls who need to but me, I admit, I have like five friends.
(ibid., 16)

This is consistent with Livingstone’s (2009) observation that younger

children construct their online identities through display, as opposed to ado-
lescents, who construct their identities through their social connectedness to
 Swimming in the Fishbowl 131
others. Surveillance was accordingly more acceptable to our younger par-
ticipants6 because they were not yet predisposed to break away from their
identity within the family and begin to explore who they were in connection
to peers.
Nonetheless, all our participants indicated that they took steps (e.g. using
privacy settings to limit what parents could see or clearing histories on
shared computers) to avoid parental monitoring. Our teenaged participants
were highly adept at using technical controls to evade ‘lectures.’ One 15- to
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

17-year-old girl’s comment that, “My mom keeps on [posting] me, ‘You’re
on Facebook! Get off! Do your homework!’ And I’m like . . . de-friend”
(17), was met with both commiseration and a flurry of stories about evad-
ing the parental gaze online. Even many of our youngest survey respondents
felt that parents should not force their children to friend them on social
media sites (56%) or read their texts (44%)7 and took steps to avoid being
watched. Sharing their passwords was one thing; having their parents con-
stantly looking over their shoulders was another. This is in keeping with
Valentine’s (2004) observation that young people make their own decisions
about risk and disrupt the kinds of controls parents put in place to protect
them from risk.
But concerns about online privacy from parents were particularly acute
for our teenaged participants. Communication with friends was a central
part of their lives, and they enjoyed the way networked tools gave them a
deep sense of connection with their friends (Licoppe, 2004; Shade, 2011).
They also articulated a strong need for autonomy from parents so they could
better explore the “public–private boundaries of the self” (Peter, Valkenburg
and Fluckiger, 2009, 85) and be free to experiment in ways that were dif-
ficult in offline contexts (Livingstone, 2009, 91). They felt that constant con-
nection to parents also made it difficult for them to accomplish the central
tasks of adolescence to

renegotiate their familial relationships . . . seek to define themselves

within a peer group . . . [and] venture out into the world without paren-
tal supervision.
(Draper, 2012, 223)

As one 13- to 14-year-old in Toronto put it,

There should be a point where parents will just like, leave you alone and
not have to know every single thing about you. Like I get, the protection
side, but they don’t need to know every single thing about you.
(Steeves, 2012a, 18)

Because of this, our teenaged participants were much more likely to

seek out private online spaces in which to interact with their peers. They
articulated a clear need for privacy and explicitly linked it to their need for
 132 Valerie Steeves
autonomy and independence, drawing on Samarajiva’s notion of privacy as
“the capacity to implicitly or explicitly negotiate boundary conditions of
social relations” (Samarajiva, quoted in Livingstone, 2009, 110). However,
many parents of teens told us they had to increase their level of vigilance
as their child entered adolescence because this age group was more likely
to do or say something inappropriate online. Moreover, they worried that
the impact of such a misstep would be magnified because so many people
would be able to see it. They therefore recruited other family members, such
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

as older siblings or cousins, to monitor their child online and report back
to them, creating a hybrid form of surveillance that was both hierarchical
and lateral.
Although all our parent participants were uncomfortable with this level
of scrutiny, almost all of them felt they had little choice because surveillance
was perceived to be a necessary tool to protect their children from their own
poor judgment. This exchange was typical:

Who was it that said you had spies out there? I have nieces [who] will
write to [my daughter], even call me to say, “uh, tell her to change . . . her
wall, her status, or whatever,” so that’s good.
(Steeves, 2012a, 14)

Others spoke of the need to read every text, every social media posting,
and every email to make sure their child was not in a position to make a mis-
take. Accordingly, just when teenagers were asking for more freedom from
the parental gaze—precisely so they could make their own mistakes—many
parents were increasing the level of online monitoring in order to shut down
any possibility of their child behaving poorly. Parental surveillance accord-
ingly moved from monitoring for protective care of a child at risk of harm
from ill-intentioned others to monitoring for behavioural control of a child
that was a source of risk himself or herself.
From our teenaged participants’ perspective, this kind of surveillance cre-
ated a great deal of conflict and interfered with their family relationships. As
one 15- to 17-year-old said,

“I blocked my little brother, he’s like a little spy for my mother” (18).
And, after cousins of one of our 13 to 14-year-old participants ‘snitched’
on her, she told us, “the same night I go and delete them . . . then [my
mom] gets mad, she’s like ‘don’t delete your family members.’ I’m like,
well, tell them to stop stalking me.”
(18)

Interestingly, it was also not in keeping with the kinds of monitoring

they experienced in offline situations. Whereas parents were comfortable
with their child’s judgment in general, they were much more concerned
about their child’s actions when networked technologies were involved. The
 Swimming in the Fishbowl 133
teenagers we interviewed found this confusing, because online (and sur-
veilled) actions were perceived by parents to be more dangerous than offline
(and unsurveilled) actions. One 13- to 14-year-old girl summarized:

My mom trusts me enough to, like, actually bring a guy home, like one
of my guy-friends home? But she doesn’t trust me enough to like, have
him up on Facebook, which kind of makes me depressed.
(19)
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

They told us that this kind of surveillance made it more difficult for
them to express themselves online and explore their various roles as friend,
romantic partner, and emerging adult because online performances for their
peer audience would often unintentionally become visible to their family
audience. This meant they were often held to account for comments that
were taken out of context or seen as more dangerous because they were
made online or by text. This in turn disrupted their “ability to disclose pri-
vate information in appropriate ways and settings” (Peter, Valkenburg and
Fluckiger, 2009, 83), both with their peers and with their family members.
Rooney’s work on trust is particularly relevant here. She argues that young
people must not only learn to trust others; they must also learn how to be

“trusted by others to be responsible, to take control and do things in

ways that extend their skills and competencies” so they can develop into
“competent, confident and active human agent[s].”
(2010, 344, emphasis added)

From this perspective, surveillance as control teaches children the wrong

lessons: family members who monitor you for your own protection do not
trust you to make the right decisions and are not to be trusted because they
exaggerate risks and interpret actions out of context.
At the same time, our participants were empathetic to their parents’ con-
cerns and acknowledged that parents were only trying to protect them from
harm. Moreover, all age groups of survey respondents indicated that parents
were a helpful resource for solving problems. Younger children tended to see
parents as a first response—three quarters of 10-year-olds indicated they would
ask their parents for help first if someone was mean or cruel to them online—
but more than one quarter (27%) of 15- to 17-year-olds continued to rely on
parents if self-help strategies (including ignoring the problem, confronting the
person who said it face to face, and asking friends for help) were unsuccessful.
The presence of parental rules about online activities (as opposed to parental
monitoring) also appears to have a protective effect for all age groups; survey
respondents who reported that they had house rules against risky online behav-
iour were statistically less likely to participate in that behaviour.
Interestingly, our teen participants who were not monitored were also the
ones who were the most likely to willingly share aspects of their online lives
 134 Valerie Steeves
with their parents and to go to them for help. Trust for them was mutual—
their parents trusted them to behave appropriately, and they trusted their
parents not to over-react (Steeves, 2012a, 19).
Kerr, Stattin, and Trost (1999) report that this kind of voluntary disclo-
sure increases trust for parents as well as for children. The more a child is
able to express feelings and talk about problems with a parent, the more the
child trusts the parent. Conversely, the more a child spontaneously reveals
about his or her daily life to a parent, the more the parent trusts the child.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Surveillance as control, on the other hand, diminishes measures of trust

and, interestingly, also correlates with higher, not lower, rates of anti-social
behaviour (Stattin and Kerr, 2002). So although parental surveillance may
be well intentioned, it does not provide a “realistic form of protection”
(Rooney, 2010, 351) for the child. It also may reduce the opportunities for
the child to explore other independent identities with a sense of competence
and confidence (Steeves, 2012a, 351).
One story in particular illustrates the importance of adult support as a
child begins to explore the world outside the home and competently navi-
gate new identities. A 13- to 14-year-old boy told us that he was being
‘stalked’ online by a 13-year-old girl who wanted to be his girlfriend. After
she posted a comment on his Facebook wall that “btw, I’m not a virgin,”
he was subjected to cruel comments and uncomfortable teasing from his
friends. He did not know what to do, so he showed the posts to his mother.
His fear was that she would take the matter out of his hands and, in fact,
she immediately wanted to intervene. However, he articulated his concerns
about exacerbating an already uncomfortable situation, and they negotiated
an appropriate strategy that met his need for protection in a way that also
respected his need for independence. He removed the comments from his
wall, and the teasing stopped (ibid., 20).
This boy’s problem was not resolved by parental surveillance—in fact, the
mother only knew of the situation because the boy told her about it. It was
resolved by dialogue structured by mutual trust and a willingness on the part
of the parent to provide the child the opportunity to solve his own problem,
with her guidance. This illustrates that the negotiations between parents and
children around technologies are complex and multifaceted and implicate
broader concerns about the competing needs to allow a child freedom and
exert control over the child.

SCHOOLS, PANOPTIC SURVEILLANCE, AND A LOSS

OF PRIVACY

Monitoring at school was pervasive and often posed problems for the young
people we interviewed as they sought to complete their schoolwork. Not
only was certain content blocked, but their online communications with
peers were closely scrutinized, and they were held to account for interactions
they interpreted among themselves as harmless. Moreover, they were given
 Swimming in the Fishbowl 135
little to no opportunity to explain themselves if one of the technical controls
indicated they had done something they were not allowed to do.
Our participants were most vituperative about this kind of panoptic con-
trol. They both actively resisted it (e.g. sharing technical fixes to get around
filters) and questioned the school’s need to police their every word, to make
sure they did not ‘swear’ or say something ‘inappropriate.’ From their per-
spective, the school should give them access to networked technologies with-
out placing them under surveillance and, instead, rely on teachers to help
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

them learn how to use them. One 13- to 14-year-old explained:

But then, again, we’re supposed to write every day everywhere else,
except for school. What’s the big deal if we do exactly what we do
at home, at school? . . . Teachers should be allowed to read what we
write, and if it’s inappropriate, they can make us take it down, but they
shouldn’t just block us out from it, that is our own right.
(Steeves, 2012a, 21)

Interestingly, the key informant teachers we spoke to agreed. From their

perspective, school filters, acceptable use policies, and keystroke loggers
made it incredibly difficult to productively use networked technologies in
the classroom because they decreased learning opportunities and restricted
the teachers’ ability to teach. As one teacher from Ontario noted,

For me it would be so much easier if it were just unblocked and the

Board trusted the teachers to show the kids how to actually use this
material. That’s how I’d prefer to teach.
(Steeves, 2012b, 12)

This form of panoptic surveillance also took away the teachable moments
in which teachers could help students learn to act as good digital citizens. A
teacher from Western Canada put it this way:

It’s not like all of a sudden you hit 18, and now you can have autonomy.
I mean, children do not learn to make good choices by being told what
to do and follow instructions. And, unfortunately, they have to be given
the opportunity to make bad choices as often as good choices. And they
need adults to be the saving, caring allies that we need to be to help them
make [good choices], to learn from their mistakes.
(12)

A teacher from Eastern Canada suggested that surveillance can interfere

with the privacy of the classroom and make it more difficult to create a safe,
confidential learning environment:

When the conversation was intended to provoke intellectual curios-

ity and you’re expected to take intellectual risks and really share and
 136 Valerie Steeves
expose your thoughts about a particular text or event, to have that
trust, that collaboration, that safe learning environment sort of ruined
from access to technology or a recording device or posting online, I
don’t know if you could overcome that to build [the] kind of classroom
[where students feel safe to experiment].

A teacher from Northern Canada agreed. He told us students

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

need to trust you in order to take risks . . . being able to answer ques-
tions and know that if I get a wrong answer, that’s okay, they won’t
laugh or make fun of me. That’s risk taking for some students. That’s
a big risk.

Interestingly, both the young people and the teachers we talked to indi-
cated that the panoptic surveillance they both experienced in school was the
most difficult to negotiate, because they had very little control over the gaze
or input into how their actions were interpreted by the school administra-
tors that monitored the use of the system.

CONCLUSION

Our research findings suggest that the young people and adults we talked
to have a complex relationship with networked technologies and continu-
ally negotiate the degree of monitoring young people are subjected to in the
socio-technical spaces they inhabit. Young people are attracted to networked
spaces because of the visibility they provide, but the monitoring they experi-
ence in these spaces makes it difficult for them to draw lines between their
various audiences and attain the level of privacy they desire. That lack of
privacy also detracts from their ability to enter into relationships of trust—
particularly with parents, teachers, and other family members but also with
friends—which in turn complicates their attempts to meet the developmen-
tal objective of individuating and ‘growing up.’
Moreover, young people’s experiences of monitoring involve different
kinds of surveillant relationships, each of which offers different opportu-
nities for freedom and control. Lateral surveillance among peers is highly
nuanced, both opening up opportunities to reflexively perform a variety
of social roles and potentially magnifying the consequences of failed per-
formances. Self-representations are accordingly carefully crafted on and
through social networks, and young people rely upon friendships within
these networks to help them manage how they are perceived and under-
stood, particularly in these contexts.
Negotiations between young people and parents are particularly com-
plex, as parents seek to balance surveillance as care with surveillance
 Swimming in the Fishbowl 137
as control. This balance becomes more challenging when children enter
adolescence—when young people seek a greater degree of privacy from
parents so they can explore their identity outside the family. However, this
is also when many parents feel they should increase the level of monitoring
to ensure that their children do not engage in relationships or practices that
they see as inappropriate or dangerous. Again, failed performances are seen
as more problematic because they take place on networked media and can
be reproduced and distributed widely. In addition, the ubiquity of being
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

watched through social networks by a variety of audiences makes it more

difficult for a teen to craft a specific persona for peers within that social
network without disrupting the expectations of family members. This dif-
ficulty is magnified when parents see these behaviours on social networks
as particularly risky.
Monitoring in school most closely aligns to the panoptic conception of
surveillance, and young people participate in a number of strategies to resist
this kind of watching. Interestingly, both the young people and the teach-
ers we spoke to lament the ways in which panoptic surveillance invades the
privacy of the classroom and detracts from the relationships of trust that are
at the heart of learning.

NOTES

1. The Young Canadians in a Wired World project began in 2000–2001 when

we interviewed parents and children and surveyed approximately 5,500 Cana-
dian students aged 10 to 17 to examine children’s use and perceptions of the
Internet. In 2004 to 2005, we conducted a similar study, broadening the tech-
nology to other forms of networked communications, including cell phones
and gaming platforms. In 2012 to 2013, we again returned to the field but
added a series of interviews with teachers to get a better understanding of
the impact of the full range of networked technologies in the classroom. This
chapter draws on the most recent data from 2012 to 2013, which includes the
results of interviews with 10 key informant teachers and 12 qualitative group
sessions (four each in Calgary, Toronto, and Ottawa) with a total of 66 young
people aged 11 to 17 and 21 parents of children and youth aged 11 to 17, and
a quantitative survey of 5,436 children and youth aged 10 to 17 from across
the country. For a full report of each phase of YCWW, see http://mediasmarts.
ca/research-policy.
2. All information about and quotes from our 2012 to 2013 qualitative research
are taken from Steeves, 2012a. The survey results are taken from Steeves,
2014.
3. Twenty percent of our survey respondents reported pranking online at least
once a day or once a week, and 28% reported pranking over a cell phone
(Tables 8, 9).
4. Although only 20% of 11-year-old survey respondents reported that they did
this, the proportion rose to 50% by age 17.
5. Vandoninck, d’Haenens, and Roe (2013) report similar concerns among Euro-
pean parents, independent of socio-economic status.
 138 Valerie Steeves
6. Seventy-nine percent of children aged 10 agreed with the statement that par-
ents should keep track of their children online all the time, compared to only
23% of 17-year-old respondents.
7. The percentages increase with age, to 77% of 17-year-olds and 83% of
17-year-olds respectively.

REFERENCES
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ANDREJEVIC, M. (2005) The Work of Watching One Another: Lateral Surveil-

lance, Risk and Governance. Surveillance & Society, 2(4). p. 479–497.
BOYD, D. (2007) Why Youth (Heart) Social Network Sites: The Role of Networked
Publics in Teenage Social Life. In Buckingham, D. (Ed.), Youth, Identity, and
Digital Media. Cambridge (MA): MIT Press, 119–142.
CANADIAN RADIO-TELEVISION AND TELECOMMUNICATIONS COM-
MISSION. (2010) Communications Monitoring Report, July 2010. www.crtc.
gc.ca/eng/publications/reports/policymonitoring/2010/cmr2010.pdf, accessed
22–09–2013.
DRAPER, N. (2012) Is Your Teen At Risk? Discourses of Adolescent Sexting in the
United States. Journal of Children and Media, 6(2). p. 221–236.
GIDDENS, A. (1991) Modernity and Self-Identity. Cambridge: Polity Press.
GOFFMAN, E. (1959) The Presentation of the Self. New York: Basic Books, Inc.
ITO, M. (2005) Mobile Phones, Japanese Youth, and the Re-placement of Social
Contact. Computer Supported Cooperative Work, 31. p. 131–148.
KERR, M., STATTIN, H. & TROST, K. (1999) To Know You Is to Trust You: Par-
ents’ Trust Is Rooted in Child Disclosure of Information. Journal of Adolescence,
22. p. 737–752.
LICOPPE, C. (2004) ‘Connected’ Presence: The Emergence of a New Repertoire
for Managing Social Relationships in a Changing Communication Technoscape.
Environment and Planning D: Society and Space, 22. p. 135–156.
LIVINGSTONE, S. (2009) Children and the Internet. Cambridge (UK): Polity Press.
———. (2008) Taking Risky Opportunities in Youthful Content Creation: Teenag-
ers’ Use of Social Networking Sites for Intimacy, Privacy and Self-expression.
New Media and Society, 10(3). p. 393–411.
MEAD, G. H. (1934) Mind, Self, and Society from the Standpoint of a Social Behav-
iorist. C. W. Morris (Ed.). Chicago: University of Chicago Press.
MEDIA AWARENESS NETWORK. (2001) Young Canadians in a Wired World:
The Students’ View. Ottawa: Media Awareness Network. http://mediasmarts.ca/
research-policy. Accessed 22–08–2013.
MENDOZA, K. (2007) “WATZ UR NAM?”: Adolescent Girls, Chat Rooms, and Inter-
personal Authenticity. Media Education Lab. www.mediaeducationlab. Accessed
22–08–2013.
OLIVEIRA, M. (2012, Feb 29). Canada’s ‘Most Socially Networked’ Title Slipping
Away. Globe and Mail. www.theglobeandmail.com/technology/digital-culture/
social-web/canadas-most-socially-networked-title-slipping-away/article550205/.
Accessed 29–02–2012
PETER, J., VALKENBURG, P. & FLUCKIGER, C. (2009) Adolescents and Social
Network Sites: Identity, Friendships and Privacy. In Livingstone, S. & Haddon, L.
(Eds.), <Kids Online>: Opportunities and Risks for Children. Bristol: The Policy
Press, 83–94.
PHILLIPS, D. (2009) Ubiquitous Computing, Spatiaility, and the Construction of
Identity: Directions for Policy Reponses. In Kerr, I., Steeves, V. & Lucock, C.
 Swimming in the Fishbowl 139
(Eds.), Lessons from the Identity Trail: Anonymity, Privacy and Identity in a
Networked Society. New York: Oxford University Press, 303–318.
ROONEY, T. (2010) Trusting Children: How Do Surveillance Technologies Alter a
Child’s Experience of Trust, Risk and Responsibility? Surveillance and Society,
7(4). p. 344–355.
SHADE, L. R. (2008) Internet Social Networking in Young Women’s Everyday Lives:
Some Insights from Focus Groups. Our Schools, Our Selves. p. 65–73.
———. (2011) Surveilling the Girl via the Third and Networked Screen. In Kearney,
M. C. (Ed.), Mediated Girlhoods: New Explorations of Girls’ Media Culture.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

New York: Peter Lang, 261–276.

STATTIN, H. & KERR, M. (2002) Parental Monitoring: A Reinterpretation. Child
Development, 71(4). p. 1072–1085.
STEEVES, V. (2005) Young Canadians in a Wired World—Phase II Trends and
Recommendations. Ottawa: Media Awareness Network. http://mediasmarts.ca/
research-policy. Accessed 19–08–2013.
———. (2012a) Young Canadians in a Wired World, Phase III: Talking to Youth
and Parents about Life Online. Ottawa: Media Smarts. http://mediasmarts.ca/
research-policy. Accessed 19–08–2013.
———. (2012b) Young Canadians in a Wired World—Phase III Teachers’ Perspec-
tives. Ottawa: Media Smarts. http://mediasmarts.ca/research-policy. Accessed
19–08–2013.
———. (2014) Young Canadians in a Wired World, Phase III: Student Survey. Ottawa:
Media Smarts. http://mediasmarts.ca/research-policy. Accessed 19–08–2013.
STERN, S. R. (2004) Expressions of Identity Online: Prominent Features and Gender
Differences in Adolescents’ World Wide Web Pages. Journal of Broadcasting &
Electronic Media, 48(2). p. 218–243.
VANDONINCK, S., D’HAENENS, L. & ROE, K. (2013) Online Risks. Journal of
Children and Media, 7(1). p. 60–78.
 7 Makers of Media Wisdom
Translating and Guarding Media
Wisdom in the Netherlands
Isolde Sprenkels and Sally Wyatt
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

Children in the Netherlands and elsewhere spend ever more time online
using various new media devices and applications to help them with their
homework, to play games, and to communicate with friends and family.
As Van der Ploeg and Pridmore discuss in the Introduction to this volume,
digitization and automation processes pervade nearly all social, political,
cultural, and economic domains in advanced industrialized societies, affect-
ing people in their various roles as citizens, consumers, and workers. Chil-
dren, as future workers, proto-consumers, and citizens-in-the-making, are
a particular focus of attention. On-going public debates about children’s
use of new media can be divided between those who draw upon discourses
of ‘celebration’ and those who draw upon discourses of ‘concern’ (Drotner,
2011). Kirsten Drotner explains that in the celebratory discourse, children
are perceived as media savvy, as pioneering experts, as digital natives, while
discourses of concern focus on the content-, contact-, and conduct-related
dangers possibly posed by new media (ibid.; see also Livingstone, 2009,
and Hasebrink et al., 2008). This distinction opposes an optimistic image
of children and their relationship with new media technology with a more
pessimistic and fearful image of the role new media technologies play in
children’s lives. Both views contain a dualistic picture of children as being
either media savvy or vulnerable (Buckingham, 2011) and a dualistic view
of new media technology posing either opportunities or risks (Livingstone,
2009). Either way, whether describing savvy or vulnerable children and
portraying technologies as offering opportunities or risks, it is increasingly
seen as necessary to educate children about the possibilities of new media.
In these contexts, new media is either seen as a medium for creative self-
expression, public participation, and cultural heritage or as a means of
encountering harmful or offensive content and people (Lunt and Living-
stone, 2012). This suggests another dualism, one that includes advocates
for empowerment versus those with a more protectionist take on media
education (Hobbs, 2011).
 Makers of Media Wisdom 141
While the need for education in itself may be undisputed, these opposing
positions make it less clear which approach should be used and how ‘respon-
sibility’ should be distributed amongst all those involved, including par-
ents, teachers, and librarians as well as those more removed from children’s
daily lives, such as policy makers, private corporations, and the various
non-governmental organizations (NGOs) developing digital material aimed
at children. There are also many terms circulating that attempt to capture
the knowledge and skills assisting people to use, interpret, and produce
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

media. These terms are a part of a range of educational activities. Terms in

circulation include not only ‘media literacy’ (European Commission), but
also ‘media and information literacy’ (UNESCO), ‘digital literacy,’ ‘infor-
mation literacy,’ and the concept of ‘media wisdom.’ Dutch policy makers
now favour this latter term. However, these terms are not neutral. As Sonia
Livingstone says, “how media literacy is defined has consequences for the
framing of the debate, the research agenda and policy initiatives” (2004, 3).
In this chapter, we trace how ‘media wisdom’ took root in the Dutch
policy context.1 Our analysis draws upon insights from the interdisciplinary
field of science and technology studies (STS), moving beyond technologi-
cal determinist views of technology (Wyatt, 2008a), and essentialist views
of either technology or of users’ identities (Oudshoorn and Pinch, 2005;
Wyatt, 2008b). We prefer to emphasize that people—adults and children
alike—engage with new media technologies in many different ways. With-
out denying the enormous power of contemporary digital technologies, we
also recognize their contingency, flexibility, and multiplicity of everyday
(Internet) use and users (Wyatt, 2005). In particular, we make use of the
concept of ‘boundary object’ in order to understand how terms like ‘media
literacy’ and ‘media wisdom’ are able to engage and mobilize social actors
across disparate areas. We suggest that the concept of media wisdom func-
tions as a boundary object (Star and Griesemer, 1989; Star, 2010) in the way
it bridges boundaries between different sectors and the organizations and
representatives operating within them. They all have different viewpoints
on media wisdom deriving from the sectors in which they operate and their
concomitant priorities and target groups. As a concept, ‘media wisdom’ is
simultaneously flexible and robust, adapting to different viewpoints and
retaining a common identity across diverse sectors, organizations, and their
representatives. As we will show later in this chapter, the concept of ‘media
wisdom’ meant different things to the Dutch Council for Culture who intro-
duced the term than to the Dutch Cabinet adopting it a few years later. But it
is also translated many more times and in many more ways. As Susan Leigh
Star and James Griesemer suggest, boundary objects are “weakly structured
in common use, and become strongly structured in individual site use . . .
they have different meanings in different social worlds but their structure
is common enough to more than one world to make them recognizable, a
means of translation” (1989, 393).
 142 Isolde Sprenkels and Sally Wyatt
In the next section, we provide some of the pre-history, outlining earlier
European and international definitions of the various terms mentioned in
the preceding paragraph. The following section outlines how ‘media wis-
dom’ emerged, starting in an advisory report to the government in 2005
and becoming more official in a Cabinet report in 2008. Next, we focus on
how this concept was translated in practice, leading to a proliferation of
initiatives across sectors, all claiming to be implementing ‘media wisdom.’
The following section explains how this initial flowering was pruned back
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

or became operationalized in a more narrow interpretation of the term, con-

tributing to the institutionalization of media wisdom into education. We
observe that media wisdom is sought as a quality to be developed in new
media users themselves, specifically in children, but such a view fails to take
into account that users and technology are constituted in relation to one
another, which means media wisdom is also facilitated or hindered by and
reliant on the design of new media. In the Conclusion, we reflect on what
this process has taught us about how concepts and definitions serve to enrol
different types of actors and distribute responsibility across all of them.

BACKGROUND: MEDIA EDUCATION AND LITERACY

Media education is generally considered to provide the means of implement-

ing the aspirations of media literacy (Lunt and Livingstone, 2012). Both
have been studied, debated, and defined in various domains and disciplines
for decades. Although debate continues on how best to define media lit-
eracy and what it means for educational practices, there seems to be some
agreement on a conceptual level. An oft-quoted definition formulated during
the National Leadership Conference on Media Literacy held in the United
States in 1992 defined media literacy as “the ability to access, analyse, evalu-
ate, and communicate messages in a variety of forms” (Aufderheide, 1993).
Most definitions within academic and policy discourse contain similar ele-
ments, relating to people’s knowledge and skills, aiding them to use, inter-
pret, and produce media.
There are two prominent approaches within debates about media educa-
tion. The first, a protectionist approach, advocates “an intervention designed
to counter the negative effects of mass media and popular culture” (Hobbs,
2011, 423), by some described as a “pedagogic equivalent of a tetanus shot”
(Bazalgette, 1997, 72). In other words, the approach emphasizes inoculating
children against the ‘damaging effects’ of media by lecturing about media’s
negative influences, manipulated messages, and so on. It is thinking and
acting for children, who are generally perceived as passive and vulnerable
media consumers. The second approach is supported by empowerment
advocates, which aims to teach children to think and act for themselves,
generally perceiving them as competent social actors. These empowerment
advocates take account of what children already know about media instead
 Makers of Media Wisdom 143
of laying out what they ought to know (Buckingham, 2011). Not only do
they strongly promote a critical analysis of media, but they also recommend
teaching “children to use the technical tools of self-expression, all the better
to participate in modern society” (Lunt and Livingstone, 2012, 118), as the
media afford “an expressive, cultural, and participatory opportunity which
brings significant benefits to those who are able to ‘read’ its codes and con-
ventions and to use its tools and technologies” (ibid., 119). We argue that
in the approach of empowerment advocates, both educators as well as new
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

media technology can be considered to be ‘empowering.’ By this we mean

that it is about empowering youth, teaching them to think critically for
themselves regarding the opportunities and risks new media possibly pose,
as well as the empowering affordances of new media technologies. Within
this empowerment approach, children sometimes are discursively positioned
as pioneers, digital natives leading the way into the future (Helsper and
Enyon, 2010; Lunt and Livingstone, 2012), considering them as naturally
autonomous and competent—as spontaneously ‘media literate’ (Bucking-
ham, 2011). David Buckingham warns that adopting such a view can pro-
vide a useful alibi for avoiding the obligation to care for children in this
matter (ibid.).
Media literacy has long received attention on international policy agen-
das. On a global level, UNESCO (United Nations Educational Scientific and
Cultural Organization) stresses the need to provide citizens with compe-
tences (knowledge, skills, and attitudes) necessary to engage with traditional
media and new technologies.2 It has published several documents related
to the topic (UNESCO, 1982, 2005, 2007) using the terms ‘media edu-
cation’ and ‘information literacy.’ Since 2011, UNESCO uses ‘media and
information literacy’ (MIL) as a composite concept (Wilson et al., 2011).
“MIL brings together Information Literacy and Media Literacy, along with
Information and Communication Technology, and Digital Literacy, as a new
literary construct that helps empower people, communities and nations to
participate in and contribute to global knowledge societies” (UNESCO,
2013, 17). Later in the same document, it explains the benefits of MIL as
it “empowers citizens to access, retrieve, understand, evaluate and use, to
create as well as share information and media content in all formats, using
various tools, in a critical, ethical and effective way, in order to participate
and engage in personal, professional and societal activities” (UNESCO,
2013, 29). Elsewhere, it states that MIL can be considered to be a way to
“expand civic education movement that incorporates teachers as principal
agents of change” (Wilson et al., 2011, 11). Its strategy includes initiatives
such as a curriculum for teachers (Wilson et al., 2011), guidelines for poli-
cies and strategies (Grizzle et al., 2013), and setting up a university network
to stimulate research that can inform policies.3
On a European Union level, ‘media literacy’ is not only considered to be
important for active citizenship, participation, inclusion, and social cohesion,
but it is also considered to be crucial for the establishment of a competitive
 144 Isolde Sprenkels and Sally Wyatt
knowledge-based economy, in which digital media and the Internet in particular
are emphasized (European Commission, 2007, 2009; European Parliament,
2008). ‘Media literacy’ entered the European policy stage when it came to be
considered essential in meeting the goals of the 2000 Lisbon Strategy (Euro-
pean Commission 2007, 2009; European Parliament, 2008), and its con-
nection with economic interests became even more important in the Europe
2020 Strategy for Smart, Sustainable, and Inclusive Growth (European
Commission, 2010a) and one of its initiatives, A Digital Agenda for Europe
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

(European Commission, 2010b). Enhancing ‘media literacy’ or ‘digital lit-

eracy’ as the Commission chooses to use in the Digital Agenda, is deemed
necessary, as “Europe is suffering from a growing professional ICT skills
shortage and a digital-literacy deficit. These failings are excluding many citi-
zens from the digital society and economy and are holding back the large
multiplier effect of ICT-take up to productivity growth” (European Commis-
sion, 2010b, 6). The EU definition of media literacy focuses on “the ability
to access the media, to understand and critically evaluate different aspects
of the media and media content and to create communications in a variety
of contexts” (European Commission, 2007, 2009). In order to develop this
definition, a public consultation among industry, media organizations, edu-
cation institutions, regulators, and citizen and consumer organizations was
held, and a Media Literacy Expert Group meeting was organized in which
academics and media professionals contributed.
Apart from describing a duality in media literacy education, the preceding
paragraphs briefly illustrate two points. First, the understandings of ‘media
literacy’ differ between key actors, for instance between UNESCO and the
EU. UNESCO focuses on civic education and participation, and the EU
highlights employability and economic growth. In this chapter, we demon-
strate that this also happens on a more local level. Second, various actors
adopt different terms such as ‘media literacy,’ ‘media education,’ ‘digital
literacy,’ and ‘information literacy,’ which carry different meanings and con-
notations. This influences the foci of the ‘media literacy’ policy agendas and
practices.
In this chapter, we refrain from assigning a priori definitions to ‘media
wisdom’ or ‘media literacy.’ Instead, we look at how they were created
and put on the policy agendas as solutions to deal with a variety of poten-
tial problems such as civic, social, and economic disadvantage and exclu-
sion, which, in turn, may occur due to developments in media technology,
and processes of digitization and globalization. Therefore, we argue, it is
important to explore and follow the concepts of ‘media literacy’ and ‘media
wisdom’ while keeping in mind that they are an outcome of relations, of
choices, practices, an interplay between negotiating actors, and not some-
thing stable with a fixed meaning across time and place. Inspired by Bruno
Latour’s first rule of method, that we should study science in action and
not ready-made science or technology (Latour, 1987), we trace ‘media wis-
dom’ ‘in the making’, in action as it emerged in Dutch policy and practice.
 Makers of Media Wisdom 145
We follow it and trace the ways in which the various meanings and con-
notations shape the ‘media wisdom’ debate and various ‘media wisdom’
practices, enrolling different types of social actors and distributing respon-
sibility across them.

FROM MEDIA EDUCATION TO MEDIA WISDOM

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

As described, in the early years of the 21st century, there were already several
terms in circulation, aimed at capturing and addressing the problems faced
by children in learning to use and in using advanced digital technologies.
Nonetheless, in its advisory report to the Dutch government in 2005 titled
Media Wisdom: The Development of New Citizenship (Raad voor Cultuur,
2005), the Dutch Council for Culture (Raad voor Cultuur: RvC) decided
not to use either ‘media education’ or ‘media literacy’ but to use ‘media
wisdom’ instead. According to the RvC, media education in the Netherlands
had been passive, defensive, and narrowly focused on youth and formal
education. It wanted to move away from such a protectionist approach. The
RvC claimed that Dutch society is saturated with media and that citizens
have to be(come) ‘media wise’ in order to be able to participate in various
domains, ranging from healthcare, leisure, and the labour market to poli-
tics, education, and commerce. It defined ‘media wisdom’ as: “The whole of
knowledge, skills, and mentality by which citizens can consciously, critically
and actively move within a complex, dynamic and fundamentally mediated
world” (Raad voor Cultuur, 2005, 2). It further stressed that the purpose
of ‘media wisdom’ is not just dealing with media as such but participating
in wider societal processes (Raad voor Cultuur, 2005). ‘Media wisdom,’ the
RvC argued, would provide a much broader perspective than media educa-
tion. The RvC noted that citizens were dealing with various technological
and societal developments, contributing to new forms and practices of citi-
zenship. It described a move from a welfare state to a society in which the
government increasingly calls upon citizens’ self-reliance in various areas for
which they need to use digital media. In light of such developments, citizens
need “tools and expertise to shape this responsibility in a satisfying way for
themselves and society” (Raad voor Cultuur, 2005, 3). They needed to be
able to participate in the societal process so that social exclusion and the
emergence of a ‘digital divide’ would be prevented, related to differences in
media use and skills. In other words, citizens needed to be ‘media wise’ in
order to fit this ‘new citizenship.’
The RvC stressed that the government has a duty of care for enhancing
such media wisdom in citizens. It did not propose to make media wisdom
obligatory in formal education, but it did suggest that it should be consid-
ered part of a school’s citizenship education and of the curricula for training
teachers. It also recommended that schools could hire media coaches to
promote the coherence of media wisdom within educational programmes, to
 146 Isolde Sprenkels and Sally Wyatt
guide other staff members, and to initiate media projects. The RvC empha-
sized that media wisdom should not only be part of formal education but
also of the tasks and practices of public media, cultural organizations, and
public libraries. Prior to this point, initiatives had been made within all these
sectors, but there was no shared policy.
The RvC considered it “of great importance to concentrate on enlarging
media competences, or rather: media wisdom of citizens” (Raad voor Cul-
tuur, 2005, 13), emphasizing the competences and skills needed for media
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

wisdom. The definition provided by the RvC makes this clear: “media wis-
dom is the whole of knowledge, skills and mentality by which citizens can
consciously, critically and actively move within a complex, variable and
fundamentally mediated world” (Raad voor Cultuur, 2005, 18). Knowl-
edge includes that which is “needed to interpret media messages, the aware-
ness that media content is constructed, the ability to detect which interests
or value systems are steering them” (Raad voor Cultuur, 2005, 19). Skills
include “the ability to watch, being able to choose, and to control buttons”
(Raad voor Cultuur, 2005, 20), so that one could not only consume but also
produce media content. Mentality has to do with “citizens’ awareness of the
way they use media, and the effect that has on themselves and others” (Raad
voor Cultuur, 2005, 2).
Most media literacy definitions contain similar items relating to people’s
knowledge and skills, helping them use, interpret, and produce media.
Mentality, on the contrary, is a more specifically Dutch element. The most
striking difference in this RvC document is the use of ‘wisdom,’ wijsheid in
Dutch, instead of the more widespread geletterdheid, the usual translation
of ‘literacy.’ The RvC justifies its choice in a single sentence, which reads,
“the more commonly used media geletterdheid is more suitable, but not
preferred by the RvC due to its linguistic associations” (Raad voor Cultuur,
2005, 18), arising from the analogy with print literacy, namely the ability
to read, write, or produce text. The RvC deliberately chose the term ‘wis-
dom,’ which it felt was broader than literacy. ‘Wisdom’ carries connotations
with various philosophical and religious perspectives dating back to ancient
classical and biblical times, namely “the ability to use your knowledge and
experience to make good decisions and judgements” (Cambridge Dictionar-
ies online). This matches the perspective of the RvC on media wisdom as
an on-going process that citizens continue to develop, in contrast to a more
binary take on literacy in which one can or cannot read and write at a basic
level. Literacy is taught in school from an early age, provided for by the gov-
ernment, through the right to education (United Nations, 1948). This is not
the case for media wisdom, which, we argue, carries a different connotation
in relation to responsibility. Through the education system, the government
is primarily responsible for teaching literacy but not for developing lifelong
media wisdom. As will become clear in the following pages, media wisdom
has not been made compulsory in education. This worries Terry, editor-
in-chief of one of the first Dutch websites on parenting and editor-in-chief
 Makers of Media Wisdom 147
at a knowledge centre for youth and (digital) media, focusing on schools,
parents, and the media industry. She claims:

So now we have a situation that at the Ministry of Education it is said

“yes it is important that citizens are made media wise and that it hap-
pens in school,” but that’s it. We are the only country in Europe that
does not have the obligation to include media wisdom in its curriculum.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Terry worries about this. She compares media wisdom education with
teaching children how to read and write, and raises concerns about the emer-
gence of an illiterate citizenry and all the concomitant consequences:

I don’t think we would accept it if children finished primary educa-

tion from one school knowing how to read and write, and finished at
another school not knowing how . . . In fact, this is the new illiteracy
which we are creating, what we allow to happen, people have no idea.

Even though the RvC chose ‘wisdom’ and not the Dutch equivalent gelet-
terdheid, Terry associates media wisdom with print literacy on the level of
knowledge and power differences, connecting it with societal disadvantage
and exclusion.
The Dutch government did follow up on some of the recommendations
of the RvC advisory report. In a response to the RvC’s advice, the Minister
of Education, Culture and Science acknowledged the importance of media
wisdom, arguing that a mediated society leads to different relationships
between media and citizens, government and citizens, and citizens and busi-
ness (Hoeven, 2006). The minister advocated cooperation and knowledge
exchange within networks and organizations already dealing with media wis-
dom in one way or another. After the report, numerous initiatives concerning
media wisdom were undertaken, including the development of courses and
teaching material and the organization of workshops. However, the minister
criticized this proliferation of initiatives for its lack of coherence. To address
this, the ministry organized a conference together with the RvC in 2006 in
order to search for “the meaning and impact of the media wisdom concept
which is difficult to grasp” (Ministerie OCW, 2006, 1). It became clear, per-
haps for the first time, that there were some difficulties with the concept, that
it carried many connotations, meanings, interpretations, and practices. The
conference “had people from various sectors and disciplines discuss the topic
at different levels of abstraction in contact with each other” (ibid.).
The Dutch Cabinet published its vision on media wisdom, called Medi-
awijsheid, three years after the RvC’s report (Plasterk and Routvoet, 2008).
It cites the RvC’s broad definition in one of its footnotes, but the notion of
‘media wisdom’ had changed. In the introduction, the cabinet states that
it wants to foster safe and responsible media use by equipping citizens to
take advantage of the opportunities new media are said to offer and to deal
 148 Isolde Sprenkels and Sally Wyatt
with the risks they are assumed to pose and to stimulate a safe media envi-
ronment, using self-regulation and a system of complaints for parents and
other carers. Immediately following this, it states that the focus of media
wisdom is no longer on all citizens but on youth, their social environment,
their media use, and positive and negative media effects. The target group for
media wisdom initiatives was narrowed down, and the goals had changed. In
addressing media wisdom, the cabinet took up the RvC’s concept but trans-
lated it in such a manner that the meaning of the concept changed, as did
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

the target group, government policy, and implicated actors. In other words,
a new configuration emerged of practices, policies, social actors in which
media wisdom is enacted into being. Youth and their wider social environ-
ment including parents, teachers, and other carers became top priority. Media
wisdom became connected to media effects and the ways in which parents,
other carers, and media producers play a part in guiding children’s media
use. In a way these actors were turned into guardians of media wisdom. The
government itself also carried out a protectionist-oriented policy, for instance
through its involvement in public broadcasting, providing support for media
self-regulation and rules and regulations to protect (young) media users.4 It
also stimulated research on the topic and supported media users in critical
and conscious media use. The latter would become a task of the media wis-
dom expertise centre the cabinet suggested starting, which had as a goal to
bring together various media wisdom initiatives mainly focusing on support-
ing youth, parents and other carers. According to the cabinet, the mission of
the media wisdom expertise centre was to strengthen media wisdom within
society through first focusing on youth, parents, carers, and teachers in order
to empower youth, to help them use media wisely and actively.
In this section, we have outlined how the concept of ‘media wisdom’
emerged in the Dutch policy landscape, how it changed from being a broad
concept capturing a lifelong process of learning to something aimed primar-
ily at empowering young people by those responsible for their development
and education, as well as protecting them by particular provisions and regu-
lations. In the next section, we turn to the many organizations that took
up the concept in myriad ways, focussing on its empowering connotations,
giving examples from different sectors.

MAKERS OF MEDIA WISDOM

As mentioned in the previous section, one of the initiatives of the Dutch gov-
ernment was to establish a media wisdom expertise centre bringing together
the variety of existing initiatives related to media wisdom. Funded by the
Ministry of Education, Culture and Science, this networking organization,
named Mediawijzer.net, Media Wiser in English, started in 2008 in order to
strengthen media wisdom in society, with a primary focus on youth between
2 and 18 years of age and their carers and teachers.
 Makers of Media Wisdom 149
Mason, a colleague of Terry, and director of the knowledge centre on
youth and (digital) media, mentions that “Mediawijzer.net has been initiated
to prevent fragmentation. I have the idea that fragmentation has grown.”
He connects this fragmentation with not making media wisdom obligatory
within the formal education system:

Mediawijzer.net is not to blame for that, they do a terrific job, they also
connected many people together, strengthening the network. A regula-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tion to stimulate projects has led to many beautiful initiatives. But the
theme is not being concretized by the government, it is not a compulsory
section of education, so the market can do with it as it likes.

Mediawijzer.net relates to the intention to make citizens more media

wise using these initiatives. It considers itself to be “an active and diverse
network with one shared interest: making The Netherlands media wise”
(Mediawijzer.net, 2013a). Mediawijzer.net describes itself as “[a] network
of makers of media wisdom” (ibid.). As will be seen in this section, the
fragmentation of Mediawijzer.net and the broad variety and struggle for
conceptual unity and understanding also “makes media wisdom.” In other
words, media wisdom is continuously shaped in a variety of ways, eventu-
ally moving towards standardizing and institutionalizing it.
Mediawijzer.net connects a variety of organizations concerned with media
wisdom, promoting cooperation and synergy between them (Kwartiermak-
ersgroep, 2008). Its mission is formulated as follows: “The expertise centre
Mediawijzer.net connects, strengthens, and inspires initiatives and organiza-
tions around media wisdom” (Mediawijzer.net, 2010a). It functions as a
platform for knowledge sharing, conducting research, organizing events and
campaigns, and advising local and central government.
Mediawijzer.net has a programme council. Five organizations form the
core of the centre: the Netherlands Institute for Sound and Vision (NIBG)
preserving audio-visual heritage, Kennisnet, a public expertise centre for
ICT in education, the Institute for Public Libraries Sector (SIOB), the pub-
lic broadcasting company (NTR), and the Electronic Commerce—or infor-
mation society—Platform (ECP-EPN). Each has a specific role. NIBG is the
national meeting point or ‘clubhouse’ and responsible for communication.
Networking meetings are held there, and visitors of the NIBG can test
their media wisdom skills in a specially equipped pavilion. SIOB provides
local counters or ‘houses of media wisdom’ for citizens and is formally
responsible for libraries. NTR is responsible for communication in gen-
eral and for broadcasters and media, and ECP-EPN coordinates research.
Kennisnet is responsible for online environments and for the website of
Mediwijzer.net, which disseminates news and information on media wis-
dom and provides information about the activities and services of the net-
work partners, or “makers of media wisdom” (Mediawijzer.net, 2013a).
The five organizations represent different sectors, including culture, media,
 150 Isolde Sprenkels and Sally Wyatt
education, and business, each of which already has a strong knowledge
base for media wisdom. Media wisdom means something slightly different
in each of these sectors and the organizations and representatives operat-
ing within them. To SIOB and its representatives, media wisdom might be
related to developing information (-seeking) skills, while NIBG and NTR
focus on the manipulation of images, and Kennisnet might support initia-
tives concerning digital bullying or setting up rules concerning smartphone
use in schools.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Since its start in 2008, more than 900 network partners (Mediawijzer.net,
2014) are part of Mediawijzer.net, and this keeps increasing. “This network
should form a representative reflection of the breadth of the media wisdom
concept in combination with its target groups” (Mediawijzer.net, 2010b,
28). Anyone can take part, and the variety of organizations is enormous:
primary, secondary, and tertiary education, research organizations, indepen-
dent researchers, companies engaged in consultancy and coaching in edu-
cation and IT, charitable organizations, media companies, communication
bureaus, libraries, cultural organizations including museums, and drama
and art workshop organizers, youth care, youth work. The number of initia-
tives are manifold, yet their link with media wisdom is sometimes tenuous.
For example, there is a reference to ‘autism training’ in the list. Following
the link, it turns out to be a ‘mind academy,’ offering ‘neuro-linguistic pro-
gramming training.’ Terry comments on this pluralism:

It is meant to be diverse, but now it is so diverse that you don’t know,

actually all different possible takes on media wisdom are represented,
no choice is made, no structure, no selection . . . Anyone can become
a member now, every month a list of twenty new partners join. Even-
tually it is about reading and writing, so after a while everybody has
joined.

We argue that in this way, media wisdom functions as a boundary object

(Star and Griesemer, 1989; Star, 2010). It bridges boundaries, not only
among the five organizations that form the core of the centre and the sectors
they represent but also among the 900 networking partners, the RvC, the
Dutch Cabinet, and many other actors, such as the news media reporting on
media wisdom–related topics. The term ‘media wisdom’ retains its identity
among all of these organizations and their representatives, being recogniz-
able to all. At the same time, it is adaptable in such a way that everyone
involved can fill in their own meanings relating to their own practices, pri-
orities, and target groups. A continuous movement back and forth between
the two forms of the boundary object, between the weakly structured and
the well-structured aspects of media wisdom, takes place (Star, 2010). As
Star explains, over time, people try to control this movement, which will be
demonstrated in the next section.
 Makers of Media Wisdom 151
MOVING FROM MEDIA WISDOM TOWARDS
MEDIA EDUCATION

According to Mediawijzer.net, the abstract way in which the RvC defined

media wisdom facilitated the proliferation of initiatives and meanings put
forward by different social actors (Madiawijzer.net, 2011a). By providing a
financial incentive, Mediawijzer.net invited its networking partners to pro-
vide clarity on the concept of media wisdom, to establish its elements and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

the measurable competences a person needs to be media wise (Wiebenga

et al., 2011).
Meten van mediawijsheid (Measuring Media Wisdom) is the report writ-
ten by six of the partners which received funding from Mediawijzer.net. They
used the RvC definition of media wisdom as a starting point and promised
to ‘operationalize’ it in order to support parties that would subsequently
measure it. They formulated a definition of media wisdom, explicitly in terms
of competences, considering it to be “a collection of competences a person
needs to master in order to be media wise” (Wiebenga et al., 2011, 28). They
translated the broad RvC definition of media wisdom into a model contain-
ing 14 competences. A person who masters all these competences is consid-
ered to be media wise, offering “an essential vehicle for self-reliant citizenship
and full participation in the media society” (Mediawijzer.net, 2011a, 3).
Mediawijzer.net argues that a model offers the possibility to order the pro-
liferation of media wisdom initiatives in a new way. “With projects, work-
shops, websites and educational material it is now possible to determine to
which competences they contribute” (Mediawijzer.net, 2011a, 8). We argue
that it not only aids in ordering, but it also serves to standardize the media
wisdom concept and its practices by focusing on particular competences. The
definition of media wisdom provided by Mediawijzer.net from that moment
on is “the set of competences which everyone needs, to participate actively
and consciously to the media society” (Mediawijzer.net, 2013b).
In 2012, a slightly different competence model was developed, consisting
of 10 competences instead of 14, divided into four groups: use, understand-
ing, communication, and strategy. The 10 competences are: awareness of
the growing influence of media on society, understanding how media are
made, understanding how the media colour reality, using equipment soft-
ware and applications, orientation within media environments, finding and
processing information, creating content, participating in social networks,
reflecting on own media usage, and achieving objectives through media.
All 10 competences can be achieved on a level from zero to four and cover
the three elements as provided for by the RvC, namely knowledge, skills,
and attitude. “The model forms an invitation to teachers and professionals
(media coaches, librarians, researchers) to apply it in their daily practice”
(Gillebaard et al., 2013). “It forms a handle in developing media wisdom
products and services, making the offer of media wisdom initiatives more
 152 Isolde Sprenkels and Sally Wyatt
insightful, and helps shaping (simple) measurement instruments“ (website
Mediawijzer5). It should be used as “[a] starting point for media wisdom
initiatives” (Mediawijzer.net, 2012).
The point here is not the exact details of the 10 competences but rather
the overall effort to create a model with a set of measurable standards and
indicators. According to boundary object theory, “over time, people (often
administrators or regulatory agencies) try to control the tacking back-and-
forth, and especially, to standardize and make equivalent the ill-structured
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and well-structured aspects of the particular boundary object” (Star, 2010,

613). Star mentions that people try to control this constant moving between
the ill-structured and the well-structured form of the boundary object and
that at a certain point, the movement between the two forms scales up or
becomes standardized. From that point, boundary objects begin to move
and change into infrastructure, into (methodological) standards. We have
described a ‘tacking back and forth’ between the ill- and well-structured
forms of media wisdom. With the development of a competence model, a
methodological standard was developed, changing media wisdom from a
loose, broad set of definitions into a standard for all network partners to use
when developing their products, services, and activities. All network part-
ners need to connect with this model and these competences and comply with
their further development in order for projects, websites, workshops, and
educational material to be transparent about the competences to which they
can contribute and develop (Mediawijzer.net, 2011a). Here Mediawijzer.net
establishes itself as an obligatory passage point (Callon, 1986),6 strength-
ening its position by standardizing media wisdom and enabling its opera-
tionalization. In the remainder of this section, we demonstrate how this
contributes to the institutionalization of media wisdom in schools.
Mediawijzer.net stimulates the integration of media wisdom in educa-
tional curricula, either as a separate course or integrated in existing courses
(Mediawijzer.net, 2010b). It explains that this is difficult, as it is not consid-
ered a compulsory subject in formal education, but that a way into education
is through influencing educational publishers, schools, and teachers. And as
described earlier, a way into the schools is through the network partners who
are stimulated to use the competence model in developing their initiatives,
which in turn may be deployed by individual schools for their media projects.
Mediawijzer.net does not develop services for media users but stimulates its
network partners to do so. For instance, it provides funding to its network
partners to develop media wisdom projects which contribute to the stimula-
tion of media wisdom through an annual incentive round providing funding
for agreed themes. Each year several projects receive funding based on an
annual topic, goal, problem, or set of spearheads. Mediawijzer.net explains
that projects in this first period were mainly focusing on an “operational
level: practical guides, lessons, policies, and descriptions of media wisdom”
(Mediawijzer.net, 2011b). From 2012, the themes of the incentives focus on
more strategic support, integrating media wisdom into education.
 Makers of Media Wisdom 153
The 2013 incentive requests the network to develop solutions for three top-
ics (Mediawijzer.net, 2013b). The first request is to develop a ‘learning tool
bank media wisdom’7 for students in primary education to be used by
teachers. The second is a learning tool bank for people undergoing teacher
training. The third is a learning tool bank for teachers to keep them up to
date. There is already media wisdom material for primary education, but
according to Mediawijzer.net, it is hard to find and difficult to use, because
it is not clear what media wisdom competences are developed, whether they
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

relate to the school’s curriculum or to the core goals for primary education
established by the Dutch government, whether it matches students’ needs
at a particular age, and there are no exams (Mediawijzer.net, 2013b). In
teacher training, not much formal attention is yet paid to media wisdom;
rather, those learning to be teachers need to develop their own skills and
learn how to teach media wisdom to children (ibid.). Finally, teachers in
primary education want to be kept up to date and continuously develop
their competences and skills. According to Mediawijzer.net, the learning tool
bank could offer this to them (ibid.). Mediawijzer.net has also developed
a step-plan to integrate or “embed media wisdom structurally in primary
education.”8 It offers its network partners such as libraries, museums, pub-
lishers, cultural organizations, and companies an approach to make their
products and services visible for primary education and suggests possible
activities to reach target groups, such as teachers, school boards, and the
school environment including parents and libraries. Most of its advice con-
nects with Mediawijzer.net initiatives, for instance participating in the incen-
tive funding and using the competence model in developing products.9
As the foregoing indicates, media wisdom is treated as a matter of
education—developing competences. Mediawijzer.net attempts to have this
incorporated into formal education. This time an empowering approach is
followed, contrary to the previous protectionist approach from which the
Dutch Council for Culture advised its government to distance itself. Sonia
Livingstone argues that even if media literacy would be taught in schools,
there would still be difficulties in providing media literacy for all. She starts
by sketching the difficulties relating to continuous technological change and
a need for lifelong learning. Contrary to print literacy, education at a young
age does not last a lifetime; therefore, she also discusses ways to reach adults
and some of the difficulties in doing so. But the biggest challenge she dis-
tinguishes in establishing media literacy for all is not a practical one; it is
“a fundamental point of definition, or to be precise, of mutual definition.
Media design and media literacy can be considered as reverse sides of the
same coin” (2011, 33). She argues that “difficulties experienced by citizens
are being framed as media literacy tasks when an alternative description
would point to failures of design, provision or regulation” (ibid.). She argues
that “the task of acquiring media literacy is unnecessarily burdened by the
insufficient design, planning and provision of media services and providers”
(2011, 32).
 154 Isolde Sprenkels and Sally Wyatt
In support of this argument regarding failures of provision or regulation,
we claim that a strong focus on media wisdom in children, empowering or
enabling children in a variety of ways, teaching them to think and act for
themselves risks losing sight on protectionist measures, thinking and act-
ing for children, by for instance regulating media and children’s media use
through legal, social, and technical measures. For a long time, there has been
a debate in the Netherlands among media wisdom practitioners whether
children should be protected or empowered (‘weerbaar maken’). Justine Par-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

doen, affiliated with one of the first Dutch websites on parenting as well as
a knowledge centre for youth and (digital) media, said in an online column:

I have changed my mind: as caretakers it is our duty to protect children

from being exposed to harmful material. And we cannot do that by solely
empowering them. We can quarrel about what exactly is harmful, but
protection has become very important, as well as proper guidance.10

Pardoen previously supported an empowerment perspective but changed

her mind due to developments such as the increase in domestic Internet
access, as well as an increase in the amount of extreme violence, harassment,
and pornography available online. She now supports technological means
for regulating children’s Internet use, for example, by installing web filter
software. She argues that choosing between protection and empowerment
is not necessary, as one can do both. We would argue that the one implies
the other, which Pardoen implicitly recognizes when she states that we can’t
protect children solely by empowering them, which means empowerment
can also be considered as a form of protection.
In addition, we argue that a tension arising with media wisdom comes
from its strong reliance on the individual skills and competences of a gener-
alized media user. Insights from the interdisciplinary field of science and tech-
nology studies (STS) help us understand the limits of relying on individual
skills and competences. Not only are the multiplicity of uses or situated use
(Bakardjieva, 2005; Wyatt, 2005; Wyatt et al., 2005) and the multiplicity of
users not adequately considered (Wyatt, 2005), nor are the character of new
media technology itself and the economic context in which it is produced.
Users and technology are constituted in relation to one another; focus-
ing on one cannot be done without also considering the other. As Leah
Lievrouw claims, “technologies and people alike would be thought of as
interrelated nodes in constantly changing sociotechnical networks, which
constitute the forms and uses of technology differently in different times
and places for different groups” (2006, 250). Similarly, Bakardjieva and
Gaden explain that “technologies come to life in the hands of human oper-
ators, practitioners who are situated in a much wider and richer network
of relations and experiences than a single technology, no matter how total-
izing, is able to encompass” (Bakardjieva and Gaden, 2011, 410). In other
words, the everyday practice or use of technology is complex, and so is
 Makers of Media Wisdom 155
the variety of users. In media wisdom policy making and practices, new
media usage is set as a ‘norm.’ Those who do not use new media or do not
use it ‘wisely’ are believed to face risks such as socio-economic exclusion.
But users of new media technology are not a homogeneous group. For
instance, the focus of media wisdom policy makers is often on the ‘have
nots,’ or ‘illiterates’ and does not take into account the ‘want nots,’ who
may have good reasons for rejecting or resisting some new media forms
(Wyatt, 2005).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

In the documents we have analyzed, new media technology remains black

boxed, and is considered to be neutral, static and homogeneous. Together
with Livingstone (2009, 2011), we argue that media wisdom is also shaped
by new media, in the sense that it is facilitated or hindered by and reli-
ant on the design of new media. Livingstone argues that “media literacy
is also medium-dependent, a co-production of the interactive engagement
between technology and user” (2004, 12). She has shown, for example,
that the design of Internet environments, can actually shape and limit media
literacy (2009). ‘Failing’ to engage in such an environment is not just caused
by a lack of competences or skills on the part of a user but by the leg-
ibility of the medium. This means one should also consider media tech-
nologies themselves and the ways in which they allow a user to ‘read’ an
environment. In some cases, for instance with particular forms of marketing
communication such as advertising disguised in games, the ability to detect
the economic interests underlying such environments is prohibited by their
design (Sprenkels and Van der Ploeg, 2014). Livingstone refers to the work
of Steve Woolgar and the notion of ‘technology as text.’11 Woolgar’s use of
the metaphor of ‘machine as text’ is part of a semiotic approach to user-
technology relations. It was introduced by actor-network theory scholars
who have “extended semiotics—the study of how meanings are built—from
signs to things” (Oudshoorn and Pinch, 2005, 7). Or, as Peter Lunt and Sonia
Livingstone claim,

technologies are also text, and the institutional purposes and culture, orga-
nizational norms and structures, and communicative design and intent are
all embedded in the very construction of the interface, as is also an implicit
conception of the user—what they know, don’t know, can and cannot do,
take for granted or need to learn. The more impenetrable, opaque or ill-
designed the text or interface . . . the more users will struggle.
(Lunt and Livingstone, 2012, 135)

CONCLUSION: CONTEXTUALISING MEDIA WISDOM

In contemporary society, various technological and societal developments

contribute to calls for a media-literate citizenry. In this chapter, we have traced
how media wisdom as a concept, goal, practice and identity in the sense of
 156 Isolde Sprenkels and Sally Wyatt
‘being media wise’ took root in the Dutch policy context. We demonstrated
how it was translated by the Dutch Cabinet, narrowing down the target
group and changing the goals set by the Dutch Council for Culture, focus-
ing on youth, their social environment, media use, and positive and negative
media effects. We briefly discussed a dualistic picture of children (vulnerable
vs. savvy), a dualistic picture of new media technology (risks vs. opportu-
nity), and a dualistic picture of media (literacy) education (protectionist vs.
empowerment). While all these perspectives are present within the media
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

wisdom landscape, it is generally the empowering affordances of technology

as well as empowering children to use media wisely and actively that are
foregrounded in Dutch policy initiatives regarding this topic.
We described how media wisdom was translated many times and in
several ways. This allowed configurations of practices, policies, goals, and
social actors to emerge, which indicated that it is not a stable, natural, and
neutral concept, but instead media wisdom is highly contingent. We have
argued that it is an outcome of relations, of the various choices, practices,
and policies, made by the social actors involved in its continuous constitu-
tion. We suggest that while the recent Dutch scene has been characterised by
a successful mobilization of many actors around media wisdom as a bound-
ary object, it has been less successful in providing for long-term education
and training needs.
In addition, we have claimed that its various meanings, connotations, and
practices, provided by the social actors involved, and its contingent fluid
character shape the ‘media wisdom’ debate, policy, and practices. We have
demonstrated that media wisdom is adaptable, with various social actors
across many sectors each adopting their own meanings and practices, but
media wisdom is nonetheless common enough to bridge boundaries between
them, recognizable to all organizations and representatives. What they share
is a focus on developing the skills and competences of media users. We have
shown how the connotations of the term ‘wisdom’ contributed to the dis-
tribution of responsibility for stimulating media wisdom across numerous
actors instead of leaving it with the government and formal education sector,
as is the case with more traditional literacy. We traced how one party estab-
lishes itself as an obligatory passage point, controlling the move between
the common identity of media wisdom and its more adaptable side, by stan-
dardizing media wisdom through the development of a specific model. This
model served to order media wisdom initiatives and standardize the concept
and its practices, first by focusing on competences set within the model, for
all to use, in order to institutionalize it, across organizations and in formal
education. And second, with this distributed responsibility, several actors
are left out of the picture, including the technology itself and its contingency,
flexibility, and the multiplicity of its use and users, and technology designers
are similarly neglected.
In the Dutch situation described in this chapter, we observed that media
wisdom is mainly sought as a quality to be developed in new media users
 Makers of Media Wisdom 157
themselves, specifically in children, with media technology considered as
neutral, static, and homogeneous. We, however, have argued that media wis-
dom is also shaped by new media, in the sense that new media technologies
are never neutral, always changing and heterogeneous. Thus media wisdom
can be facilitated or hindered by the design of new media. Media wisdom
policies and practices need to take into account the full range of relation-
ships between technology and users, children as well as adults—in other
words, the network of relations between all of the human and non-human
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

actors involved, as responsibility for media wisdom can be sought in all of

them.

NOTES

1. The chapter draws on a selection of national and international policy docu-

ments, academic studies on ‘media literacy’ originating from various disci-
plines, and professional and popular literature focusing on developments in
technology, business, and education in relation to youth. The latter tend to
inspire many actors dealing with media wisdom. Semi-structured, face-to-face
interviews were conducted between January and June 2012 with five repre-
sentatives from five different Dutch media literacy initiatives, focusing on
their background, goals, work methods, teaching and research materials. In
addition, the annual networking event for media wisdom representatives was
attended in May 2012. All interviewees have been given pseudonyms. Inter-
views were conducted in Dutch and translated by Sprenkels. This is discussed
more fully within Sprenkels’s PhD thesis (forthcoming).
2. www.unesco.org/new/en/communication-and-information/media-development/
media-literacy, visited March 26, 2014.
3. www.unesco.org/new/en/communication-and-information/media-development/
media-literacy/mil-university-network/, visited March 26, 2014.
4. Although the Dutch government tends to be reluctant in its actions (Valken-
burg, 2005), it is stimulated by article 17 of the Convention on the Rights of
the Child on access to information and mass media to initiate action regard-
ing harmful media content. Section e reads: “State Parties shall encourage the
development of appropriate guidelines for the protection of the child from
information and material injurious to his or her well-being, bearing in mind
the provisions of articles 13 and 18” (United Nations, 1989). Important actors
in this matter are the Ministries of Education, Culture and Science; Health,
Welfare and Sport; and Security and Justice; the Dutch Media Authority; the
Netherlands Institute for the Classification of Audio-Visual Media; the Pan-
European Game Information age rating system; and the Advertising Code Com-
mittee (Nikken, 2013).
5. www.mediawijzer.net/competentiemodel-voor-mediawijsheid-gepresenteerd,
visited March 26 2014.
6. Similar to the case of the scallop researchers in Callon’s much-cited paper,
Mediawijzer.net imposes itself and its take on media wisdom on other actors
such as its network partners. It defines the identities and interests of other actors
that are consistent with its own interests and establishes itself as an obligatory
passage point through which the other actors must pass, rendering itself indis-
pensable. The actors are persuaded to identify with their roles and start speaking
on behalf of Mediawijzer.net, thus further strengthening its position.
 158 Isolde Sprenkels and Sally Wyatt
7. ‘leermiddelenbank mediawijsheid’
8. Mediawijzer.net brochure, year of publication unknown, available at www.
mediawijzer.net/wp-content/uploads/2013/05/PO-Stappenplan.pdf, visited
April 15 2014.
9. Mediawijzer.net also wants to initiate initiatives to integrate media wisdom in
secondary education (Evers, 2013). The Royal Netherlands Academy of Arts and
Sciences also wants to do so, as suggested in its report on digital literacy (KNAW,
2013). However, contrary to Mediawijzer.net, it wants schools to offer separate
courses in ‘information and communication’ for younger students and ‘informat-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ics’ for older students instead of integrating digital literacy in existing courses.
This digital literacy should focus on ‘basic knowledge,’ ‘use,’ and ‘behaviour.’
10. www.katholiekgezin.nl/index.php?option=com_content&task=view&id=
89&Itemid=110, visited May 8 2014.
11. According to Nelly Oudshoorn and Trevor Pinch, Woolgar “suggested that
how users ‘read’ machines is constrained because the design and the produc-
tion of machines entails a process of configuring the user” (Oudshoorn and
Pinch, 2005, 8).

REFERENCES

AUFDERHEIDE, P. (1993) Media Literacy: A Report of the National Leadership

Conference on Media Literacy. Aspen: Aspen Institute.
BAKARDJIEVA, M. (2005) Internet Society. The Internet in Everyday Life. London:
Sage Publications.
BAKARDJIEVA, M. & GADEN, G. (2011) Web 2.0 Technologies of the Self. Phi-
losophy & Technology, 25(3). p. 399–413.
BAZALGETTE, C. (1997) An Agenda for the Second Phase of Media Literacy Devel-
opment. In Kubey, R. (Ed.), Media Literacy in the Information Age: Current
Perspectives, Information and Behavior (Vol. 6). New Brunswick: Transaction
Publishers, 69–78.
BUCKINGHAM, D. (2011) Studying Children’s Media Cultures: A New Agenda
for Cultural Studies. In van den Bergh, B. & van den Bulck, J. (Eds.), Children
and Media: Multidisciplinary Approaches. Apeldoorn: Garant/Centrum voor
Bevolkings- en Gezinsstudie, 49–66.
CALLON, M. (1986) Some Elements of a Sociology of Translation: Domestication
of the Scallops and the Fishermen of St Brieuc Bay. In Law, J. (Ed.), Power, Action
and Belief: A New Sociology of Knowledge? London: Routledge, 196–223.
DROTNER, K. (2011) Children and Digital Media: Online, On Site, On the Go. In
Qvortrup, J., Corsaro, W. A. & Honig, M.-S. (Eds.), The Palgrave Handbook of
Childhood Studies. London: Palgrave Macmillan, 360–374.
EUROPEAN COMMISSION. (2007) Communication from the Commission to the
European Parliament, the Council, the European Economic and Social Commit-
tee of the Regions. A European approach to media literacy in the digital environ-
ment. COM(2007) 833 final. Brussels.
———. (2009) Commission Recommendation on Media Literacy in the Digital
Environment for a More Competitive Audio-visual and Content Industry and an
Inclusive Knowledge Society. C(2009) 6464 final. Brussels.
———. (2010a) Communication from the Commission. Europe 2020. A Strategy
for Smart, Sustainable and Inclusive Growth. COM(2010) 2020. Brussels.
———. (2010b) Communication from the Commission to the European Parliament,
the Council, the European Economic and Social Committee and the Committee
of the Regions. A Digital Agenda for Europe. COM(2010) 245 final. Brussels.
 Makers of Media Wisdom 159
EUROPEAN PARLIAMENT. (2008) Report on Media Literacy in a Digital World.
(2008/2129(INI). Brussels.
EVERS, F. (2013) Mediawijsheid Structureel Inbedden in Gehele Curriculum. Retrieved
April 02, 2014, from www.kennisnet.nl/themas/mediawijsheid/mediawijsheid-
structureel-inbedden-in-gehele-curriculum/
GILLEBAARD, H. et al. (2013) Kennispositie van Mediawijsheid Competenties: Inven-
tarisatie Onderzoek 2005—Heden. Utrecht: Nextvalue Research & Dialogic.
GRIZZLE, A. et al. (2013) Media and Information Literacy. Policy and Strategy
Guidelines. Paris: UNESCO.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

HASEBRINK, U. et al. (2008) Comparing Children’s Online Opportunities and

Risks across Europe: Cross-national Comparisons for EU Kids Online. London:
EU Kids Online.
HELSPER, E. J. & ENYON, R. (2010) Digital Natives: Where Is the Evidence?
British Educational Research Journal, 36(3). p. 503–520.
HOBBS, R. (2011) The State of Media Literacy: A Response to Potter. Journal of
Broadcasting & Electronic Media, 55(3). p. 419–430.
HOEVEN, M.J.A. VAN DER. (2006) Brief aan de Tweede Kamer over Mediawijsheid:
Burgerschap in de Informatiemaatschappij. Den Haag: Ministerie van OCW.
KNAW. (2011) Digitale Geletterdheid in het Voortgezet Onderwijs. Vaardigheden en
Attitudes voor de 21ste Eeuw. Amsterdam: Koninklijke Nederlandse Academie
van Wetenschappen.
KWARTIERMAKERSGROEP MEDIAWIJSHEID. (2008) Programma Mediawi-
jsheid. De Inrichting van een Expertisecentrum voor Mediawijsheid. Den Haag:
Ministerie van OCW
LATOUR, B. (1987) Science in Action. How to Follow Scientists and Engineers
through Society. Cambridge (MA): Harvard University Press.
LIEVROUW, L. A. (2006) New Media Design and Development: Diffusion of Inno-
vations v Social Shaping of Technology. In Lievrouw, L. A. & Livingstone, S.
(Eds.), The Handbook of New Media. Social Shaping and Social consequences of
ICTs. Updated Student Edition. London: Sage, 246–265.
LIVINGSTONE, S. (2004) Media Literacy and the Challenge of New Information
and Communication Technologies. The Communication Review, 1(7). p. 3–14.
———. (2009) Children and the Internet. Great Expectations, Challenging Realities.
Cambridge: Polity Press.
———. (2011) Media Literacy for All? On the Intellectual and Political Challenges
of Implementing Media Literacy Policy. In Livingstone, S. (Eds.), Media Literacy:
Ambitions, Policies and Measures. Brussels: COST (European Cooperation in
Science and Technology, 31–35.
LUNT, P. & LIVINGSTONE, S. (2012) Media Literacy. In Lunt, P. & Livingstone, S.
(Eds.), Media Regulation. Governance and the Interests of Citizens and Consum-
ers. London: Sage, 117–142.
MEDIAWIJZER.NET. (2010a) Meerjarenplan 2011 t/m 2014.
———. (2010b) Jaarplan 2011.
———. (2011a) Helderheid over Mediawijsheid. De Betekenis van “Meten van
Mediawijsheid” voor het Netwerk.
———. (2011b) Stimuleringsregeling 2012.
———. (2012) Competentiemodel: 10 Mediawijsheidcompetenties.
———. (2013a) Mediawijzer.net: Expertisecentrum voor Mediawijsheid.
———. (2013b) Stimuleringsregeling 2013.
———. (2014) Partners Mediawijzer.net Februari 2014.
MINISTERIE van OCW. (2006) Mediawijsheid, Leven in de Gemedialiseerde
Samenleving. Den Haag: Ministerie Onderwijs Cultuur en Wetenschap.
NIKKEN, P. (2013) Media-risico’s voor Kinderen. Een Verkenning. Utrecht: Neder-
lands Jeugdinstituut.
 160 Isolde Sprenkels and Sally Wyatt
OUDSHOORN, N. & PINCH, T. (2005) Introduction: How Users and Non-Users mat-
ter. In Oudshoorn, N. & Pinch, T. (Eds.), How Users Matter. The Co-construction
of Users and Technology. Cambridge (MA): MIT Press, 1–28.
PLASTERK, R. & ROUVOET, A. (2008) Kabinetsvisie Mediawijsheid. Den Haag:
Ministerie van OCW.
RAAD VOOR CULTUUR (2005) Mediawijsheid. De Ontwikkeling van Nieuw
Burgerschap., Den Haag: Raad voor Cultuur.
SPRENKELS, I. & PLOEG, I. VAN DER (2014) Follow the Children! Advergames
and the Enactment of Children’s Consumer Identity. In van der Hof, S., van den
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Berg, B. & Schermer, B. (Eds.), Minding Minors Wandering the Web: Regulating
Online Child Safety. The Hague: Asser Press, 173–190.
STAR, S. L. (2010) This Is Not a Boundary Object: Reflections on the Origin of a
Concept. Science, Technology & Human Values, 35(5). p. 601–617.
STAR, S. L. & GRIESEMER, J. R. (1989) Institutional Ecology, ‘Translations’ and
Boundary Objects: Amateurs and Professionals in Berkeley’s Museum of Verte-
brate Zoology, 1907–39. Social Studies of Science, 19(3). p. 387–420.
UNITED NATIONS. (1948) United Nations Universal Declaration of Human Rights
1948.
———. (1989) Convention on the Rights of the Child.
UNESCO. (1982) Grunwald Declaration on Media Education. Grunwald, Germany.
———. (2005) Beacons of the Information Society: The Alexandra Proclamation
on Information Literacy and Lifelong Learning. Alexandria, Egypt: UNESCO.
———. (2007) Paris Agenda or 12 Recommendations for Media Education. Paris,
France: UNESCO.
———. (2013) Global Media and Information Literacy Assessment Framework:
Country Readiness and Competencies. Paris, France: UNESCO.
VALKENBURG, P. (2005) Schadelijke Media en Weerbare Jeugd: Een Beleidsvisie
2005–2010. Amsterdam: Amsterdam School of Communications Research.
WIEBENGA, F. et al. (2011) Meten van Mediawijsheid. Een Studie naar een Raam-
werk, Meetmiddelen en Toepassing Hiervan. Amsterdam: Mediawijzer.net.
WILSON, C. et al. (2011) Media and Information Literacy. Curriculum for Teach-
ers. Paris, France: UNESCO.
WYATT, S. (2005) Non-Users Also Matter: The Construction of Users and Non-
Users of the Internet. In Oudshoorn, N. & Pinch, T. (Eds.), How Users Matter.
The Co-construction of Users and Technology. Cambridge (MA): The MIT Press,
67–80.
———. (2008a) Technological Determinism Is Dead; Long Live Technological
Determinism. In Hackett, E., Amsterdamska, O., Lynch, M. & Wajcman, J.
(Eds.), The Handbook of Science & Technology Studies. Cambridge (MA): MIT
Press, 165–180.
———. (2008b) Feminism, Technology and the Information Society. Learning from
the Past, Imagining the Future. Information, Communication & Society, 11(1).
p. 111–130.
WYATT, S. et al. (2005) The Digital Divide, Health Information and Everyday Life.
New Media & Society, 7(2). p. 199–218.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Part III

(Mis)Behaving
Suspects and Deviants
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 8 Data Mining ‘Problem Youth’
Looking Closer But Not Seeing Better
Francisca Grommé
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

In 2011, the Dutch municipality of Burgcity conducted a pilot study about

data mining. It aimed to find out whether this statistical technique could be
used to improve its understanding of ‘problem youth,’ loosely defined by the
city as youth below 23 years of age who are likely to commit minor offences
such as vandalism, littering, or shop theft.1 In particular, it aimed to learn
whether a combination of municipal data, police data, and commercial data
about consumption could lead to new insights for youth crime policy.
Especially salient in this pilot study was the policy makers’ use of the
metaphor of ‘zooming in.’ The policy makers expected data mining to pro-
vide knowledge that was local, particular, and timely. This understanding
of data mining is not unique to Burgcity. Indeed, proponents of data mining
promise increased detail and granularity. As was stated in a project plan,
data mining techniques would generate “local theories” that “the general
theories of social science” cannot provide.
Vision metaphors are never innocent, however, as Donna Haraway has
famously argued (1991). The metaphor of zooming draws on the imagery of
a mechanical lens, which reveals more detail about an object. The resulting
‘close-ups’ are assumed to have a high truth status because of their implied
precision. In this chapter, I am interested in how digital data are put to use
according to a rationale of zooming in to profile problem youth. Profiles can
be understood as identities that are ascribed to youth and used as bases for
government intervention (Pridmore, Chapter 2, this volume).
I aim to challenge zooming as an underlying principle in data mining
practices and scholarly and professional accounts. I do so because sticking
to this metaphor risks mobilization of data mining in an argument for the
acquisition of ever more personal data. Namely, if a technology by itself can
provide ever more detailed representations of youth, all it needs are more or
better data. Furthermore, the metaphor obscures the normativities that are
part of the practice of data mining. With this I mean that it diverts attention
from the ‘goods’ and ‘bads’ that come into play when ‘detailed knowledge’
is produced.
 164 Francisca Grommé
My questions are how zooming in was done in the Burgcity data min-
ing pilot and what norms were embedded in and produced through these
practices. I attend to the bodies of knowledge, discursive practices, and arte-
facts that were part of zooming in. Adopting a material-semiotic approach,
I focus on the heterogeneous relations that bring objects, such as problem
youth, into being (Mol, 2002; M’charek, 2013). From this approach, it fol-
lows that by zooming in, one does not simply see the same thing in more
detail. Instead, the practices of zooming in bring new objects into being
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

(Strathern, 2005). I draw on Charles Goodwin’s work to describe data min-

ers’ practices as ‘situated improvisation’ (1995, 1996).
This chapter is based on fieldwork that I conducted before, during, and
after the Burgcity pilot for a period of 18 months. I focus on the interac-
tive sessions at the core of the pilot, in which policy makers and corporate
experts analysed the data. We will follow the participants through various
attempts to zoom in. First, from zooming in to the level of the sub-city dis-
trict, we learn about two modes of situated improvisation the participants
engaged in: evocation and comparison. Second, from zooming in to the level
of the neighbourhood, we learn from the trouble that the participants ran
into: they could not zoom in any further without losing sight of the problem
youth. This part if the pilot draws out the regimes of evidence that are part
of zooming in.
Attending to the practicalities of zooming in also allows us to learn
that it is normative work. I show that results needed to be made relevant
as surprises; that the ‘lens’ focused on the neighbourhood as a source for
truth; that city youth were constituted as a norm; that detail is produced
by the application of general categories; and that zooming in includes mak-
ing judgements about good knowledge for government. These normativi-
ties suggest that taking seriously the metaphors by which technologies are
brought into practice might be a good starting point to change the terms by
which digital identities are produced.

ZOOMING IN AS SITUATED IMPROVISATION

Mining for Local Knowledge

Ever more digital data are available for analysis. People produce increas-
ing amounts of data through activities as simple as browsing on the Inter-
net and using a chip card on public transport. This development is joined
by a growing capacity to search and analyse these data. Data mining is a
statistical technique often used for the analysis of big datasets. It is com-
monly referred to as “the automatic or semi-automatic process of discover-
ing patterns in data” (Witten, Eibe, and Hall, 2011, 5) or “the application
of specific algorithms for extracting patterns” (Fayyad, Piatetsky-Shapiro,
and Smyth, 1996, 39). In everyday usage, it can refer to both software and
analytical skills.
 Data Mining ‘Problem Youth’ 165
Data mining is argued to challenge traditional science, because, in contrast
with statistical techniques such as regression analysis, the software allows
the analyst to search for relations in the data without defining hypotheses
and limiting the number of variables in advance (Witten, Eibe, and Hall,
2011; Hildebrandt, 2008). Instead, an algorithm is applied to automatically
find co-occurrences in a large dataset that can comprise thousands of vari-
ables.2 Industry therefore advertises data mining as ‘digging’ into the data
to find ‘nuggets of gold’ (yet other metaphors).
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

It needs to be noted that although data mining provides analysts with

new possibilities, it is often practiced as a combination of old and new sta-
tistical techniques. In the application studied in this chapter, for instance,
conventional geodemographic marketing techniques are combined with
data-mining algorithms.3 Furthermore, in practice, patterns are often not
found automatically but rely on the expert’s insight to choose variables (Ang
and Goh, 2013) Following boyd and Crawford, it therefore seems appro-
priate to approach data mining not as a “higher form of intelligence and
knowledge that can generate insights that were previously impossible,” but
as a mythology as well as a technological development (2012, 663).
In policy practice, data mining is most often used to create profiles: sets of
correlations that can be used to identify or represent individuals or groups
(Hildebrandt, 2008, 19). It is applied in a variety of policy domains, such
as anti-terrorism programs, programs against tax evasion, welfare policy
and, as this chapter describes, local crime policy. In these cases, the promise
of data mining is specificity: More data and diverse data sources generate a
closer view. Policy makers are seldom interested in general patterns, policy
ethnographies show (Yanow, 2002; Choy, 2005). They are in search of par-
ticularities, and commercial data mining is brought in for this purpose. This
is also reflected in the marketing analytics industry, which has specialized in
the provision of ever more locally specific knowledge (Phillips and Curry,
2003).
Empirical scholarly work on the use of data mining and large datasets by
(local) policy agencies mainly describes the limitations of data mining. It is
argued that digital data do not necessarily lead to descriptions of individu-
als and groups that are more precise. In Gary T. Marx’s words, users of this
technology “see hazily (but not darkly) through the lens” (2005, 339). A
fundamental reason is that profiles give very little information about indi-
viduals and small groups of people because they aggregate data. The charac-
teristics of the profile may therefore not be applicable to all the individuals
it represents (Curry, 2004; also see Custers, 2004). It is also argued that the
data do not represent individuals correctly because government datasets are
often incomplete or simply wrong (Pleace, 2007). Moreover, not all types of
large datasets are valuable or reliable. This is the case for some social media
data, such as statements on Twitter (boyd and Crawford, 2012). In acquir-
ing data, furthermore, local governments are limited by privacy restrictions.
An example is the use of police data. Last, the possibilities of acquiring
 166 Francisca Grommé
commercial datasets are often limited for financial reasons and because com-
panies are not always allowed or willing to share them (Pleace, 2007; boyd
and Crawford, 2012).

Localized and Embodied Visions

These considerations put data mining into perspective. The reason I raise
them here, however, is because they bring to mind a similar issue in anthro-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

pology. In Partial Connections (2005), Marilyn Strathern identifies the

problem that there never seems to be sufficient data to adequately describe
social life. Whenever one looks closer or from another angle, there still does
not seem to be enough data. One would expect complexity to decrease when
a smaller part of society is examined; instead, there always seems to be
something new to discover. The similarity between these two practices—
data mining and anthropological observation—is that they are both based
on the premise that seeing better requires more or better data.
Strathern takes issue with the idea that there are an infinite number
of perspectives, each requiring more data. She argues that all perspectives
exist as “localized, embodied visions” (40). These visions do not present
different versions of reality but enact the object in different, partially con-
nected ways (also see Mol, 2002). To change perspective, therefore, does
not amount to seeing part of a whole. For zooming in, this means that one
does not see a part of the same object in more detail. When scientists, pro-
fessional analysts, policy makers, or others zoom in, they bring an object
into being.
This is a relevant distinction because it avoids the notion that technologies
by themselves reveal the characteristics of populations that exist indepen-
dently of the practices by which they are mapped (M’charek, 2005; Rup-
pert, 2011). This notion of technology is problematic in academic discourse
because it feeds into the idea that lack of data or analytical power obstructs
perfect close visions. It therefore fails to question data mining in terms other
than those of an information problem. I suggest examining how zooming in
is done in practice in order to overcome this limitation.

Situated Improvisations
To be sure, to insist on a critical attitude towards metaphors of vision is
not new.4 Donna Haraway is especially known for her contribution to this
project. In Situated Knowledges (1991), she argues that, with contemporary
visualizing technologies, “Vision in this technological feast becomes unregu-
lated gluttony; all perspective gives way to infinitely mobile vision, which no
longer seems just mythically about the god-trick of seeing everything from
nowhere, but to have put the myth into ordinary practice” (189). Claims
for scientific objectivity are based on a ‘view from nowhere,’ thus creating a
 Data Mining ‘Problem Youth’ 167
science that is detached and totalizing. What is needed, Haraway argues, is
a commitment to situated knowledges: knowledges that are understood as
local, embodied, and partial.
Inspired by this body of work, Christopher Gad and Peter Lauritsen (2009)
suggest that social scientists with an interest in surveillance practices such as
digital identification should study them as situated phenomena. To observe
something, actors in the field of surveillance need to draw together bod-
ies of knowledge, artefacts, and human actors at a certain place and time.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Surveilled objects thus achieve their status in terms of the relations between
heterogeneous entities (M’charek, 2013).
I use this notion of situatedness to study zooming in as a metaphor com-
parable to Haraway’s vision metaphor of ‘seeing everything from nowhere.’
Goodwin’s work about situated improvisations (1995) provides a point
of departure for understanding how the participants of the Burgcity pilot
zoomed in on youth. Goodwin coined the term to describe how an object of
knowledge emerges through the interplay of screens, common knowledge,
lay theories, everyday artefacts, and professional repertoires (606). Impor-
tantly, with improvisation I do not refer to random actions but to actions
that are informed by bodies of knowledge, artefacts, and emerging priorities
at a certain place and time.
Following Goodwin, I will attend to talk and gestures by which the rela-
tions between these heterogeneous entities were achieved. For instance,
Goodwin describes how marine biologists highlighted the information pre-
sented to them on a screen by pointing at it with a pencil or by making
comments such as “that nice feature again” (262). In his work on archaeol-
ogy, he shows that some of the talk makes some results more relevant than
others; it sets out the conditions of relevance (1996). Furthermore, from
Goodwin’s analysis of the Rodney King trial, we learn that some actors may
use the status of their profession to influence enquiries. This is what he refers
to as professional privilege (1996).

DATA MINING PROBLEM YOUTH: A PILOT STUDY

Marketing Intelligence Methods

Burgcity is a medium sized Dutch city of about 200,000 inhabitants. The pilot
study took place at the Department of Community safety, which employed
16 persons at the time I was present. Data Inc. is a data analyst company
employing no more than 10 persons. It had mainly accepted assignments
from the Dutch police force and corporations in the Netherlands. Data Inc.
provided the software, collected and cleaned the data, and performed the
analyses.
The case description and the case study that follow are based on observa-
tions of the pilot meetings, the city’s policy makers’ everyday routines, and
 168 Francisca Grommé
interviews. As the reader might suspect, Burgcity and Data Inc. are fictitious
names, as are the names of all other organizations, places, and persons men-
tioned in this chapter (with the exception of the international corporation
that delivered the commercial data used in this pilot: Experian). I agreed to
anonymity because confidential information regarding suspects, their back-
grounds, and crime-control practices often passed my eyes on the Burgcity
work floor. I also agreed because I felt a responsibility to protect the liveli-
hood of my informants. I believe this is an especially sensitive issue because
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

both the company and the city were involved in an experimental activity,
making themselves vulnerable as they treaded uncertain ground.5
The pilot study’s aim, as stated in Data Inc.’s project proposal, was to
develop “more efficient and effective approaches”6 to non-criminal nuisance
and minor criminal offences committed by youth. The use of the Experian
dataset, which contains ‘lifestyle data,’ was advertised as the pilot’s main
innovative feature. ‘Marketing intelligence methods’ were to lead to new
insights into problem youth, their life worlds, and their motivations. If one
knows that the “most troublesome youth” wear Nikes, a new approach in
policy might be to involve Nike stores in a campaign against vandalism, one
of the project’s initiators argued in a 2010 presentation.
The Experian lifestyle dataset incorporates about 1,500 variables on basic
information about household income and age; interests and activities, such
as membership to a sports club; and media usage, such as Internet usage
and newspaper subscriptions. These variables also include pre-set profiles,
referred to as Mosaic groups, such as ‘free spirits,’ ‘mini machos,’ or ‘digital
families.’ The data come from sources such as property or telephone regis-
trations and surveys. Experian datasets are sold for marketing purposes, to
develop strategies for the collection of bills and debts, and to manage finan-
cial risk.
These data were combined with police data about suspects and offences
and with municipality data including variables such as age, income, and
school attendance. A remark about the police dataset is in order here. It
contains records about individuals suspected of one or more offences. Even
if police investigation has pointed out that an individual was not involved
in an incident, he or she remains registered. For the purposes of the pilot,
young suspects were used as an approximation of problem youth, which
raises the obvious issue of whether these data can really help Burgcity learn
about the group it is interested in.
Data mining was expected to help the city understand problems more
“thoroughly and quickly,” understand patterns that characterize problem
youth at neighbourhood level, identify the groups that need attention, and
identify and influence causal relations. Thus, it would allow the city’s policy
makers to “zoom in” on local youth and generate “local theories” (Inter-
view, former CEO Data Inc., November 12, 2010). In policy maker Anna’s
words, it could be another method to “reach out” to Burgcity’s inhabitants
(Interview, September 5, 2011).7
 Data Mining ‘Problem Youth’ 169
Pilot Results
Prior to the pilot, Burgcity’s Department of Community Safety had already
considered data mining to improve the city’s ‘information position.’ As the
department’s information specialist put it, “data mining is definitely going to
happen in Burgcity, there is no doubt about it. The head of our department
asks me how data mining is coming along every week” (Interview, April 21,
2011). So when the department head was asked by an innovation platform
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

to participate in the pilot, he consented. The funding was supplied by a grant

from the Dutch Ministry of Internal Affairs.
The innovation platform chose to focus on problem youth, which was
and still is a hot topic in Dutch crime policy. Next, it was up to Burgcity to
select a case study. It chose its newest and demographically youngest city dis-
trict, Molendistrict, as a case study. Molendistrict has about 40,000 inhabit-
ants. The district seemed to suit the purposes of the pilot because one third
of Burgcity’s youth lives here, and nuisances in the area are of considerable
concern for the city’s staff at the Department of Community Safety.
Data Inc. insisted that the policy makers’ input was crucial to assess which
results were of value for Molendistrict. Three of Burgcity’s civil servants
were involved in the pilot: Mieke, Anna, and Liesbet. Mieke is the depart-
ment’s information specialist, Anna the department’s specialist on youth,
and Liesbet the district manager of Molendistrict. The group would ideally
explore the possibilities of the software together to learn what insights could
be gained by using it. Data Inc. suggested taking an “iterative approach,”
whereby Burgcity’s policy makers would formulate questions, find answers,
and “learn how to ask better questions” (Fieldwork notes, February 28,
2011).8
Next to preparatory and strategic meetings, the city’s civil servants and
Juriaan, Data Inc.’s analyst, met four times for interactive sessions in Burgc-
ity’s city hall. During these sessions, Juriaan operated a laptop and the city
employees faced him. Everybody in the small meeting rooms could see the
computer interface on an LCD screen. During the meetings, a question for
analysis would be formulated by the policy makers or by Juriaan. Juriaan
would then attempt to answer the question, showing the results on the
screen.
Three types of results were generated. First, frequency tables presented
counts, for example, of the number of suspects living in each neighbour-
hood. Second, profile analysis was used to compare two or more popu-
lations. The resulting table displayed the variables for which significant
differences between the populations were found. Third, the data mining
system was used to generate ‘hotspot’ maps: clusters of offences committed
in a certain location on the map.
The pilot did not result in a set of stable and generally acknowledged con-
clusions about problem youth and youth crime policy. Data Inc. did present
a number of conclusions in the final report, amongst others about the types
 170 Francisca Grommé
of neighbourhoods that have a relatively high risk of youth delinquency,
such as neighbourhoods in which household income varies strongly (Data
Inc. 2012, internal document). In the end, however, the report stated that it
could not characterize problem youth using the lifestyle data due to privacy
restrictions (on which I will elaborate later in the chapter).
In what follows, I describe how zooming in was done and what bodies of
knowledge and objects informed the efforts to obtain a closer view. I attempt
to understand how, once the results had been presented on the screen, a close
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

view of youth “emerged through the interplay between a domain of scrutiny

(. . .) [and] a set of discursive practices (. . .), being deployed with a specific
activity” (Goodwin, 1996, 606). Two themes that regularly returned in the
sessions are presented to the reader: the place of residence of youth that
commit offences in Molendistrict and the use of lifestyle data to character-
ize youth that commit offences. I have chosen to elaborate on these themes
because they were discussed most frequently and thoroughly during the ses-
sions and therefore are the most instructive about the practice of zooming
in. For each of these themes, I focus on the participation of one of the policy
makers.

ZOOMING IN FROM CITY TO DISTRICT

Surprise
The first story is about Liesbet’s question for the technology. It demon-
strates the particularities of zooming in from the city to the district level.9
Liesbet is the district manager for Molendistrict. It is her job to facilitate
communication between the central city and Molendistrict’s local politi-
cians, interest groups, and case workers. As she puts it, she defends the
district’s interests at city level. Of all the participants in the pilot, she visits
the district most frequently. This is what data mining should answer for
Liesbet:

Where in Molendistrict do the young people who commit offences

live? And how can useful policies be developed for youth who commit
offences in Molendistrict but live in other districts?
(Data Inc., internal document, June 18, 2010)

These questions require zooming in. It is necessary to look closer, she

explains, because, although she is familiar with the area and its inhabitants,
it is difficult to learn about the causes of youth crime. Statistics, she argues,
often obscure local circumstances because they aggregate data. Further-
more, because the neighbourhood is relatively new (construction started in
the early 1980s), she feels that the social networks that should theoretically
help her stay informed do not yet exist.
 Data Mining ‘Problem Youth’ 171
To find the answer to Liesbet’s question, Juriaan starts out by producing
a frequency table about Molendistrict. A frequency table is a basic list that
presents the number of occurrences in a query.

Liesbet, Mieke and Anna take a good look at the results presented on
the LCD screen. It shows a list of numbers: the number of young sus-
pects in Molendistrict, broken down by neighbourhood. “Neighbour-
hood H, this surprises me,” Liesbet says. “Neighbourhood K makes
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

sense because it has a shopping mall. But neighbourhood H is only a

small residential area.”
Anna asks her to show neighbourhood H on the map on the wall.
Liesbet points at it: “There it is, the park is also part of it, nowadays.”
“Is there a sports park nearby?” Anna wonders. This is not the case, but
Liesbet notes that there is a swimming pool in the area.
Anna has another suggestion: perhaps these are conflicts between
neighbours? “No, not in neighbourhood H,” Liesbet replies, “this is not
a neighbourhood known for fights between neighbours.” Next, Liesbet
wants to know more about neighbourhood H: “This is making me curi-
ous, what types of offences were committed here?”
(Fieldwork notes, April 7, 2011)10

This tells us that not all results count. A result counts when it is surpris-
ing, as in the case of neighbourhood H. Surprise, in Goodwin’s terms, is a
condition of relevance for data mining. A close view is a view that is reveal-
ing. It provides a look under the surface.

Evoking
The group returns to the question of problem youth’s place of residence at the
end of the meeting. This time they take a different approach. To learn more
about the questions that neighbourhood H raises, they need to know more
about the registered suspects. Therefore, Juriaan creates a frequency table that
shows information about the suspects for offences committed in Molendis-
trict. This table is different from the previous one because it does not present
youth living in Molendistrict but youth that have committed an offence there.

Many of the young suspects come from other districts, Liesbet notes.
She points at the map: “Neighbourhood G, this is quite far away, it is in
another district.” But “distance is not so important to youth,” she adds.
Juriaan agrees: “Youth have a large radius of activity.” They look a little
longer at the list. “These are neighbourhoods with a lot of social hous-
ing,” Liesbet notes. “Indeed, one of these neighbourhoods is a weak
neighbourhood,” Juriaan adds.
(Fieldwork notes, April 7, 2011)
 172 Francisca Grommé
So youth travel from other places to Molendistrict. They might come
from neighbourhoods in other districts with fewer facilities, the group adds,
and when they come to Molendistrict, they cause a nuisance.
This is the type of account that was often generated in this pilot: proposi-
tions, statements, descriptions, and explanations that loosely hang together.
An important part of producing an account of Molendistrict’s suspects is
‘evocation’; the data need to come to life and make sense, to be made ‘tel-
lable’ (see also Curry, 2004; Ziewitz, 2011). This first of all required maps.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

During the same session described in the first fragment, Anna asks Liesbet
to show neighbourhood H on the paper map on the wall. Almost every
meeting room has one of these maps. They are large, about one and a half
meters by one meter, and detailed. They show the borders of the districts and
neighbourhoods as designated by the Dutch central government (National
Statistics Netherlands or CBS).
The map displays what the screen does not: the borders between neigh-
bourhoods, the distances between the neighbourhoods, and a detailed street
plan. The data mining software can plot the results on a digital map, but this
is an oblique map: It does not show borders and specificities. The detailed
paper map helps Liesbet locate the neighbourhood and tie it to her previ-
ous experiences: For instance, neighbourhood H is not known for conflicts
between neighbourhoods. Maps also help to fill in the particulars of the
neighbourhood. They show whether there is a park nearby, or a shopping
mall. Furthermore, they show distances, revealing that youth travel. So
pointing at a map is part of bringing the data to life.
The second, and related, part of making the data come to life is to tie
the results to policy theories, categories, and local knowledge. This might
be policy’s equivalent of what Mariana Valverde refers to as adminis-
trative knowledge in legal practices: “in-between knowledge” used by
officials that is neither scientific nor lay (Valverde, 2003, 20). When they
talk through these policy theories, categories, and local knowledges, the
results on the screen are pieced together. For instance, neither the fre-
quency tables nor the maps give information about social housing, yet
the policy officers apply this characteristic to the area on the map. Subse-
quently, they apply the notion that neighbourhoods with social housing
are poorer, and poorer neighbourhoods are more troublesome. Juriaan’s
additional remark about ‘weak neighbourhoods,’ moreover, pieces the
account together. He refers to the 40 neighbourhoods the national gov-
ernment has pointed out as the most problematic in the country in 2007
(also known as Vogelaarwijken).
My point here is not that the city’s knowledge base is not ‘scientific’
enough. Policy practitioners need at hand knowledge to do their jobs. Rather,
I argue that this mode of improvisation shows that producing such accounts
under the banner of more granularity introduces more general categories. In
this example, moreover, a neighbourhood category invokes judgments about
youth behaviour.
 Data Mining ‘Problem Youth’ 173
This is also relevant for the outcome of these sessions: namely, that youth
who commit offences in Molendistrict are not from Molendistrict but from
other, ‘bad’ neighbourhoods. This outcome was partly informed by local
politics. At the time of the pilot, Liesbet was involved in a discussion about
the construction of new facilities in Molendistrict, such as practice rooms
and bars. The dominant notion was that more facilities would help reduce
complaints about youth behaviour, as youth could engage in more activities.
Liesbet reasoned that facilities might in fact cause more complaints because
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

problem youth would come from other places to use them.

Comparing
At the next meeting, the group returns to the theme of the Molendistrict sus-
pects. The question is how neighbourhoods within Molendistrict differ. Juriaan
performs a ‘profile analysis’—a comparison between young suspects living in
Molendistrict neighbourhoods with all inhabitants of the district. He states:

Now we see that of all persons we are looking at, all 772 [registered sus-
pects aged 23 and younger], 5.3 per cent lives in neighbourhood K. Of all
Molendistrict inhabitants, 3.4 per cent lives in neighbourhood K. This is
a difference, but not dramatically so. The difference in percentage is 1.9.
(Fieldwork notes, April 26, 2011)

Not everything counts as a surprise. From Juriaan’s quote, we learn that

surprise also comes with difference. Neighbourhood youth need to be dis-
tinguished from district youth.
Comparison is another mode of situated improvisation, next to evoca-
tion. In order to zoom in, one needs to find a difference. This is built into
Data Inc.’s software as ‘profile analysis.’ As with evocation, however, the
results on the screen do not make sense by themselves. This is an instance
in which the pilot’s participants aimed to zoom in from city to district:

Juriaan is asked to compare young suspects living in Molendistrict with

suspects living in the city as a whole. First, the results on the screen show
that 17% of the Molendistrict group is suspected of vandalism, against
11% in the city.

“So it must be the case that city youth commit different types of
offences, whereas Molendistrict youth mostly commit vandalism,”
Liesbet argues.

Next, Juriaan turns to the absolute numbers. He shows that there

were about 7,000 young suspects in Burgcity in 2007–2010, of which
about 600 suspects live in Molendistrict. Juriaan asks Liesbet for the
total number of inhabitants. Liesbet answers that 40,000 people live in
the district, compared to 200,000 in the city.
 174 Francisca Grommé
“So the city is only five times as big, while the number of suspects
is about ten times larger.” This means that, relatively, Molendistrict
youth are not that bad, Juriaan concludes.
(Fieldwork notes, April 7, 2011)

The issue of size is central in this fragment, and Liesbet’s knowledge of

the numbers is basic but crucial. In this case, comparison serves to estimate
the size of the problem of Molendistrict youth delinquency. Interestingly, the
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

group concludes that the problem of youth delinquency is relatively small.

At other times, a paper map was used to compare. Liesbet was not always
sure of the neighbourhoods’ sizes in terms of numbers of inhabitants, but a
map could be used to estimate the geographical size. In the one meeting in
which no maps were present, Anna got up, walked to the whiteboard, and
drew a map for Juriaan. She did so to demonstrate to Juriaan the relatively
large size of a neighbourhood compared to the district as a whole.
The two modes of improvisation, evocation and comparison, indicate
that the results on the screen did not establish close views on their own. Part
of the work of zooming in only started when the results had appeared on the
screen. It was done by applying professional and everyday knowledges and
reading paper maps. Problem youth, therefore, was by no means an identity
established by algorithms alone.
Zooming in, moreover, was normative work. First, the use of maps in this
section draws attention to a, perhaps obvious, characteristic of this practice:
the idea of formal neighbourhoods as units for policy. Consequently, the
nature of magnification is geographical; it focuses on a magnification of
neighbourhood processes. This is telling, as income or educational levels
could have also served as focal points for zooming in. Instead, the project
group focused on the physical and spatial characteristics of the neighbour-
hood, such as facilities.
Neighbourhoods are taken to be the determining factor in many urban
processes, and the neighbourhood is central in policy practice in the Nether-
lands as a whole, as exemplified by national programs that focus on ‘weak
neighbourhoods.’ This focus on the neighbourhood derives from an era of
social science in which it was also a marker of social class. As David Phillips
and Michael R. Curry note, it brings into practice the notion that “you are
where you live” (2003, 143). The latter was especially true for the Burgcity
policy makers, as they were not only looking for pragmatic information to
sell products, but they wanted “deeper information” to “get close to youth”
(Interview, November 15, 2010).
Second, through comparison, the entity representing the ‘whole’ is con-
stituted as the norm. So when Molendistrict youth are compared to city
youth, the city is the norm. In this case, Molendistrict’s deviation from the
norm was positive: The problem of youth delinquency was, comparatively
speaking, small.
 Data Mining ‘Problem Youth’ 175
ZOOMING IN TO NEIGHBOURHOODS

Lifestyle Analysis: A Discovery

For Anna, the city’s policy specialist on youth, results at district level are
not specific enough. She wants to know what characterizes problem youth
in specific neighbourhoods. This closer view of problem youth, however,
was never established in the pilot study. In fact, Data Inc. finally suggested
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

zooming out in order to zoom in.

Anna participated in the meetings as the city’s project’s manager. Her
main concern was to find out whether the software could be of added value
to policy. Her point of departure was ‘the funnel model’ (Interview, July 18,
2011; Burgcity 2012, internal document). According to this model, 20% of
all youth is in the lower end of the funnel’s cone; these youth are at risk of
committing an offence. The 4% in the tip of the cone is out of the police’s
and government’s reach. Anna’s aim was to find proactive approaches for
this 4%.
To this end, Anna contributes the following question to the pilot
study: What types of problem youth can be found in the neighbourhoods
in terms of combinations of characteristics and behaviour? (Data Inc.,
June 18, 2010, internal document). This question is to be answered with
the lifestyle data. During the meetings, Anna poses her question with
some humour: “Do girls that read Tina [a Dutch teen magazine] set fire
to litterbins more often?” On April 26, 2011, the group finally discovers
something:

Partly joking, Anna introduces what has been her main question for the
past few meetings:

“Which offences do girls that read Tina commit? Or, in other words,
can we find a relation between offences and the Mosaic [Experian]
data?” When Juriaan does not reply straight away, she rephrases the
question: “So can you say that people who own a Jaguar or whose
parents own a Jaguar are guilty of different offences than youth
whose parents drive a Mini?”

Juriaan’s first step in generating an answer to this question is to find

the characteristics that distinguish young suspects in Molendistrict from
all inhabitants of Burgcity. He presents a long list of characteristics that
differentiate these youth on the LCD screen. We only see the character-
istics for which the difference is statistically significant. Strong correla-
tions with high values are shown at the top of the list; weak correlations
are at the bottom. Among the latter are the Experian lifestyle variables.
Anna and her colleagues had been waiting for Juriaan to use the lifestyle
 176 Francisca Grommé
data, so quite excitedly they ask him to scroll down. Suddenly Anna
exclaims:

The sport darts! Yes, so the pilot is a success!

(Fieldwork notes, April 26, 2011)

Anna finds what she is looking for: an association between lifestyle and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

youth crime and nuisance. To be sure, she does not believe this relation to
be causal. In fact, the nature of the relationship is not relevant to her. But
if it would be possible to find a ‘segment’ of Molendistrict problem youth
that could be located in a darts club, that reads Tina, or whose fathers drive
Jaguars, she would know how to reach them (Interview, November 15,
2010).
In the meeting, Anna continues to discuss how these findings could be
used to inform policy. She reasons that, on the basis of these and similar pat-
terns on the neighbourhood level, one could initiate a marketing campaign.
One could address darts associations, for instance, to reduce vandalism.
Anna does not expect this approach to lead to drastic reductions of vandal-
ism rates; however, if offences would decrease from 100 to 80 incidents, this
would be a satisfactory result. “We can ask a marketing intern whether this
is a useful tool for a communicative government,” Anna concludes. (Field-
work notes, April 26, 2011)

How Surprise Disappeared

Anna’s discovery did not make it to the end report. When I asked Juriaan
about it shortly after the pilot group’s last meeting, he did not remember
the finding. In fact, Juriaan had discarded it soon after its discovery, arguing
that there were simply not enough young suspects living in the individual
Molendistrict neighbourhoods to perform such an analysis.
There was an additional problem. The police had not granted the proj-
ect group access to the names and individual addresses of the young sus-
pects due to privacy restrictions. This was a problem, because the lifestyle
information needed to be related to the suspects on the basis of residential
address, while the police had only provided information about the neigh-
bourhoods these youth live in. The lifestyle data therefore also needed to
be aggregated to neighbourhood level.11 Accordingly, a correlation between
darts and registered suspects only meant that a number of people played
darts in the suspects’ neighbourhood; it did not mean that the suspects actu-
ally played darts themselves.
Anna was disappointed. Not only were the results statistically significant,
but in a previous meeting, one of Juriaan’s Data Inc. colleagues had argued
that even if a difference in terms of lifestyle data was found on the basis of
 Data Mining ‘Problem Youth’ 177
only 4 suspects in one neighbourhood and 10 in another, this could be an
interesting lead for a policy maker. Even though the numbers are small, the
difference between 4 and 10 is telling, so this analyst argued.
There were two ways in which using small numbers as a basis for policy
were discussed. The first of these comes from the world of marketing. As
Juriaan explained:

Well, in marketing, it’s like this . . . When you are flyering, for instance,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

it is about making very small differences. So if you start a campaign that

leads to a difference [in sales] of one per cent in one neighbourhood and
0.7 per cent in the other . . . you could say this is insignificant but it
actually is a very good result.
(Fieldwork notes, February 28, 2011)

Marketing had inspired Anna to think about lifestyles in the first place.
To Juriaan she responded, “It is the same for us, really. I mean, even if one
person is prevented from becoming a suspect or a criminal, this would be a
welcome result.” In terms of the darts club example, even if just one person
matched the darts profile, this could be interesting.
The second way in which the use of small numbers as evidence for pol-
icy was discussed comes from policing. In detective practices, a difference
between 4 and 10 provides a ‘lead.’ It means that one has a starting point
for investigation. For instance, if the profile for pickpocketing is a tennis
player that reads glossy magazines, a detective can start investigating ten-
nis clubs.
At work here are ‘professional regimes of evidence’ (Kahn, 2013): profes-
sional standards for the type of evidence acceptable as a basis for interven-
tion. In this case, the regime of evidence for policy had not yet crystallized.
Juriaan, in his role as analytical expert, had final decision-making power
over the acceptability of the regimes of evidence that were applied. In other
words, he had the professional privilege (Goodwin, 1996) to decide on this,
as he stated in the final pilot project meeting:

BART: So what about interventions, as in the Tina example?

JURIAAN: Well, if we are talking about a neighbourhood of 300–800 peo-
ple, and maximum five per cent are registered suspects, should one
start a campaign?
(Fieldwork notes, September 26, 2011)

The question was put forward by Bart, a member of the innovation plat-
form that supported the pilot. In response to Juriaan’s question, Bart, Anna,
Liesbet, and Mieke all shook their heads.
Crime, Juriaan argued, is too serious a topic to flyer for if one can only
expect the numbers to range from 10 to 20 persons. Moreover, they could
 178 Francisca Grommé
not even be certain about the actual relation between the young suspects and
the lifestyle data, as explained. Juriaan furthermore rejected the detective
logic; in the case of the Burgcity pilot, the results were not reliable enough
to justify a visit to the darts club (Interview April 13, 2012).

Zooming Out
In the final meeting, Juriaan and his senior colleague Frank argued that
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

lifestyle analysis was not a feasible option. Nevertheless, Frank suggested

changing to a “helicopter perspective.” He advised acquiring data about
young suspects in other cities. Youth from similar neighbourhoods in dif-
ferent cities, in terms of income similarities, for instance, could be grouped
together and compared to other types of neighbourhoods. More variety
between neighbourhoods and larger numbers would benefit the analysis
and lead to statistically significant results: “to uncover the real processes on
a micro level, one needs more material for comparison” (Fieldwork notes,
September 26, 2011). Once a pattern was found for a general type of neigh-
bourhood, it could be applied to Molendistrict neighbourhoods. It follows
that to zoom in, they would first need to ‘zoom out.’
The reasoning here seems surprisingly similar to what Data Inc. and
Burgcity’s policy makers had earlier referred to as ‘general social science.’
With this, they referred to national statistics and criminological theory, as
well as knowledge of a more universal kind (not specifically tailored to the
neighbourhood). This was the type of knowledge that “should inform pol-
icy,” Frank argued.
Data Inc. seemed to have decided that the regime of evidence suitable for
policy should be based on large numbers and statistical significance. Ironi-
cally, this was exactly the type of analysis that Data Inc. had promised to
avoid at the outset of the pilot when it proposed formulating ‘local theories.’
Anna therefore argued against Data Inc.’s new suggestion, claiming that this
type of analysis lacks specificity. For instance, it overlooked the fact that
Molendistrict does not have sufficient facilities for youth between 12 and
18 years of age. With that, the pilot meetings came to an end, and Data Inc.
commenced writing the end report.

PROBLEM YOUTH

At the outset of the chapter, I proposed that by zooming in with data mining
one, does not obtain a more detailed view of the same object. Rather, zoom-
ing in brings a new object into being. In Burgcity, profiles of problem youth
were never stabilized, that is, accepted by all participants and included in the
end report as an outcome of the pilot study. Yet we observed several tenta-
tive profiles in this case study. In the first instance of zooming in described in
this chapter, for example, problem youth were evoked with crime statistics
 Data Mining ‘Problem Youth’ 179
and physical and spatial neighbourhood characteristics, such as the presence
of swimming pools. Later on, we learned how problem youth came into
being by assembling more general categories and aggregates, such as weak
neighbourhoods and city youth. When Data Inc. suggested zooming out,
problem youth were related to youth in similar neighbourhoods in other
cities; they became part of a national trend.
A particularly contentious issue is how corporate data and methods affect
government policies. The ways in which corporations categorize and dif-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ferentiate in order to increase profits may be at par with egalitarianism and

the provision of social justice as principles of government (Gandy, 2007). I
will further elaborate on the profiles of problem youth that were created in
Burgcity with this issue in mind. First, I discuss the use of pre-set Experian
lifestyle profiles. Next, I discuss the profiles in terms of their regimes of
evidence.
The Experian lifestyle profiles played an important role in this case study.
The application of consumer profiles was expected to reveal a more person-
alized and closer view of youth. Problem youths would be consumers, to be
identified by their media usage or their parents’ cars. Yet the profiles did not
always seem to easily fit into a crime-control environment, as is illustrated
by the following fieldwork fragment. A local police officer attended part of
an interactive meeting at Anna’s invitation. She suggested how the Experian
categories might be used:

POLICE OFFICER: Well, I think you can learn from this [the Experian
data]. Those “quiet radio listeners” will be very annoyed by kids
playing soccer outside, and they will surely call the city or the police
and say: “I can’t hear the radio because of the noise.”
LIESBET: I think you should see this as a marketing profile (. . .) It is a
characteristic of the neighbourhood, where residents use the radio
more often than the Internet.
(Fieldwork notes April 26, 2013)

‘Quiet radio listeners’ refers to an Experian profile. Here it was used to

learn about the types of people that file complaints with the police and the
city about youth. The short dialogue illustrates how marketing profiles can
be used to explain and predict behaviour. The profile’s name invited con-
notations about a certain type of person: quiet and peaceful (see also Curry,
2004).
The fragment also indicates that practitioners do not necessarily accept
the categories they are presented with. Liesbet argued that these profiles are
tailored to marketing usage, not to understand other types of behaviour—
let alone those of problem youth. Marketing variables and categories with
less obvious titles, however, may be more easily integrated into government
practice. In another meeting, for instance, Liesbet argued that persons in the
category of “two times average income and self-employed” complain more
 180 Francisca Grommé
often and insist that the government should solve their problem (Fieldwork
notes, April 21, 2013).
With regard to the regimes of evidence that were applied, we learn two
things about corporate intervention. The first is that Data Inc. approached
problem youth as an information problem, thereby justifying the collection
of more data (Schinkel, 2009). It argued for the collection of more data as it
aspired to a universal truth value in the guise of large numbers and statisti-
cal significance.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Second, we should nevertheless be careful in assuming that marketing meth-

ods simply ‘contaminate’ government practices. With regard to the regime of
evidence, Data Inc. decided that a more ‘traditional’ social science analysis on
the basis of more cases would be more suitable for local government. It was
therefore Data Inc. that reified the ethics of government intervention. In Data
Inc.’s view, a local government cannot deal lightly with the issue of youth
crime. Juriaan and Frank emphasized the city’s responsibility for careful and
effective action in the field of youth crime policy. Burgcity, in contrast, was
rethinking its own role in marketing terms, as a ‘communicative government.’

CONCLUSIONS

My interest in this chapter was in how a digital identity of problem youth

was created according to the rationale of zooming in. When brought into
practice, I argued, this metaphor suggests that digital representations of
youth have a high truth status. It therefore justifies the collection of ever
more data and the use of profiles. I set out to challenge zooming in as a data
mining metaphor by showing how it was done in practice and by drawing
out the norms that were embedded in and produced through this work.
The Burgcity case shows the limitations of the metaphor. It had one
obvious limitation: there simply were not enough registered suspects in a
neighbourhood to perform an analysis on. Yet this did not discourage the
analysts; Data Inc. argued that simply more data were needed.
I demonstrated, moreover, that data mining is not a practice based solely
on digital data. Far from a smooth and technical operation, data mining was
a situated practice. I identified two modes of situated improvisation: evo-
cation and comparison. These were conducted by the interplay of screens,
professional knowledges, paper maps, local politics, and regimes of evi-
dence. By relating these heterogeneous entities, one does not acquire a better
view of a smaller part of the same object, but a new object of intervention
is brought into being. In this case, problem youth shifted from relations
between categories of administrative everyday knowledge and objects in the
neighbourhood, such as swimming pools, to relations between youth from
comparable neighbourhoods in different cities.
We learn about several normativities that were part of zooming in. First,
results needed to be made relevant as surprises. Useful knowledge was
 Data Mining ‘Problem Youth’ 181
unexpected knowledge. Yet the results could not be too unexpected because
data mining results needed to be made tellable. Furthermore, surprise
depended on difference: Local problem youth needed to be distinguished from
the problem youth in the larger geographical area they are part of, such as a
city or district. This requirement also produced a norm: Local problem youth
is invariably constituted as a deviation and the city or the district as the norm.
Second, zooming in had a focus: Useful knowledge was knowledge at
the level of the neighbourhood. This had a consequence for problem youth:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

They were taken to “be where they lived” (Curry, 2004). Because crime
policy is focused on neighbourhoods in Burgcity, categories of good and bad
neighbourhoods already existed. These labels were easily transferred to the
young suspects living in them.
Third, as the previous remark also indicates: Zooming in depended on
the application of more general categories and aggregated data. Problem
youth identities only became more particular by assembling generalities. The
categories and knowledges that were applied were normatively laden them-
selves: They told of good and bad neighbourhoods, and they were mobilized
in relation to local policy discussions about facilities (as these might attract
youth from ‘weak neighbourhoods’). Furthermore, standardized profiles
were introduced from the domain of marketing, thereby equating problem
youth identities with consumer identities. In this pilot, however, these pro-
files did not stabilize. It also needs to be noted that the policy makers did not
accept every standardized profile.
Fourth, establishing a closer look depended on decisions about what
counts as good evidence for policy practice. Two regimes of evidence were
introduced into Burgcity’s policy practice: a detective regime and a market-
ing regime. The marketing regime was appealed most to because it related
to the idea of local government as a ‘communicative government.’ Data
Inc., however, decided that government intervention needs a more ‘scientific
basis.’ Intervention on the basis of small numbers would not be effective or
justifiable. In this case, a corporation reified what it thought of as the ethics
of government intervention.
If anything, these findings point out that improving the use of data min-
ing for a better view of problem youth involves not only decisions about
which digital data to use and how much. Producing knowledge about prob-
lem youth that can inform a fair policy practice will include rethinking the
issues discussed. At the very least, it involves rethinking the relation of policy
knowledge to social science and the corporate sector.
To conclude, a range of metaphors circulates digital identification prac-
tices such as data mining. Aside from zooming in, actors use ‘connecting
the dots’ (Amoore and De Goede, 2008), ‘deep knowledge,’ and ‘obscured
knowledge hidden in the data.’ These metaphors, I suggest, help perform the
seemingly endless analytical possibilities of these technologies. We need to
attend to them as situated practices in order to change the terms by which
digital identities are produced.
 182 Francisca Grommé
NOTES

1. My informants used ‘problem youth’ and ‘at-risk youth’ intertwiningly. To

avoid confusion, I will only refer to ‘problem youth.’ Although I will hence-
forth use the term without quotation marks, I emphasize that it is by no
means a natural category and that applying this category can negatively affect
individuals and groups (see Bowker and Star, 1999). Some of my informants
shared this view.
2. In Burgcity, Data Inc. applied a nearest-neighbour associative algorithm. This
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

algorithm is understood to be suitable for a wide variety of practical applica-

tions because it allows for the analysis of databases with many missing values.
3. Some would argue that this is not ‘real’ data mining. In using the term ‘data
mining,’ I adopt my informants’ terminology.
4. Surveillance Studies has had its own issues with the Panopticon, a model for
visual surveillance that Michel Foucault (1994 [1975]) adopted as a theoreti-
cal model (see for instance Lyon, 2006). In this chapter, I am less concerned
with theoretical models in social science and more with knowledge practices.
5. For this reason, I have chosen not to include the full titles of the policy docu-
ments that I cite in the bibliography. Instead, I mention the document types in
the text.
6. All fieldwork quotes have been translated from Dutch by the author.
7. The participants focused on finding patterns. They did not aim to introduce
an automated signal.
8. This process is referred to as the CRISP-DSM method; see Custers (2004).
9. The levels my informants worked with are, from high to low: city, (sub-city)
district, postal code, neighbourhood, sub-neighbourhood, postal code-6
(about 20 households), individual.
10. The interactive meetings could not be recorded due to the sensitive nature of
the data. The quotations from these meetings are based on notes that were
typed out within 48 hours of the meetings.
11. The original lifestyle data were aggregated to postal code-6 level, which means
this dataset provided information about the average characteristics of about
20 persons.

REFERENCES

ANG, R. P. & GOH, D. H. (2013) Predictive Juvenile Offending: Comparing Dif-

ferent Data Mining Methods. International Journal of Offender Therapy and
Comparative Criminology, 57(2). p. 191–207.
AMOORE, L. & DE GOEDE, M. (2008) Transactions After 9/11: The Banal Face
of the Preemptive Strike. Transactions of the Institute of British Geographers,
33(2). p. 173–185.
BOWKER, G. C. & STAR S. L. (1999) Sorting Things Out: Classification and Its
Consequences. London: MIT Press.
BOYD, D. & CRAWFORD, K. (2012) Critical Questions for Big Data. Information,
Communication & Society, 15(5). p. 662–679.
CHOY, T. K. (2005) Articulated Knowledges: Environmental Forms After Universal-
ity’s Demise. American Anthropologist, 107(1). p. 5–18.
CURRY, M. R. (2004) The Profiler’s Question and the Treacherous Traveler: Narratives
of Belonging in Commercial Aviation. Surveillance and Society, 1(4). p. 475–499.
CUSTERS, B. (2004) The Power of Knowledge. Ethical, Legal, and Technical
Aspects of Data Mining and Group Profiling in Epidemiology. Nijmegen: Wolf
Legal Publishers.
 Data Mining ‘Problem Youth’ 183
FAYYAD, U., PIATETSKY-SHAPIRO, G. & SMYTH, P. (1996) From Data Mining
to Knowledge Discovery in Databases. AI Magazine, 17(3). p. 37–54.
FOUCAULT, M. (1977) Discipline & Punish. The Birth of the Prison. New York:
Random House.
GAD, C. & LAURITSEN, P. (2009) Situated Surveillance: An Ethnographic Study of
Fisheries Inspection in Denmark. Surveillance and Society, 7(1). p. 49–57.
GANDY, O. H. (2007) Data Mining and Surveillance in the post–9/11 Environment.
In Hier, S. P. & Greenberg, J. (eds.) The Surveillance Studies Reader. Maidenhead
(UK): McGraw-Hill, 147–157.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

GOODWIN, C. (1995) Seeing in Depth. Social Studies of Science, 25(2). p. 237–274.

GOODWIN, C. (1996) Professional Vision. American Anthropologist, (3). p. 606–633.
HARAWAY, D. (1991) Situated Knowledges: The Science Question in Feminism
and the Privilege of Partial Perspective. In Haraway, D. Simians, Cyborgs, and
Women. The Reinvention of Nature. London: Free Association Books, 183–202.
HILDEBRANDT, M. (2008) Defining Profiling: A New Type of Knowledge? In
Hildebrandt, M. & Gutwirth, S. (eds.) Profiling the European Citizen: Cross-
Disciplinary Perspectives. Dordrecht: Springer, 17–46.
KAHN, J. (2013) Race in a Bottle: Law, Commerce and the Production of Race.
Lecture June 24, 2013. University of Amsterdam.
LYON, D. (2006) Theorizing Surveillance: The Panopticon and Beyond. Devon
(UK): Willan Publishing.
MARX, G. T. (2005) Seeing Hazily (but Not Darkly) Through the Lens: Some Recent
Empirical Studies of Surveillance Technologies. Law & Social Inquiry, 30(2).
p. 339–399.
M’CHAREK, A. (2005) The Human Genome Diversity Project. An Ethnography of
Scientific Practice. Cambridge: Cambridge University Press.
M’CHAREK, A. (2013) Beyond Fact or Fiction: On the Materiality of Race in Prac-
tice. Cultural Anthropology, 28(3). p. 420–442.
MOL, A. (2002) The Body Multiple: Ontology in Medical Practice. Durham (NC)
and London: Duke University Press.
PHILLIPS, D. & CURRY, M. R. (2003) Privacy and the Phenetic Urge: Geodemo-
graphics and the Changing Spatiality of Local Practice. In Lyon, D. (ed.) Sur-
veillance as Social Sorting. Privacy, Risk and Digital Discrimination. London:
Routledge, 31–56.
PLEACE, N. (2007) Workless People and Surveillant Mashups: Social Policy and Data
Sharing in the UK. Information, Communication & Society, 10(6). p. 943–960.
RUPPERT, E. (2011) Population Objects: Interpassive Subjects. Sociology, 45(2).
p. 218–233.
SCHINKEL, W. (2009) De Nieuwe Preventie. Actuariële Archiefsystemen en de
Nieuwe Technologie van de Veiligheid. Krisis, (2). p. 1–21.
STRATHERN, M. (2005) Partial Connections. Updated Edition. Walnut Creek
(CA): Altamira Press.
VALVERDE, M. (2003) Law's Dream of a Common Knowledge. Princeton, Princeton
University Press, 2003.
WITTEN, I. H., EIBE, F. & HALL, M. A. (2011) Data Mining: Practical Machine
Learning Tools and Techniques. Amsterdam: Morgan Kaufmann Publishers.
YANOW, D. (2002) Constructing “Race” and “Ethnicity” in America: Category-
Making in Public Policy and Administration. Armonk (NY): M. E. Sharpe.
ZIEWITZ, M. (2011) How to Think About an Algorithm: Notes from a Not-Quite
Random Walk. Draft paper, September 29. Available from http://ziewitz.org/
papers/ziewitz_algorithm.pdf. [Accessed 10–01–2014]
 9 Sorting (Out) Youth
Transformations in Police Practices
of Classification and (Social Media)
Monitoring of ‘Youth Groups’
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Vlad Niculescu-Dinca, Irma van der Ploeg,

and Tsjalling Swierstra

INTRODUCTION

This chapter highlights a set of methodological and normative issues related

to technological developments and policy transformations in contemporary
policing. It draws from data in the Netherlands, where the prioritization of
the issue of ‘problematic youth groups’ on the government agenda entailed
a proactive crackdown approach to sort out the problems generated by these
youth.1
The crackdown approach uses a classification method to categorize these
groups into ‘criminal,’ ‘nuisance,’ ‘annoying,’ and ‘acceptable’ depending on
the gravity of their offences (Bureau Beke, 2010). The so-called ‘shortlist-
ing method’ uses questionnaires filled in by field officers, who periodically
assess group composition, activities, and behaviour.2 Using this classifica-
tion, police map the activities of groups in the first three categories with geo-
graphic information systems in order to maintain a view on the phenomena.
In addition to geographic mapping, the proactive crackdown approach
promotes early signalling and the exchange of information about groups
and their members among the multiple partners involved in interventions
(e.g. police, municipalities, public prosecutor, child protection agencies).
The information used in these practices is gathered from multiple databases
of partners, as well as from Internet/social media monitoring of these youth.
In 2013, the approach was evaluated as generally being on a good track
(Van Montfoort, Office Alpha and Dutch Youth Institute, 2013). Yearly
statistics of classified groups show a drop in numbers of all categories at
national level since the government introduced the approach in 2010/early
2011 (Bureau Beke, 2013). Nevertheless, the approach is set to continue. It
now promotes extra emphasis on ‘quick and effective crackdown’ on the
less problematic ‘annoying’ and ‘nuisance’ groups to prevent them ‘slipping
down to the status of a criminal group’ (Van Montfoort, Office Alpha and
Dutch Youth Institute, 2013).
According to the shortlisting method, the phrase ‘annoying youth groups’
refers to youth that hang around the neighbourhood and are occasionally
 Sorting (Out) Youth 185
noisy or have incidental small arguments, which are usually finished quickly.
At the same time, they form the largest category of ‘problematic youth groups’
(e.g. in 2012 they formed 731 out of all 976 shortlisted youth groups). Given
that the crackdown approach is set to focus more on these much larger cat-
egories of groups, bigger numbers of youth enter the scope of proactive sur-
veillance. As the monitoring and intervention processes are influenced by the
way in which groups are classified and counted, this chapter asks how the
police classify, map, and monitor cases of problematic youth groups and what
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

normative issues emerge in these technologically mediated policing practices.

The chapter draws on ethnographic research performed throughout the
year 2012 at a Dutch police station, which I name in this document Dutch
Urban Police (DUP).3 There I was granted access to interview and observe
officers involved in mapping youth groups of concern with geographical
information systems and in monitoring them with Internet/social media
monitoring technologies (SMM). Additionally, the chapter analyses an inter-
view with one of the main designers of a dedicated police solution for Inter-
net and social media monitoring. The solution enables Dutch police forces
to collect and analyse large amounts of Internet information in an automatic
way. The following gives an illustration of these practices and technologies.

Proactive Policing at Work

Working in front of her computer screen, agent Anna is responsible for
monitoring ‘problematic youth groups’ on the Internet and social media (i.e.
Twitter, Facebook, etc.):

“The main objective of Internet surveillance is being [there] before the

criminal offense happens. You don’t have to have criminal offenses to
use the Internet as an information source,” she tells me when I ask about
her tasks.

In between describing her practices, she comments on the very method

used by the police to shortlist groups:

“We’re thinking that maybe we are running behind [with the short-
listing method]: In some situations we are saying: ‘we cannot find our
youth on the street. The groups are not there anymore. How is that
possible?’ And then she quickly answers herself: “There is much more
contact on the Internet. They play a Playstation game, all that kind of
stuff, online. They play against each other from their own homes. And
they also make more contacts [with other groups]. That’s what I want
to show: the groups are mixing . . .”

I would later find out (see below) that this insight is shared by other police
forces. What seemed to be the view of one officer appears to be spread among
 186 Vlad Niculescu-Dinca et al.
more officers and more police forces with implications for their approach
to the phenomenon.
For Officer Bart, the Internet strategist of the police force, not seeing the
groups on the streets is not a reason to let them pass without surveillance
but to put more effort into monitoring them on social media. Even if they
are off the street, he thinks that they merely changed their way of setting up
whatever they used to do:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

“That’s what you get on the Internet. You do not see that many youths
hanging out on the streets anymore. Normally [the problematic activ-
ity] was brewing: okay, they sit at the bench, talk, and then do it! And
now there’s nobody at the bench. But they’re still making contact, trying
to communicate [through social media]. And that’s the different thing.
Normally, we used to know the group by seeing them: okay, there they
are. But now, where are they?”

I remembered Officer Bart’s challenges when, later that year, I interviewed

Officer Cees, from the Dutch National Police Services Agency. He is one of
the main designers of a dedicated police solution for Internet/SMM. The
solution aims to enable Dutch police forces (and other governmental agen-
cies) to automatically gather, store, and analyse information from the Inter-
net, including social media. He explains that the technology differs from
other (commercial) solutions that monitor all social media activity:

“You can’t get a big net on the Internet, 24 hours a day . . . You will
violate privacy; you will violate all legal restrictions. I strongly believe
you shouldn’t do that. It should always be case related.”

He translates his views about how to account for ‘privacy violation’ by

only enabling monitoring when it is related to cases (as opposed to moni-
toring all social media activity). These restrictions (such as who can access
which features of the system, which information, how many times, etc.) are
to be delegated to the monitoring technology. In his view, this is the way to
avoid undesired consequences:

“[The solution] lets user A only look at public Facebook profiles and
lets user B, because he has a lot more possibilities from a legal point
of view, get behind the front door of Facebook. This should always be
clear: when does the technology decide to do that and how does the
technology do that; otherwise it’s very scary stuff!”

Together with the views of Anna, Bart, and others and with observa-
tions about officers working on cases of ‘problematic youth groups,’ I take
a closer look at the hidden normativities of these technologically mediated
policing practices.
 Sorting (Out) Youth 187
First I provide a more elaborate context, giving an account of policing
‘problematic youth groups’ in the Netherlands and of the evaluation report
of the crackdown approach. The next section outlines the analytical stance
of the chapter. It shows the ways in which actor-network theory (Akrich and
Latour, 1992; Bowker and Star, 1999; Law, 2009) is particularly relevant for
understanding how youth group classifications, their statistics, problematic
character, suspicious behaviour and the very notion of ‘youth group’ are
constructed or performed within technologically mediated practices. The
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

subsequent two sections analyse police practices promoted by the crack-

down approach, illustrating the ways in which youth groups are classified,
mapped, and monitored. The analysis shows that depending on how youth
are defined as ‘cases,’ how they are performed as ‘problematic,’ and/or how
they are categorized, wider or more restricted data gathering is legitimated.
The last section links these analyses and discusses the pivotal role of the
notion of ‘case’ and its normative implications for proactive surveillance.
While the notion of case is traditionally associated to reactive policing and
to aspects such as committed crime, case (re)definition, and case closing, the
chapter shows that these features are significantly transformed when they
are translated in software code and employed in proactive policing practices.
In conclusion, the chapter argues that a particular alignment of technologi-
cal, policy, and organizational factors engenders the translation of the issue
of ‘problematic youth groups’ into an information problem, fostering auto-
matic monitoring of a growing number of youth.

THE ISSUE OF ‘PROBLEMATIC YOUTH GROUPS’

For decades, the issue of ‘problematic youth groups’ has been addressed by
the police in the Netherlands. These youth have been a major source of irri-
tation for many Dutch citizens due to their loud noise, street nuisance, graf-
fiti, disrespectful and intimidating behaviour, and criminal activities. The
terminology and definitions, however, changed over the years. Terms such
as ‘jeugdbendes’ or ‘gangs’ were sometimes used but also avoided. When
avoided, they were criticized for not capturing the variety of behaviour and
for not being adequate for the Dutch situation, bringing too much reference
to stereotypes of American gangs (Frank van Gemert, 2012). Over time, the
issue of definitions had consequences for policies, influencing the view on
the statistics and approach to the phenomenon. For instance, Spergel writes:
“Definitions determine whether we have a large, small, or even no problem,
whether more or fewer gangs and gang members exist, and which agencies
will receive funds” (Spergel, 1990, 177, quoted in Frank van Gemert, 2012).
In the past decade, the more neutral definition of ‘problematic youth groups’
has been adopted in official reports (Bureau Beke, 2009), and since 2009,
all regional police forces in the Netherlands are required to centralize the
statistics from their area.
 188 Vlad Niculescu-Dinca et al.
The shortlisting method enables the police to have a more fine-grained
characterization and classification of youth groups, depending on their
street-related behaviour and activities.4 Besides their type, groups are
characterized by the location(s) where they are seen by patrolling officers
(who usually give the group a name based on this location), composition
(size, ethnicity, age range), daily activities, risky habits (alcohol and drug
use), and recent criminal behaviour. This information is fed into forms
by the officers on the beat. To shortlist a group, the method specifies that
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

there must be repeated nuisance on the street, caused by more than two
members, and the groups must be more than ‘virtual groups, active on
the Internet’ (Bureau Beke, 2009, 18). New members or youth that are
seen together with the group are registered as well upon encounters with
the officers on the beat. Once sorted, the police begin registering their
activities with geospatial technologies in order to maintain a view on the
phenomenon.

Approaching ‘Problematic Youth Groups’

Early in 2011, the Dutch Ministry of Security and Justice put the issue of
‘problematic youth groups’ high on the agenda. It was argued that they
needed a dedicated focus with an intensified crackdown approach. The
‘seven-step approach’ (Bureau Beke, 2010) goes beyond mapping and clas-
sification to involve comprehensive interventions at group level but also at
individual level. Depending on information on each group member and also
on how a group is classified, cases are handled in particular ways and with
specific partners. A broad set of partners from the criminal justice chain and
child protection chain are involved (municipalities, public prosecutor, child
protection agencies, and the police). For example, for the ‘annoying’ and
‘nuisance’ groups, the lead role in interventions is taken by municipalities
and in cases of ‘criminal’ groups by the public prosecutor. Depending on the
analysis and information gathered, a juvenile offender in a group may face
supervision and social reintegration measures but also investigation, pros-
ecution, trial, and sentence execution (Borst, 2013).
An important step in the approach is building a common understanding
between the involved partners in order to get a clear view of the group com-
position, relations, and activities. Partners are required to consult and agree
on which measures need to be taken in each case. To inform these decisions,
they share information between them and use information brought forth by
the police. The kind of information varies widely from details on street activ-
ity to the inner motives and hopes of youth. For example—as can be found in
a police poster supporting practitioners with the implementation—officers
need to ‘get the most complete picture possible on the person’ guided by a
wide range of questions such as: ‘What motivates a person? What are his/her
dreams/talents/competencies? Is X active on the Internet? Digital identity/
profile? Does he/she have multiple identities on the Internet?, etc.’ Partners
 Sorting (Out) Youth 189
are required to look at information from geographic information systems,
police databases, partner databases, the Internet, and social media (Bureau
Beke, 2010, 9). Therefore, dedicated police officers have been assigned also
to the task of monitoring them on Internet and social media.

Evaluation Report of the Approach to ‘Criminal Youth Groups’

At the beginning of 2013, a report commissioned by the Ministry of Security
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and Justice (Van Montfoort, Office Alpha and Dutch Youth Institute, 2013)
evaluated the crackdown approach on criminal youth groups. It concludes
that the multi-level approach is generally on a good track. Drawing from the
annual statistics (Bureau Beke, 2013), it reports a decrease in the number of
youth groups that were listed as problematic.
As Figure 9.1 indicates, the report shows a drop in the number of cases
in all categories. Split by category, this is a drop from 92 to 59 ‘criminal’
groups, from 327 to 186 ‘nuisance’ groups, and from 1341 to 731 ‘annoy-
ing’ groups. Despite the drop, the report recommends further measures to

2000

1800

1600

1400

1200

1000

800

600

400

200

0
2009 2010 2011 2012

Annoying Nuisance Giving Criminal

Figure 9.1 Evolution of problematic youth groups in the Netherlands by number

and type
Source: Bureau Beke, 2013. Problematic youth groups in Netherlands, extent and nature in
autumn 2012. Arnhem: Ministry of Security and Justice.
 190 Vlad Niculescu-Dinca et al.
be taken concerning ‘annoying’ and ‘nuisance’ groups to prevent them from
‘slipping down to the status of a criminal group’ (Van Montfoort, Office
Alpha and Dutch Youth Institute, 2013, 12), because ‘criminal’ groups are
harder to dismantle than ‘annoying’ and ‘nuisance’ groups. These figures
also show that the ‘annoying’ and ‘nuisance’ groups remain the largest cat-
egories at the national level. With group sizes varying widely between 10
and 100 members (Van Montfoort, Office Alpha and Dutch Youth Institute,
2013, 5), this is a total number ranging between 10,000 and 100,000 youth
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in these two categories in 2012.

The report also identifies challenges faced by the approach. It notes a
lack of resources and personnel as a main challenge, affecting the capacity
of the police to keep up with the dynamics of the phenomenon. Moreover, it
notes the lack of a shared sense of urgency among the partners in prioritiz-
ing the approach, leading sometimes to late actions. The evaluation notes
also that “a growth or drop in numbers of shortlisted problematic youth
groups does not necessarily imply that the actual number of groups has
grown or dropped” (Van Montfoort, Office Alpha and Dutch Youth Insti-
tute, 2013, 3). Analysing possible causes of inaccuracies, the report identifies
the lack of expertise of police officers working with the shortlisting method
or the “increased police attention [which] may, for example, well result in
more groups being identified” (Van Montfoort, Office Alpha and Dutch
Youth Institute, 2013).
Although the evaluation report raised many special points of attention,
it did not have as a focus to evaluate the role of the technologies used
in the approach: in the classification and geographical mapping; in the
information-gathering process, in the Internet and social media monitor-
ing process. Technologies are largely taken for granted or mentioned very
briefly. For example, although the report notes possible inaccuracies in sta-
tistics, it does not account for the way in which these numbers are locally
produced, how the groups that get counted are created, visualized, and
classified in a category or another, and what role technologies play in this
process. Another example is the way in which social media are mentioned
(once), only as one of the factors contributing to the formation of criminal
youth groups:

In the forming, disappearing and transforming of criminal youth groups,

three factors or processes seem to play a role: [. . .] (3) spread of ‘gang
culture’ by (social) media.
(Van Montfoort, Office Alpha and Dutch
Youth Institute, 2013, 10)

Although the approach includes a step for information-gathering prac-

tices, including collection from Internet and social media, the evaluation
addresses their role neither in generating statistics nor in the approach
itself.
 Sorting (Out) Youth 191
ANALYTICAL STANCE

Though a large body of policing literature seems to either take technologies

for granted or render them as tools in the hands of police officers—that is, it
helps them do their job quicker and better as for example in Chu (2001) and
Williams and Williams (2008)—a growing number of studies have begun
to pay attention to the more substantive roles technologies play in policing.
These analyses began to look at technologies as more than just instrumental
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in relation to policing models, codes, and practices (Chan, 2001; Manning,

2008; De Pauw et al., 2011; Van Brakel and De Hert, 2011; Leman-Langlois,
2012) and sought to analyse the ways in which these technologies become
active ‘agents’ in policing processes.
The idea that technologies are not merely neutral tools but actively medi-
ate human action has been developed already for a few decades in science
and technologies studies and in particular in actor-network theory (ANT;
e.g. Callon, 1991; Akrich and Latour, 1992; Latour, 1994; Law and Mol,
2001). In analysing the agency of technological artefacts to the same extent
as that of human beings, ANT aims to avoid reducing technologies to neu-
tral tools or to clear-cut results of particular social practices without, at the
same time, viewing them as a determinant of social relations (whether for
better or worse). From an ANT perspective, actors, defined as entities that
do things, should be conceived in relations with other actors, for it is in these
networks of relations and associations that they are enacted in specific ways,
becoming what they are and doing what they do.
Therefore, in order to understand the active role of technologies in polic-
ing and their outcomes, we need to look into these networks of actors and
relations. This implies analysing the entities that emerge from these net-
works (e.g. classifications, suspicion, groups, etc.) and which are influenced
by technology. It follows that normativity of technology would also have to
be located within this interplay of (f)actors that constitute the socio-technical
arrangements. For example, it means going beyond discursive claims about
neutrality of technology to potentially disclose issues of normative relevance
embedded in technology design or analysing laws, policies, and value state-
ments and the way they are implemented in technology (e.g. the value of
privacy translated and built in technologies).
A particular set of technologies not only pervading policing but penetrat-
ing a whole array of processes of human life are systems of classification.
Sometimes explicit and mostly invisible, classifications are powerful arte-
facts ordering material and social realms alike (Bowker and Star, 1999, 4).
We can feel their force instantly if we would try, for instance, to ignore
the signposts at public toilets indicating male/female, at borders indicating
types of passports, or at supermarkets indicating cash/PIN only, and so on.
Already implicit in these examples is that classifications can be built into
infrastructures and procedures. They often embody highly normative charge
and shape people’s identities (Bowker and Star, 1999, 4).
 192 Vlad Niculescu-Dinca et al.
In order to understand how they are made and how they are often ren-
dered invisible, an exploration of these bureaucratic and technological infra-
structures is necessary (Bowker and Star, 1999, 5). The approach expands
on Foucault’s argument that an archaeological approach is necessary to dig
up the origins and consequences of the categories with which our lives are
imbued. In our case of policing technologies, I draw from this approach to
analyse classification practices and technologies, entailing stronger or lighter
monitoring and intervention.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Classifications not only organize modern life but also lie at the base of
many statistics and surveys of various kinds. Things and people alike are
often categorized and counted at the same time. Statistics based on various
classifications also have normative consequences when taken up in policy
making and translated into action as descriptions of reality. Building also
on Foucault’s practical archaeology, John Law (2009) argues for a perfor-
mative understanding of social science methods. That is to say, rather than
merely describing reality, the methods through which statistical knowledge
is produced also enact these realities into being. Instead of only reporting
on phenomena, the methods make them into ‘realities.’ For instance, in our
case, the police are the ones filling in the forms that entail sorting youth
groups into one category or another and, at the same time, they are the ones
responsible for sorting out problematic behaviour. Their classifications and
conceptualizations produce in a way the phenomena, while the aggregation
of numbers about these classified entities performs them at that level.

PRACTICES OF CLASSIFICATION AND MAPPING

Therefore, this section looks more closely to these classification and mapping
practices that precede monitoring. It illustrates the various translations occur-
ring between ‘groups hanging out on the street’ and the entities they work
with on their computer screens, which become the focus of policy making and
intervention. The section draws from interviews and participant observation
sessions with two officers—renamed Dirk and Erwin—in their daily routines.
In policing problematic youth groups, the Dutch Urban Police (DUP)
use a geographic information system to map the incidents in their area. The
system draws from data registered and stored in police databases. Amongst
other possibilities, it allows the registration and visualisation of the groups
and the individual members related to them, representing these groups
with colour-coded icons that correspond with the gravity of their offences
(‘annoying’ with grey, ‘nuisance’ with orange, and ‘criminal’ with red).
When asked about his tasks, officer Dirk explains how youth groups are
represented in their systems and how this helps him get a clear view of the
phenomenon.

“So when I go to the group [Officer Dirk points to an icon on the screen],
what it shows is the location, the last location known, that the officers
 Sorting (Out) Youth 193
on the beat put in. I know that mostly they hang out there. [. . .] So the
data they put in is shown here: the name of the group, the sort of group,
whether they are criminal or annoying or nuisance.”

Upon selecting a period of interest, the geographical information system

(GIS) displays a map with icons representing events with listed groups. Each
icon of a group can be clicked to open a list of members that were registered
as belonging to that group at each particular encounter with the police. The
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

maps can also display hotspots of various crimes, enabling analysts to form
a picture of the phenomenon and potentially relate it to a group’s activity.

“After a while we also identified the group here [Officer Dirk points to
an area on the map designating another department], in the neighbour-
hood of some bars and we saw a pattern of vandalism from here to here
in the same time that they normally gathered here. So, by identifying the
group here and here and knowing the timestamps, you could also iden-
tify them as the probable perpetrators of this vandalism trail . . . With-
out any analyses, we saw that it could probably be the group responsible
for that trail.”

Working in front of his screen when drafting his analysis, Officer Dirk
needs to rely on the information put in by field officers, ‘always bound to
a location,’ and on maps produced by the system. What he visualizes are
the coded entities on his screen. Their mapping is contingent upon a set of
factors that need to be aligned in order for the system to make these groups
visible. This depends on officers on the beat, who translate whatever they
see on the street into program fields; the officers need to have ‘a group,’ ren-
dered as a stable entity; they need to translate the place they saw the group
into the location marked in the system as the field designating the place
they usually hang out. Highlighting this process shows how the phenomena
in the street are translated into other, software-enabled entities visible on
computer screens.
As it turns out, this process faces a set of challenges that give a good
glimpse into the technologically mediated character of this work. The fol-
lowing quote is taken from a participant observation session with Officer
Erwin, who needed to do an analysis of the situation of problematic youth
groups in his area. He was about to issue a report showing a spike in ‘nui-
sance,’ based on generated statistics. After I questioned him on the probable
reasons for the spike, Officer Erwin searched and read the database registra-
tions to try to understand. He soon needed to discard the report after find-
ing that many registrations were categorised as ‘nuisance’ when other codes
were more fitting, designating lower offences:

“It’s good that we looked into that. Now we know. [. . .] They are all
registered as ‘melding overlast jeugd’ (youth nuisance). It’s the same.
You see? He went without a valid ticket [in public transportation].
 194 Vlad Niculescu-Dinca et al.
The problem with this kind of thing . . . I bet it’s the same [the officer
checks another set of registrations]. It’s the same. So that’s really a big
issue there. So then in fact you could say: ‘do we have a problem with
“nuisance” here?’ No, I don’t think so. Because all the registrations are
about the tickets, that they don’t buy a ticket. I just want to check the
other ones for 2011. This is another one: ‘Young guy was sitting with
his feet on the opposite chair.’ Not allowed. So they have to find another
incident number to put this in the system, because this is really screw-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ing up the numbers like crazy. [. . .] So this really changes our view on
nuisance. Same. [. . .] Oh my God. This means in fact that I can throw
the whole analysis down the drain.”

The quote does not merely highlight that there are challenges in working
with police systems but illustrates how systems enact certain realities. In this
case, the report showing a significant increase of ‘nuisance’ in that area clas-
sified ‘people that go through the gate without a valid ticket’ as ‘nuisance,’
which designates a higher level of criminality than would be appropriate for
free riding. As Officer Erwin mentions, this proved to be a repeated clas-
sification problem and consequently a counting issue, which affected their
view on the phenomenon, performing it in a category that entails the need
for more intense monitoring.
These situations point out some of the contingencies involved in the socio-
technical processes in the police, through which ‘groups’ are classified and
mapped. Without inferring from this reduced set of examples that national
statistics are affected, the section does remind us that what we count are
software-enabled entities, prone to partiality and ambiguity (Gerson and
Star, 1986). And it is these entities that lie at the basis of intervention and
monitoring.

PRACTICES OF SOCIAL MEDIA MONITORING

In fact, the Internet monitoring procedure in the DUP force specifies that
‘having a group’ in the GIS is the starting point. Agent Anna, whom we met
in the beginning of this chapter, explains the necessary condition to begin
her monitoring tasks. She has to

“always have the group, like this [she points to an icon of a group on
the screen], always start with a group that’s already in the system.”
Being practiced within a proactive approach, there is no need “to have
criminal offenses, but use it as an information source.”

‘The group’ defined as a case becomes an obligatory passage point in their

working processes (Latour, 1987): necessary to begin monitoring but also
sufficient, in the sense of legitimizing further data gathering.
 Sorting (Out) Youth 195
While the monitoring of youth on social media has been made legal by the
Ministry of Security and Justice, its legitimization builds on the idea that a
form of displacement is at work. That is, an assumption is made that group
members continue or increase their problematic behaviour as they spend
time on the Internet and social media, even if they keep a low profile on the
streets. The idea is known in policing as ‘crime displacement’ or—in infor-
mal police circles—as ‘waterbed effect’ or ‘balloon effect.’ This hypothesis
implies that crime tends to shift or move, either upon police actions or due
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

to changes in the societal structure, but remains the same in volume. The
notion makes officers more prone to translate the absence of criminal activi-
ties into their shifting to other places, times, tactics, or types.5 The quote of
Officer Bart at the beginning of the chapter is indicative of the way in which
the displacement hypothesis is sometimes taken up at an operative level. Not
encountering a certain phenomenon (in this case youth hanging out at the
street corner) leads the officer to use displacement for explaining the change.
The quote performs youth groups as still ‘out there’ setting up their activi-
ties, which continue to be of a criminal nature. Thus the underlying assump-
tion of this theorization of crime dynamics is that crime among these youth
remains a constant. In this way, it becomes an argument to increase the
distribution of resources towards monitoring these youth on social media.
Once engaging with this practice, officers look at individual group mem-
bers to interpret various indicators specific to social media in order to infer
suspect behaviour. For instance, in the following quote, Agent Anna reads
several Twitter messages, pictures, and accounts leading one to the other. The
situation is about a picture of a weapon held by a boy, who reportedly used
the picture to threaten his girlfriend. The officer traced the picture through
various social media accounts to find more pictures. Given their content, the
police decided to bring the young boy to the station for questioning:

“I follow one group on Twitter—and then there was this message about
a boy being arrested for threatening his ex-girlfriend with a weapon. It
was retweeted many times. [. . .] I found it looking into his followers on
Twitter. And then somebody had this picture of a weapon [the officer
points to a picture with a boy holding a gun]. I went to his Twitter and
I found more pictures with weapons. So, on his account I found these
pictures. I called the department and said ‘Look what I found. We have
to interrogate him again because of what I found here.’ And we should
confront him with that because he was this little boy crying and he’s
sorry and didn’t know that it was wrong to have it. But if you take pic-
tures like this, you’re very well aware of that.”
Interviewer: So he’s dangerous now?
“Well, he’s not dangerous, but . . . he deserves the attention, for sure.”

This example shows how officers engaging in SMM not only monitor
youth groups for street-related offences but assign suspicion to new kinds
 196 Vlad Niculescu-Dinca et al.
of behaviour or entities. Besides looking for the risky activities specified
by the shortlisting methodology (e.g. graffiti, alcohol and drug use, street
nuisance), officers now need to interpret pictures, social media relations,
or status updates. Depending on the prior knowledge they may have, they
infer suspect behaviour from new entities with their specific normativity:
too-high numbers of ‘followers’ for the declared age, too many messages on
a topic, the clothing in a profile picture, and so forth. This kind of behaviour
is not necessarily problematic according to the shortlisting method, based on
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

which the group was included for monitoring in the first place. Still, being
related to a case legitimizes monitoring.
However, in this research, it became apparent that youth don’t always
continue their problematic behaviour. They also change their behaviour, and
spending time on social media is part of this shift. They engage in gaming,
picture sharing, sending messages, status updates, and chatting instead of
spending time on the street. Even if these activities can also generate prob-
lematic behaviour, they are not necessarily suspect in the sense defined by the
shortlisting methodology. This is also attested by Agent Anna:

“Because they are not out on the streets—or less so—makes their way
of interacting different.”
“[. . .] They are not criminal groups”
“They put everything on here. ‘Eating soup.’ ‘In a moment there will
be the match,’ they put it in there. ‘Finished with internship.’ This one’s
obviously sick, because he says: ‘being sick is the most awful thing,’ but
then with a bad word, there it is. They say they buy things, so it gives
me information about what they do.”

This is also signalled by reports of other police forces:6 According to a

Dutch portal on youth and safety, the police in the city of Arnhem noticed
a sharp drop in nuisance, and they link this with the increasing time spent
by youth on social media and online gaming. These observations are also
in line with several critiques of the crime displacement hypothesis. Multiple
studies show that, upon successful prevention programs, displacement does
not necessarily occur (Clarke and Weisburd, 1994; McLennan and Whit-
worth, 2008). Others, such as Barr and Pease (1990), argue that the notion
is inadequate in the first place, as not all crime prevention results in crime
being displaced. They propose the notion of ‘deflection,’ which encapsulates
the idea that crime can also be prevented even if it may translate in other
forms of crime.
Thus, the police shortlisting method needs to account more dynamically
for changes in youth behaviour. Otherwise, once enacted as a problematic
group, the youth have a hard time invalidating the conditions for which they
were placed under monitoring, even when the reasons for which they were
selected in the first place are not present anymore.
 Sorting (Out) Youth 197
New Forms of Groups
Social media monitoring practices challenge not only the ways in which
groups are assessed as problematic but the very police notion of youth group.
As the quote from Agent Anna, in the very beginning of this chapter, suggests,
police officers doing SMM begin challenging the shortlisting methodology. In
the following quote, Agent Anna questions the size of a group, as registered in
the system, and does that by checking its associated members on social media:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

“When they were in a group they usually stayed for a longer time, the
changes were not that big. But now, if I check Twitter, for example, I
look at the Keizerstraat group, 35 people are in [the system] as contacts
of each other. I say: ‘I do not believe that they are really contacts in
the sense of friends.’ So I will go and search for those 35 people on the
Internet and see if I can find whether they are really having contact
through the Internet and whether I can say, because of the messages
they are sending, that they are friends. I dare to say that there are up
to a maximum—and it’s a real maximum—10 people of that group are
really in contact with each other.”

Besides the different group size, the following quotes illustrate that groups
are not well defined and stable but mix and merge. Although the procedure
needs to start from ‘a group,’ officers find connections with other shortlisted
groups.

“They often change more in the way they are constructed as a group.”
“[. . .] Then I saw the retweet and the retweet led me to him. And he
is not part of that group but he belongs to a different group.”
“[. . .] I believe that this group is . . . [officer stops to explain] you have
‘hinderlijk’ (‘annoying’) and overlast (‘nuisance’) . . . it’s one of these
two. The group I follow is called Koningstraat, that’s this one [officer
points to an icon on a screen]. And they are members of Hoogstraat.”
“[. . .] So the groups we had in the past, I think it’s changing. And we
cannot say in a short amount of time that this is that group, depending
on that location. They are less location dependent and the loyalty is dif-
ferent because of the contact through the Internet.”
“[. . .] I think they are dissolving in a way that they are not that loyal
to each other as they were when they were on the streets.”

These quotes perform a much more fluid notion of group compared to

the one produced by the shortlisting methodology, which treats the group
as much more fixated through name, location, or category. These ‘dissolv-
ing’ relations of youth in online environments have been also highlighted
by other analyses of surveillance of young people on the Internet (Steeves,
 198 Vlad Niculescu-Dinca et al.
2012). These studies suggest that not only do Internet surveillance practices
perform youth differently, but youth change their online behaviour over
time, in part due to awareness of police monitoring.
In both situations, these new relations challenge the definition of groups and
their classification. From the above quotes, we can see the easiness with which
‘annoying’ and ‘nuisance’ are interchanged by the officer. A big set of group
identity attributes—as defined by the shortlisting method—are performed dif-
ferently by social media monitoring practices. The size of the group, its compo-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

sition, kinds of relations, their stability or their relation to location are enacted
differently, indicating significant transformations of the police notion of group.
This has implications for the aggregated statistics that are based on these
groups. More mingled relations may mean fewer groups. Thus, the under-
representation of social media relations in the shortlisting method leaves
room for misunderstandings when reading numbers in statistics. This is
especially relevant for the higher number of cases of ‘annoying’ groups. As
the evaluation report shows (Van Montfoort, Office Alpha and Dutch Youth
Institute, 2013), they remain the largest but at the same time the least prob-
lematic category. Their characterization, according to the method, can easily
encompass large numbers of youth that ‘hang around the neighbourhood
and are occasionally noisy or have incidental small arguments, which are
usually finished quickly.’ Their reinforced enactment as ‘problematic’ entails
the need for their further monitoring. In its turn, this practice is gaining
significant breadth and depth when amplified by the technological develop-
ments that afford increasing automation of this process.

AUTOMATING SMM

As mentioned by Officer Bart, the police engage with real-time monitoring

afforded by the Internet and social media monitoring technologies. Depend-
ing on the task at hand, practitioners see them as powerful tools to automate
the labour-intensive, time-consuming processes of information gathering.
These solutions allow officers to program automatic alerts on certain events,
crises, persons, or groups and to analyse the data even after it has been
removed from the servers of the social media provider.
Officer Bart describes such a technological solution for social media mon-
itoring that was purchased by their police force from a commercial vendor.7
The solution is able to automatically monitor a broad set of social media
providers, Internet forums, and websites:

“It keeps everything. They [the social media monitoring provider] have
380.000 websites in the Netherlands. And that’s a lot. For example, if
there is an author that’s very interesting, I want to make a search for
everything the author said. [Or I can ask the program to] ‘give me a
signal whenever he’s talking ‘house party.’”
 Sorting (Out) Youth 199
With such a large amount of Internet resources and power to automate
the searching and analysis, the solution is interesting for the police. But the
license to use it, Officer Bart explains, is expensive for their local police
force, and it also poses a security problem, as they need to send their queries
of interest outside the police domain, to the servers of the provider.
In part for these reasons, the Dutch police chose to develop an Inter-
net/SMM solution themselves with similar functionality. Initially aiming
for investigative purposes, the scope of the solution expanded to enable
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

more generic research by a larger set of governmental agencies and police

departments. The solution is set to allow gathering of information, stor-
age, time stamping, retrieval for forensic research, and presentation of
information in a Google-like interface. Nevertheless, as Officer Cees men-
tions, the same features that make this solution interesting for the police
make it prone to trigger privacy infringements of various kinds, as also
pointed out in several reports (Oerlemans and Koops, 2012; Bert-Jaap
Koops et al., 2012).
In order to account for ‘privacy violations,’ Officer Cees explains that the
solution is built with the idea of ‘privacy by design’ from the start.8 That
is, in his view, the system will be programmed such that it can be used only
in relation to specific cases, and only by the officers that have the proper
authorizations:

“So if we get your personal information from your Facebook page and
we have no case or there’s no legal reason for us to keep this informa-
tion, it should be gone after 24 hours. If it’s not related to a case, we are
not allowed to keep it [. . .] There’s no restriction to getting it again the
next day, reanalyse it and draw the same conclusion: we’re not allowed
to keep it and throw it away.
“[. . .] As soon as you suspect that you want to investigate possible
incidents related to bike gangs or whatever, you have a case and you
have defined it as case, even if it’s not related to picking up people and
everything else. It’s just mining information and analysing information
related to a possible incident or whatever. So it’s a case, it’s always a
case.”

Thus, the practice and design of the notion of ‘case’ becomes a crucial
factor regulating what information is legally stored and what’s not. The way
a case’s scope is defined has far-reaching implications for the amount of data
that can be gathered. This is because only data not related to a ‘case’ may
qualify for deletion on a short-term basis (24 hours). The information that is
gathered as part of a case ends up stored for longer periods, and potentially
this is exposed to the scrutiny of officers.
To prevent the exposure of data to a wide range of officers, Officer Cees
mentions that there are restrictions built into the design of the technology.
Police officers are given specific regimes of use, depending on their ‘legal
 200 Vlad Niculescu-Dinca et al.
profiles.’ These are specifications that limit the system’s affordances depend-
ing on each type of user:

“Their legal profile decides whether they have all the widgets or just a
few. So you can expect a policeman to have more legal possibilities than
a tax investigator or somebody at a lower government.”
“. . . [The technology will] let user A only look at public Facebook
profiles and user B, because he has a lot more possibilities from a legal
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

point of view, get behind the front door of Facebook.”

As pointed out by Akrich (1992), these quotes highlight that designers

define anticipated users, endowing them with particular capabilities when
inscribing predictions about the context of use. However, it may still be
that actors define their own roles even though the technical artefact may
prescribe certain imperatives or invite default patterns of action (Akrich and
Latour, 1992). While ‘privacy by design’ may offer guidance with respect to
information management, it remains abstract or silent in terms of what a
user/police officer is able to define as a ‘case.’ When I asked the designer to
give more examples of cases, these issues related to case design became clear:

“Right now we’re testing on the topic of the Olympics [. . .] to see what
it’s doing with all the information related to the Olympics. I can assure
you that’s a lot, even based on a few keywords. [. . .] All the information
related to the Olympics is stored as a history. But it’s always case related.
We don’t store all the information that is not related to it.”
Interviewer: “What if in the course of a case and its development, the
boundaries of the case are changing because you acquire new insights?”
“That’s very flexible, because the investigators in control can always,
at any moment in time, add new information or add new searches or
add new source material of which they think: hey, we know something.”

Thus a case’s scope and its technological translation can be loosely

defined. Larger or smaller amounts of information end up legally stored as
part of a case at hand. This aspect is not regulated by design but by users and
procedures. Police officers can define and redefine cases as small or as large
as they need. While the notion of case is traditionally associated to reactive
policing and to aspects such as committed crime, the case’s scope, case (re)
definition, and case closing, these features are significantly transformed or
lose their ground when they are translated in software code and employed
in proactive policing practices.
Having ‘cases’ as starting points for monitoring, these technological
solutions are thus compatible with the practice and procedure of monitor-
ing problematic youth groups. This is because shortlisted youth groups are
already defined as cases in the crackdown approach. Each shortlisted group
is analysed and prioritized, with some getting selected for closer monitoring,
 Sorting (Out) Youth 201
including on Internet and social media. As police departments are already
interested to extend their technological solutions towards Internet and social
media (i.e. they acquired and use commercial social media monitoring), the
internally developed solution is an option for them.
Several points can now be brought together. SMM practices perform new
forms of groups and new ways of inferring suspicion. The transformed prac-
tices may not require having ‘a group’ as a case for monitoring, as cases
can be easily expanded, and each new suspect can be defined as a case (“He
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

deserves the attention, for sure”). In a policy context which recommends

‘getting the most complete picture possible on the person’ (police poster) and
an organizational context lacking resources (as analysed in the evaluation
report), automated solutions predispose the police to delegate more ‘cases’ to
be monitored with automated technologies. Even if surveillance is announced
to be only case based, as opposed to monitoring all social media activity, this
does not limit the expansion of their number or loose definitions.
This section highlights the pivotal nature of the notion of ‘case’ in pro-
active surveillance practices. This includes their scope (what belongs to a
‘case,’ how and who expands it), their lifecycle (how is it redefined, when is
it closed), and their translation in technological implementation.

CONCLUSION

This chapter analysed practices and technologies involved in policing ‘prob-

lematic youth groups.’ In line with the evaluation report presented in the
first section (Van Montfoort, Office Alpha and Dutch Youth Institute,
2013), it showed how classifications and aggregated statistics on shortlisted
youth groups do not simply mirror the dynamics of the phenomenon. That
is, a variation in the number of groups reported by the shortlisting method
may not necessarily mean that there are in fact fewer or more groups from a
certain category. Rather, technologies and practices play an active role in the
ways in which youth groups are performed as ‘problematic’ and classified
in one category or another. As it was illustrated, practices of classification
produce the entities that are counted. This insight not only points towards
difficulties in the method but proves relevant for the amount of resources
allocated to monitoring.
Proactive policies promote enhancing information exchange and early
signalling on the larger sets of ‘annoying’ and ‘nuisance’ groups (Van Mont-
foort, Office Alpha and Dutch Youth Institute, 2013). Against the back-
ground of a broad adoption of social media by youth, the police tend to use
the notion of ‘crime displacement’ to explain why they see fewer groups on
the street and why they are still justified to monitor their Internet behaviour
in order to ‘get the most complete picture possible on a person.’
Once engaging with SMM, officers encounter new youth relations and
new forms of groups. At the same time, they define new indicators for
 202 Vlad Niculescu-Dinca et al.
suspicion. What a suspect situation, person, or behaviour is depends on
metrics and features specific to social media: how many followers for the
declared age, the clothing in a picture, and so on. Even if these are unac-
counted for in the shortlisting method that was used to select a group in the
first place, they predispose the police to find more cases and thus reinforce
incentives for monitoring.
The chapter shows how these issues are exacerbated with new technological
developments for social media monitoring. These developments promise to par-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tially automate monitoring and reduce labour-intensive police processes. At the

same time, the police aim to account for criticism of social media surveillance
when employing these solutions. In order to account for ‘privacy violation’ of
social media users, police designers delegate their views of privacy protection
towards technology design. The monitoring solution is used only on ‘cases’
that require it, as opposed to monitoring all social media activity, and this is to
be regulated by technology. However, ‘problematic youth groups’ are already
defined as cases. This quickly expands the scope and depth of their monitoring
and easily integrates into organizational procedures and technological designs.
The analysis of this arrangement shows the ways in which ‘cases’ legiti-
mize proactive surveillance. The norms shaping this arrangement are not
only to be found in legal provisions and policy documents but are also hid-
den in the details of software design and use contexts. Therefore, despite
restrictions built into design, particular alignments of technological, orga-
nizational, and policy factors predispose increased automated monitoring
and data gathering. These factors together engender a translation of the
issue of ‘problematic youth groups’ into an information problem, fostering
automatic (social media) monitoring of a growing number of youth.

ACKNOWLEDGMENTS

The research leading to these results has received funding from the Euro-
pean Research Council under the European Union’s Seven Framework
Programme (FP7 2007–2013)/Grant. No. 201853. Besides the support of
the DigIDeas project, the author wants to thank Irma van der Ploeg, Jason
Pridmore, Govert Valkenburg, Annelies Falk, Karolina La Fors, and Isolde
Sprenkels (the members of the Digideas project) and Tsjalling Swierstra for
their useful comments on previous versions of this chapter. Moreover, the
author wants to especially thank the police staff, municipality officials, and
technology developers for their generous collaboration.

NOTES

1. ‘Problematic youth groups’ is a term used in Dutch policy making to designate

groups of youth (considered between 12 and 25 years of age) that have been
associated with street nuisance and criminality and have been flagged by police
and other governmental agencies for closer monitoring and intervention.
 Sorting (Out) Youth 203
2. The method has been developed by the police in collaboration with Bureau
Beke, a Dutch consultancy bureau specialising in providing advice for policies
on crime and safety issues.
3. To protect the confidentiality of interviewees and of youth registered in
police systems, their names have been anonymised in this study. Officers were
renamed alphabetically (Anna, Bart, Cees, Dirk, and Erwin). This measure
was also taken with respect to the name of the visited police forces, with
the exception of the Dutch national police force, which is one for the whole
country.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

4. According to the shortlisting method, ‘annoying youth groups’ refers to groups

that are seen to hang around the neighbourhood and are occasionally noisy
or have incidental small arguments, which are usually finished quickly. Some
of the members engage in mild violence and property crime. ‘Nuisance youth
groups’ are groups with somewhat more pronounced, provocative behaviour
(e.g. insulting bystanders). They may vandalize things more regularly and may
not shy away from violence. They engage in minor crime and make an effort
to ensure they do not get caught. The last category, ‘criminal youth groups,’
contains youth who (at least in part) have committed more serious crimes (e.g.
burglary, robbery, pimping). They often come into contact with the police and
are not afraid to use force. They usually commit crimes not just for status but
also for financial gain.
5. Crime displacement has been classified into five types: ‘temporal,’ meaning
that the intended crime is committed at a different time; ‘spatial,’ referring to
the intended crime being committed in other places; ‘tactical,’ which implies the
commitment of the same crime but in a different way; ‘target,’ which is about a
different target then the one originally planned; and finally ‘type’ displacement,
referring to another kind of crime from that initially intended. Simon Hakim
and George F. Rengert, Crime Spillover (Beverly Hills, CA: Sage, 1981).
6. www.wegwijzerjeugdenveiligheid.nl/nieuws/2012/050912_jeugd-gaat-liever-
gamen-dan-hangen
7. Coosto is a Dutch-based social media monitoring company whose solutions
enable “monitor[ing of] social media, including Twitter, blogs, forums and
Hyves. More than 380,000 sources are visited daily and indexed. Coosto gives
rapidly a clear answer to the questions: who, what, where and when? The
results of Coosto are always available” (Coosto website, 2013).
8. Privacy by Design is a framework which advances the view that privacy cannot
be assured solely by compliance with legislation and regulatory frameworks.
Privacy and data protection need to be embedded throughout the entire life
cycle of technologies, from the early design stage to their deployment, use,
and ultimate disposal. Ann Cavoukian, “Privacy by design: The 7 foundational
principles” (Canada Information and Privacy Commissioner of Ontario, 2009);
European Commission, “COM 609, A comprehensive approach on personal
data protection in the European Union” (European Commission, 2010).

REFERENCES

AKRICH, M. (1992) The De-scription of Technical Objects. In Bijker, W. E. & Law,

J. (eds.) Shaping Technology/Building Society—Studies in Sociotechnical Change.
Cambridge (MA): The MIT Press, 205–224.
AKRICH, M. & LATOUR, B. (1992) A Summary of a Convenient Vocabulary for
the Semiotics of Human and Nonhuman Assemblies. In Bijker, W. E. & Law, J.
(eds.) Shaping Technology/Building Society: Studies in Sociotechnical Change.
Cambridge (MA): MIT Press, 259–264.
 204 Vlad Niculescu-Dinca et al.
BARR, R. & PEASE, K. (1990) Crime Placement, Displacement, and Deflection.
Crime & Justice, 12. p. 277–318.
BORST, W. L. (2013) Privacy by Design on the Crossroads of Chains, Lessons from the
Chain of Criminal Justice in the Netherlands. Journal of Chain-computerisation,
Information Exchange for Chain Co-operation, 4(5). p. 3–10.
BOWKER, G. C. & STAR, S. L. (1999) Sorting Things Out: Classification and Its
Consequences. Cambridge (MA): MIT Press.
VAN BRAKEL, R. & DE HERT, P. (2011) Policing, Surveillance and Law in a Pre-
Crime Society, Understanding the Consequences of Technology Based Policing. In
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

De Pauw, E., Ponsaers, P., Vijver, K. V. der & Deelman Piet (eds.) Technology-Led
Policing. Antwerpen: Maklu Cahiers Politiestudies, 163–192.
BUREAU BEKE. (2009) Shortlist Methodology in Seven Steps. Arnhem: Bureau
Beke.
BUREAU BEKE. (2010) Approach to Problematic Youth Groups, Guidelines for Munic-
ipalities. The Hague: Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.
BUREAU BEKE. (2013) Problematic Youth Groups in Netherlands, Extent and
Nature in Autumn 2012. The Hague: Ministerie van Veiligheid en Justitie.
CALLON, M. (1991) Techno-Economic Networks and Irreversibility. In Law, J.
(ed.) A Sociology of Monsters: Essays on Power, Technology and Domination.
London: Routledge, 134–164.
CAVOUKIAN, A. (2009) Privacy by Design: The 7 Foundational Principles. Ontario:
Information and Privacy Commissioner of Ontario.
CHAN, J.B.L. (2001) The Technology Game: How Information Technology Is Trans-
forming Police Practice. Journal of Criminal Justice, 1. p. 139–159.
CHU, J. (2001) Law Enforcement Information Technology: A Managerial, Opera-
tional and Practical Guide. New York: CRC Press LLC.
CLARKE, R. V. & WEISBURD, D. (1994) Diffusion of Crime Control Benefits:
Observations on the Reverse of Displacement. Crime Prevention Studies, 3(2).
p. 165–178.
EUROPEAN COMMISSION. (2010) A Comprehensive Approach on Personal Data
Protection in the European Union. COM(2010)609 final.
VAN GEMERT, F. (2012) Five Decades of Defining Gangs in the Netherlands: The
Eurogang Paradox in Practice. In: Esbensen, F. A. & Maxson, C. L. (eds.) Youth
Gangs in International Perspective: Results from the Eurogang Program of
Research. Dordrecht: Springer. p. 69–83.
GERSON, E. M. & STAR, S. L. (1986) Analyzing Due Process in the Workplace.
ACM Transactions on Information Systems, 4(3). Special Issue: Selected Papers
from the Conference on Office Information Systems.
HAKIM, S. & RENGERT, G. F. (1981) Crime Spillover. Beverly Hills (CA): Sage.
KOOPS, B., et al. (2012) Juridische Scan Openbrononderzoek. Een Analyse Op
Hoofdlijnen Van De Juridische Aspecten Van De Irn/Icolumbo-Infrastructuur
En Hdief-Tools. Tilburg: Tilburg University, TILT—Centrum voor Recht, Tech-
nologie en Samenleving.
LATOUR, B. (1987) Science in Action. Cambridge (MA): Harvard University Press.
LATOUR, B. (1994) On Technological Mediation: Philosophy, Sociology, Geneal-
ogy. Common knowledge, 3(2). p. 29–64.
LAW, J. (2009) Seeing Like a Survey. Cultural Sociology, 3(2). p. 239–256.
LAW, J. & MOL, A. (2001) Situating Technoscience: An Inquiry into Spatialities.
Environment and Planning D: Society and Space, 19(5). p. 609–621.
LEMAN-LANGLOIS, S. (ed.) (2012) Technocrime, Policing, and Surveillance.
Abingdon, Oxon, New York: Routledge.
MANNING, P. (2008) The Technology of Policing: Crime Mapping, Information
Technology, and the Rationality of Crime Control. New York and London: New
York University Press.
 Sorting (Out) Youth 205
MCLENNAN, D. & WHITWORTH, A. (2008) Displacement of Crime or Dif-
fusion of Benefit: Evidence from the New Deal for Communities Programme.
London: Communities and Local Government Press.
VAN MONTFOORT, OFFICE ALPHA, AND DUTCH YOUTH INSTITUTE. (2013)
Evaluation of the Approach to Criminal Youth Groups. The Hague: Ministry of
Security and Justice.
OERLEMANS, J. & KOOPS, B. (2012) Surveilleren En Opsporen in Een Inter-
netomgeving. Justitiële verkenningen, 38(5). p. 35–49.
DE PAUW, E., et al. (eds.). (2011) Technology-Led Policing. Journal of Police Stud-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ies, 2011–3(20). p. 1–32.

SPERGEL, I.A. (1990). Youth gangs: Continuity and Change. In M. Tonry and N.
Morris (eds.) Crime and Justice: A Review of Research, vol. 12, Chicago, IL:
University of Chicago, pp. 171–275.
STEEVES, V. (2012) Hide and Seek: Surveillance of Young People on the Internet. In
Lyon, D., Haggerty, K. & Ball, K. (eds.) The International Handbook of Surveil-
lance Studies. New York: Routledge, p. 352–359
WILLIAMS, V. & WILLIAMS, B. (2008) Technology Applications: Tools for Law
Enforcement. In Ruiz, J. & Hummer, D. (eds.) Handbook of Police Administra-
tion. London, New York: Taylor & Francis, 224–258.
 10 Identifying the Perpetrator
An Ethnographic Study of CCTV
in Police Work in Denmark
Peter Lauritsen
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

Within the last two decades closed-circuit television (CCTV) has become an
integral part of police work in Denmark. It is used to solve a range of crimes
from murder to shoplifting and has made the police develop new ways of
thinking and acquire new competencies. Thus new practices revolving around
the use of CCTV have come into being. When police officers arrive at a crime
scene, they instantly look for surveillance cameras to see if the incident could
have been filmed. With approximately 500,000 surveillance cameras in the
country there is a fair chance that this is in fact the case (SecurityUser no. 5,
2013). At the police station the footage is handed over to technical experts
who are employed by the police and know how to retrieve and handle the
information. If some of the pictures are useful they are included in the further
investigation of the incident, and sometimes the footage is used in court as
a way of presenting what happened before, during, and after the crime was
committed or as a direct means of identification.
This chapter focuses on the actual use of CCTV footage in crime inves-
tigation. This is an interesting area of study because the use of CCTV in
police work is supported by a strong claim that the technology will make
police work much more effective (Lauritsen, 2011). The basic and often-
stated argument is that more crimes will be solved at a low cost because of
the cameras. The article treats this claim as a folk theory (Rip, 2006). Like
other folk theories the theory of CCTV effectiveness is characterized by
two important aspects: (1) It lacks empirical support but (2) still serves as
an engine for future action. Thus many important decisions and debates in
Denmark rest on the theory of the effectiveness of CCTV, and it has become
increasingly difficult for privacy advocates and other concerned citizens to
have their voices heard. If crime can be fought with CCTV, why not use it?
The aim of the chapter is to use an ethnographic study to critically engage
in a discussion of the effectiveness of CCTV in police work. Although it
cannot be concluded that CCTV is never effective, the result is nevertheless
a weakening of the folk theory. Or rather it is a more nuanced theory that
points to CCTV as a valuable tool in police work but also shows how it
often takes skills and hard work to make CCTV productive.
 Identifying the Perpetrator 207
Theoretically the chapter uses the concept of oligopticon to interpret the
empirical findings. The oligopticon has its roots in actor-network theory
(ANT) but puts a specific focus on surveillance practices (Latour, 2005;
Latour and Hermant, 2006). In particular, the concept is used to stress the
networked character of surveillance. Thus an important insight is that the
working of surveillance systems depends on a plethora of actors and that it
sometimes can be difficult to align and coordinate these actors. In the pro-
cess, surveillance itself becomes limited and fragile.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The oligopticon adds important nuances not only to folk theories of

police use of CCTV but also to prevalent academic understandings of sur-
veillance, most obviously perhaps to the concept of the panopticon devel-
oped by Jeremy Bentham in the 18th century (Bentham, 2011) but made
famous by Michel Foucault (1991). As is well known, the panopticon is, in
its original form, a model of a prison which is built in a way that makes it
possible for a guard to watch the inmates, but the inmates cannot watch the
guard. Accordingly the inmates assume that they are under constant surveil-
lance, and this has, the theory states, disciplining effects.
The panopticon remains influential in academic discourse and has been
at the centre of many interesting analyses within surveillance studies. The
panopticon is an all-seeing and effective surveillance machine, and very little
is heard about breakdowns in the machinery or the hard work that is put
into maintaining the apparatus. However, where the panopticon is robust,
the oligopticon is fragile. This does not entail that the oligopticon cannot
be productive but it is always an achievement and a result of successful col-
laboration between actors (Gad and Lauritsen, 2009).
The chapter progresses in the following manner: First the folk theory of
CCTV is sketched out in more detail, and two significant examples from
the history of CCTV in Denmark are included to illustrate the folk theory.
The second section focuses on how CCTV is used for solving crimes. The
conclusion is that although CCTV is a valuable tool, it is not as smooth to
use as many people seem to think. In the third section the concept of the
oligopticon is introduced and used to interpret the empirical observations.
I conclude by making clear that surveillance is (1) always carried out by a
network, (2) limited in scope, and (3) often fragile.

A FOLK THEORY OF CCTV

According to Arie Rip, a folk theory “evolves in ongoing practices, and

serves the purposes of that practice. [. . .] What characterizes folk theories
is that they provide orientation for future action” (Rip, 2006, 349). Rip
continues,

Folk theories can be more or less explicit; this also depends on whether
or not they are challenged. They are a form of expectations based in
 208 Peter Lauritsen
some experience, but not necessarily systematically checked. Their
robustness derives from their being generally accepted, and thus part of
a repertoire current in a group or in our culture more generally.
(ibid., 349)

When it comes to the use of CCTV by the police a particular folk theory
is clearly dominating at least in a Danish context. According to this theory,
CCTV is an effective tool for fighting crime. The claim is not further devel-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

oped or backed up by empirical data. Contrarily, the positive impact of

CCTV in police work is presented as self-evident and self-explaining. For
example, the minister of justice in 2009 stated that “it is my conception that
CCTV can be an effective tool that can be used by the police for solving
crime.” Most leading politicians share this view, and surveys have consis-
tently shown that the majority of the population is positive about increased
use of CCTV (Det Kriminalpræventive Råd, 2005).
Perhaps surprisingly the theory of CCTV effectiveness has two oppos-
ing conclusions. In one variant the effectiveness of CCTV leads to the
suggestion that it should be used even more, specifically in order to solve
a range of problems ranging from terrorism, gangs, and rape to feelings
of lack of safety in residential areas. The other variant of the theory does
not disagree that CCTV is effective but warns that the technology has a
flipside. According to this view the effectiveness of CCTV will not only
lead to the capturing of more criminals, it will also lead to a more totali-
tarian and oppressive regime. Often Orwell’s 1984 is used as a reference
for this argument (Lauritsen, 2011). Most of the debate in Denmark can
be traced back to these two opposite viewpoints. They are, however, basi-
cally expressing the same conviction that CCTV is a very effective tool. In
the following I will give two examples that serve to show how the theory
of effectiveness works.
The first example is from 1980, when private citizens in the town of
Hobro installed CCTV in a street that for some time had experienced a lot
of vandalism (Lauritsen and Feuerbach, 2015). In order to end this, a local
trade organization, led by a businessman, installed CCTV in an attempt
to identify the perpetrators. The CCTV system consisted of three or four
(the sources disagree on this matter) cameras. No monitors were installed.
If nothing happened the tapes would be deleted every morning. Otherwise
they would be handed over to the police. However, after some months the
cameras were removed. According to the businessman in charge of the proj-
ect, this happened as the result of pressure from his employer (a bank)—
they believed his engagement in the project hampered their image and could
result in a loss of customers.
The installation of the cameras drew national attention, but politicians
were divided on the subject. A high-ranking member of a right-wing party
agreed that CCTV was unpleasant but found it necessary as long as there
were not enough police on the streets. Other politicians were far more
 Identifying the Perpetrator 209
sceptical and made references to Orwell’s 1984, which was “just around the
corner.” In the end the prime minister promised that the government would
make it illegal to use CCTV in public places, saying “the tendency is danger-
ous and must be stopped.” Accordingly, a law was passed in which the first
paragraph reads: “TV surveillance must not be used on a road, site or the
like, that is used for ordinary coming and going.”
The other example is from January 2008, when a young man was stabbed
to death on the main pedestrian street in Copenhagen. Three young men
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

demanded that the victim gave away his hat, and when he refused, he was
killed. In the following days people tried to cope with the incident. How
could such a thing happen in a country like Denmark, and who was respon-
sible? The media looked for politicians that could be held accountable. A
newspaper asked the Minister of Justice how she would prevent such a thing
from happening again. Her answer was CCTV in public places. Apparently
nobody noticed that surveillance cameras in fact had captured the kill-
ing of the young man without preventing the incident (Dahl and Bolther,
2012). In a matter of days everybody seemed to call for more surveillance
cameras. Mayors, citizens, and members of Parliament all promised that
CCTV would soon be in place in the major Danish cities. Of course CCTV
was already there, for instance at the facades of banks and inside cafes and
shops, but now everybody wanted the whole street to be filmed with new
cameras operated by the police.
The examples clearly illustrate the theory of CCTV as effective. But one
can also see the two opposite interpretations of this effectiveness at work.
In the 1980s, politicians raised concern about the use of CCTV and passed
legislation in order to bring surveillance under control. Almost 30 years later
politicians actively pushed for more surveillance. Now the effectiveness of
the technology was seen as something that should be used in the fight against
crime.
Like other folk theories, the theory of ‘CCTV as effective’ lacks empirical
support. There is, for instance, no empirical evidence that a few cameras in
a small Danish town will lead to an Orwellian surveillance society, just as
there is no evidence that a few more cameras operated by the police will pre-
vent crime and ensure that the police will capture more criminals. However,
as Mads Borup and colleagues state, “In the case of widely shared expecta-
tions [. . .] legitimation is hardly required” (Borup et al., 2006, 289). Also,
the theory is clearly a vehicle for action. In the first example, politicians are
prompted to constrain CCTV. In the other example, they encourage the use
CCTV even more. Finally, it is interesting to note that the folk theory of
surveillance is based on a rather strong technological determinism. Along
these lines it is the cameras, although there are very few of them, which are
seen to drive society in one direction or the other. This focus on technology
has the implication that the context that is decisive for the working of the
technology is left out of the picture. In a more general discussion of tech-
nology and expectations Borup and colleagues state, “Early technological
 210 Peter Lauritsen
expectations are in many cases technologically deterministic, downplaying
the many organizational and cultural factors on which a technology’s future
may depend” (Borup et al., 2006, 290). As will become clear in the follow-
ing, this is exactly what happens when it comes to the use of CCTV by the
Danish police.

CCTV IN SOLVING CRIME

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

CCTV is a well-researched surveillance technology, and there is a vast lit-

erature on the topic. A significant part of the literature has tried to decide
whether CCTV is preventing crime (Welsh and Farrington, 2007) or if the
police solve more or fewer crimes when CCTV is put to use in certain areas
(Ditton and Short, 1999). However, taken as a whole, this literature does not
support the claim made by the folk theories that CCTV is an effective tool.
In fact CCTV only seems to have very limited effect on preventing crime
(Welsh and Farrington, 2007; see also Ditton and Short, 1999).
Although studies of the quantitative sort clearly are of relevance they do
not shed light on how CCTV is integrated in police practices. In fact, prac-
tice is exactly what these studies try to eliminate in their attempt to isolate
the effect of the cameras. An explicit focus on practice is, however, found in
several interesting studies of operators in CCTV control rooms. Using obser-
vation as a primary method, these studies unravel new practices that have
emerged as a result of CCTV. For example, it has been shown how suspicion
is constructed in the control room and how collaboration between CCTV
observers and members of the police force works—or rather does not work
(MacHill and Norris, 2003; Goold, 2004).
These detailed studies of policing practices are interesting, but they
reflect an important difference between the UK (and perhaps other coun-
tries) and Denmark. While control-room practices are widespread in some
countries and almost have become the emblem of CCTV use, this is not the
case in the work of the Danish police. Although they have the opportunity
to watch CCTV pictures live in control rooms, they only do so on certain
occasions—for example, in relation to soccer matches or huge demonstra-
tions. As a senior police officer stated, “we cannot afford spending man
hours on watching a screen where nothing happens. There is simply not
enough action going on.” However, the fact that Danish police only on some
occasions invest resources in actually watching CCTV live does not mean
that the use of CCTV footage is a marginal activity. In fact, the coverage of
CCTV in Denmark is now so high that it is more than likely that pictures
from surveillance cameras are included in the investigation. Thus one can be
certain that the police look for CCTV footage when they arrive at a crime
scene.
The following departs from most research on CCTV in police work as it
investigates how CCTV is used by the police when solving crime. As part of
 Identifying the Perpetrator 211
the study I carried out fieldwork in several Danish police districts. Observa-
tions were made, and police officers were interviewed both formally and
informally. In addition, various documents including media reports were
collected. The fieldwork focused on two questions: (1) How do the police
handle the process of installing CCTV in public areas and (2) How do they
use CCTV footage in crime solving? It is experiences from the latter part
that are presented here. Furthermore, I concentrate on the process that
starts when the police collect footage from surveillance cameras and stops
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

when the suspect is identified or the case is completed or left unsolved. How
CCTV evidence becomes stronger or weaker through negotiations in court
is an interesting topic, but it is not dealt with here.
The police officers I encountered had a highly nuanced understanding
of CCTV. Nobody believed that CCTV can prevent serious crimes such as
homicide, rape, or physical assault, especially not in the many instances
in which these crimes are committed as an act of passion or desperation.
Neither was it a widespread opinion that CCTV would make the public
feel safer. But the police officers readily shared stories of how CCTV had
helped in solving crime. In some instances, these stories showed how CCTV
provided the police with high-quality “portraits” which made it possible to
solve the case almost immediately:

We have plenty of examples. There was this very unfortunate event

where a young moped driver was out practicing with his driving school.
He was run over in a traffic junction and killed. A witness saw that the
people from the car jumped on a bus. But we got hold of CCTV footage
from the bus and were thereby able to identify the culprits.

In this case pictures were used to identify the driver immediately, and the
police could simply pick him up at his home address. However, the typical
impression was that although useful and often described as a “highly valu-
able tool,” the process was often more complicated than this. This can be
seen in the following, in which a police officer presents an example in which
he thinks the use of CCTV was a success.

A person was stabbed to death . . . and at some point we get a hold of

CCTV from a taxi. I looked at the footage, and I could see that there
was a person sitting in the backseat. But it was only a dark spot because
it was dark and the quality of the pictures was not good enough. But
using a piece of software I analysed the footage frame by frame. And
there was one single picture where the person did like this [shows how
the suspect looks up]. But still one could only see a dark shadow. How-
ever, using photo-editing software we succeeded in developing a picture
of the face of the person. We showed the portrait to the guys in our
special unit and they said “We know this guy.” And it is was in fact
the killer.
 212 Peter Lauritsen
To the police officer this is a story of success. It took time and skill to
analyse the pictures, but in the end the killer was identified and captured. It
is experiences like these that make most police officers conclude that CCTV
is a valuable tool. However, as already indicated, the use of CCTV can be
difficult and demanding. Often the process runs into obstacles that must be
dealt with. Although the police are often successful in doing this, the effect
of CCTV is uncertain. It cannot be taken for granted that the use of CCTV
produces new evidence or helps much in identifying the perpetrator.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Whether the use of CCTV falls apart or turns out to be productive depends
on many things, but in the following three important themes are discussed.
The themes are at least partially overlapping, but they are all important to
the police when using CCTV, as they all challenge the identification of the
perpetrator.
How is the quality of the footage? One problem the police often face is
that the surveillance pictures are of low technical quality. Also, cameras are
sometimes out of order, or the incident happened outside the reach of the
cameras and hence there are no pictures at all. A truck, leaves on the trees,
or a marquee can hide the incident. In other situations the cameras may
be old and unable to film at night. It is also a common experience that the
perpetrator wears a hood or a mask and thus tries to avoid identification.
In the same manner, license plates turn out to be stolen, or they are covered
in one way or another.
Although surveillance pictures in low quality undoubtedly threaten the
effectiveness of the surveillance system, one cannot conclude, however, that
this necessarily makes the surveillance pictures useless. An experienced police
officer states,

Take the robbery of [the bank] for example. There you can see how they
crash into the building. They drive a car through the fence and three
people dash in. But the quality [of the CCTV footage] is very bad. But I
improved contrast and lighting on a single picture and things like that.
And then one could see parts of their clothing that matched clothing
found at the place where the robbers shared the plunder.

Thus, surveillance pictures are sometimes used for other purposes than
directly identifying the perpetrator. For example the police can get knowl-
edge about how a sequence of events unfolded or if there were witnesses to
the incident. Sometimes they obtain knowledge about the height and stat-
ure of the perpetrator and the clothes he is wearing. This may not lead to
identification of the perpetrator in itself, but it can be important knowledge
in linking, for instance, the clothes to the scene of crime, which can then
strengthen the evidence found by DNA or other forensic analyses.
How can the footage be used? Even in situations in which the police have
pictures of high quality of the face of the perpetrator, things can be difficult.
Often the police do not know the identity of the person in the picture, and
they have to find out. One option is then to ask the media to publish the
 Identifying the Perpetrator 213
picture. However, the police do not do this automatically; this is especially
the case if they are uncertain that the picture in fact shows the perpetrator.
They do not want to blame an innocent person for a serious crime. In other
instances, the media are unwilling to publish a picture because they find
the case too difficult to explain or too small. However, recently the police
have started to distribute surveillance pictures using Twitter, but the impact
of this strategy is still unknown. Another much-used strategy is to publish
the pictures on the intranet of the police. Theoretically this will expose the
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

picture to every police officer in the country, but some police officers admit
that they seldom look at the pictures on the intranet. Knowing this, police
officers with a picture of an unknown suspect often ask specific colleagues if
they know the person. Often they do. Thus, according to the police, a num-
ber of crimes are solved because of the personal knowledge of police officers.
They simply know the perpetrators. However, if the suspect is unknown
things become more difficult. For example, in one instance, the police had
very good pictures from a bank robbery. But after listening to the sound of
the recording the police had reason to believe that the robbers were from a
foreign country. This made the police conclude that the robbery most likely
would not be solved, because it is often difficult to engage police authorities
in other countries in solving crimes committed in Denmark.
How many resources can we invest? A third problem is related to
resources. Often CCTV is discussed as either working or not working, but
in the daily life of police practice, it is a common experience that CCTV in
some situations can be made to work. For example, every gas station has
CCTV capturing the license plates of the cars. Consequently a person driv-
ing away without paying for his or her gasoline can be identified quite easily.
The police simply look up the license plate in the vehicle register and then
call the driver. But if the plates are not clearly visible, or they are stolen from
another vehicle, the police often choose not to investigate the incident. Thus
the surveillance pictures are simply not used. The reason is that it would take
too many resources to solve a rather harmless crime. Or, as a police officer
stated, “You don’t bring in Scotland Yard just because a person has stolen
gasoline worth 300 kroner [40 Euros].” If however the driver continues to
steal gas, an investigation would be opened. Thus when it comes to many
minor crime incidents, CCTV is not ‘working’ because the police choose not
to activate other technologies and personnel in order to make the technology
work. If however a murder or a rape is committed, one can be certain that
every surveillance camera near the crime scene is checked, even if there is only
a remote chance that anything that can help finding the murderer has been
captured on video. One such incident is described in the following:

We had a couple of really serious incidents of rape. We didn’t know who

the rapist was but at some point we got hold of CCTV footage. And we
got a picture of a man that matched the description, and we thought it
was him. But we didn’t know who he was; nobody knew him. Then we
could see on the footage that he spat. We hurried to the scene and found
 214 Peter Lauritsen
the gob of spit. It showed that we had him in the DNA database, and
then we took him in.

In this case, as in many others, the police often spend many hours look-
ing at recordings without knowing precisely what they are looking for, but
sometimes they actually find something that can be of help. In this case it is a
gob of spit that enables the police to activate other resources, specifically their
DNA database. It is this plethora of resources that identifies the perpetrator.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The question of resources is important. Like any other public institution,

the police have to prioritize their resources, and they do not consider CCTV
to be a technology that automatically makes police work more effective. In
fact, CCTV is seen as something that generally requires resources, and there-
fore it should be used with caution. When for example a prominent Danish
Islam critic was shot at, the police reported that they had looked through
1,400 hours of film without finding anything useful. When they were criticised
for not getting hold of even more footage the police answered that it prop-
erly wouldn’t be of much use since they were certain that the culprit wore a
bicycle helmet in the attempt to disguise himself from the cameras. Of course
the police didn’t complain about spending many hours unsuccessfully. It was
an attempted murder, and such cases get the highest priority. But the example
questions the effectiveness of CCTV that is so often taken for granted.

CCTV AS OLIGOPTIC SURVEILLANCE

There is clearly a gap between the way the dominating folk theory and the
ethnographic understanding developed above depict CCTV in police work.
It is also obvious that the kind of surveillance practice described does not
fit neatly into either the concept of Big Brother nor the concept of panoptic
surveillance which dominates many academic discussions of surveillance.
Thus, the practice described above is not about controlling or disciplining
but about identifying subjects. This process of identification is not auto-
matic and smooth running. Often it takes hard and complicated work that
involves different kinds of resources, and still the result is uncertain. In
some situations identifying a person is easy; in others it is difficult or even
impossible.
The concept of oligopticon offers a performative understanding of surveil-
lance that seems to fit much better with the way CCTV is actually used by the
police. In an oligopticon surveillance is situated, specific, and limited, which
at least to some extent puts the oligopticon in opposition to the panopticon:

Oligoptica [. . .] do exactly the opposite of the panoptica: they see much

too little to feed the megalomania of the inspector or the paranoia of the
inspected, but what they see, they see it well.
(Latour, 2005, 181)
 Identifying the Perpetrator 215
The oligopticon has its theoretical roots in actor-network theory (ANT)
and is thereby built on a “relational materiality” (Law, 1999). It holds that
nothing exists outside of relations, that to exist is to be related. The conse-
quence of this view is that reality (including technologies, science, power,
agency, and of course surveillance) is constructed and maintained in net-
works of relations. It is these networks that produce agency and allow actors
to act, hence the term actor-network. Importantly, these networks are fur-
thermore heterogeneous—that is, they consist of human and non-human
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

actors, who are made to work together.

Bruno Latour illustrates the workings of an oligopticon with the work
of Mrs. Baysal (Latour and Hermant, 2006). Her job is to coordinate the
teaching activities at the university by planning and booking rooms for
lectures. She does this from the confined space of her office without ever
attending lectures, only by consulting paper documents and computer
files containing information about names of teachers, titles of lectures,
and availability of rooms. These technologies participate in the construc-
tion of the situation and make it possible for Mrs. Baysal to inspect and
intervene. But this happens in a very specific way. If a map is lost or an
important document is missing the situation will break down or at least
change. In this way oligoptica are fragile constructions. Or as Latour
states, “From oligoptica, sturdy but extremely narrow views of the (con-
nected) whole are made possible—as long as connections hold.” How-
ever, “the tiniest bug can blind oligoptica” (Latour, 2005, 181). Thus,
an oligopticon is a much more fragile construct than the panopticon.
It consists of specific spaces that allow detailed observation, but only
within a narrow frame. It involves, for instance, the use of maps, docu-
ments, or computer programs and is dependent on such technologies; if a
technology malfunctions or stops working, the vision changes (Gad and
Lauritsen, 2009).
Thus from a position in ANT, and in particular the oligopticon, it comes
as no surprise that it is not CCTV in itself that identifies a perpetrator but a
network in which CCTV is only one node. This network consists of cameras,
police officers, different kinds of software, the DNA database, and so on.
It is only in situations in which the work of these actors is coordinated in
certain ways that identification is successful and hence CCTV is “a valuable
tool.”
The networked character of surveillance implies that the use of CCTV
is fragile. If one of the actors does not perform as expected or simply stops
working the vision of the police is challenged. They may literally become
unable to see the perpetrator. This does not render CCTV useless, but in
order to make identification possible, more resources and more work must
be put into the process. Thus what the police gain from watching the CCTV
footage is not a total or all-encompassing view. What they strive at is a much
more limited and narrow view that is relevant for the task at hand, identify-
ing the perpetrator.
 216 Peter Lauritsen
CONCLUSION

This article has reported from fieldwork aimed at understanding the way
Danish police use CCTV to identify perpetrators and hence solve crime. It
showed how CCTV can be effective in some instances, but this only happens
if a plethora of devices and actors act together and are coordinated in a cer-
tain way. In other instances CCTV is useless and leads to nothing else than a
waste of resources. In between these extremes are, however, many situations
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

in which CCTV is made to work by relating CCTV to other resources and

actors.
Drawing on the concept of oligopticon the article offers three conclusions
that are of more general relevance: First, it takes work to establish surveil-
lance. Surveillance technologies do not work by themselves. They must be
placed in a network, and they are depending on this network in order to
work. Second, surveillance can be fragile. Many things can go wrong and
disturb the picture. This does not deem CCTV useless, but it implies that we
cannot take for granted that surveillance is a smooth-running machine with-
out bugs and friction. Finally, surveillance is not necessarily about getting
the big picture or establishing an all-encompassing situation of surveillance.
When the police use CCTV they are not even interested in the big picture but
in establishing a relevant vision. That is a situation that makes it possible to
identify a suspect.
An understanding of surveillance as oligoptic has important implications
also on a policy level. In Denmark, CCTV and other types of surveillance
systems are applied to all sorts of problems partly because it is believed that
it is rather easy to establish effective situations of surveillance. If, however,
political debate and policy initiatives were based on the recognition that it
is difficult and requires many resources to establish even fragile situations
of surveillance, things would look different—and politicians might start to
look for other relevant solutions to problems of security and safety.

REFERENCES

BENTHAM, J. (2011) Pantoptikon—Magtens Øje. Aarhus: Klim.

BORUP, M. et al. (2006) The Sociology of Expectations in Science and Technology.
Technology Analysis & Strategic Management, 18(3/4). p. 285–298.
DAHL, O. & BOLTHER, S. (2012) Drabschefen. Nye Sager. København: Ekstrabladets
Forlag.
DET KRIMINALPRÆVENTIVE RÅD. (2005) Tv-overvågning. Fakta om Tv-
overvågning i Danmark. Glostrup: Det Kriminalpræventive Råd.
DITTON, J. & SHORT, E. (1999) Yes It Works, No It Doesn’t: Comparing the
Effects of Open-Street CCTV in Two Adjacent Scottish Town Centres. Crime
Prevention Studies, 10. p. 201–223.
FOUCAULT, M. (1991) Discipline and Punish. The Birth of the Prison. London:
Penguin Books.
GAD, C. & LAURITSEN, P. (2009) Situated Surveillance. An Ethnographic Study of
Fisheries Inspection in Denmark. Surveillance and Society, 7(1). p. 49–57.
 Identifying the Perpetrator 217
GOOLD, B. J. (2004) CCTV and Policing: Public Area Surveillance ad Police Prac-
tices in Britain. Oxford: Oxford University Press.
LATOUR, B. (2005) Reassembling the Social: An Introduction to Actor-Network-
Theory. Oxford/New York: Oxford University Press.
LATOUR, B. & HERMANT, E. (2006) Paris: Invisible City. Available from www.
bruno-latour.fr/sites/default/files/downloads/viii_paris-city-gb.pdf. Accessed 14–
04–2013.
LAURITSEN, P. (2011) Big Brother 2.0. Danmark som Overvågningssamfund.
København: Informations Forlag.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

LAURITSEN, P. & FEUERBACH, A. (accepted 2015): CCTV in Denmark: 1954–

1982. To appear in Surveillance & Society.
LAW, J. (1999) After ANT: Complexity, Naming and Topology. In Law, J. & Has-
sard, J. (eds.) Actor Network Theory and After. Oxford: Blackwell.
MCHILL, M. & NORRIS, C. (2003) CCTV Systems in London. Their Structure and
Practices. Working Paper no. 10, Urbaneye.
RIP, A. (2006) Folk Theories of Nanotechnologists. Science as Culture, 15(4).
p. 349–365.
WELSH, B. C. & FARRINGTON, D. P. (2007) Closed-Circuit Television. Surveil-
lance and Crime Prevention. A Systematic Review. Stockholm: Swedish Council
for Crime Prevention.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Part IV

On the Move
Migrants and Travellers
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 11 The Digital Evacuee
Mediation, ‘Mobility Justice,’
and the Politics of Evacuation
Peter Adey and Philip Kirby
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

INTRODUCTION

Two weeks before Hurricane Sandy devastated the Eastern Seaboard of

the US, as well as Cuba and Jamaica, the New Jersey Office of Homeland
Security and Preparedness announced that it had awarded a contract to the
group Radiant RFID, who were commissioned to develop an ‘evacuation
solution.’ This would work in advance of and during an emergency to track
and therefore assist in the management of people, pets, vehicles and other
emergency assets. Radiant’s RFID tags were designed to surveil or trace
these subjects and things across the cycle of evacuation mobility, as people
moved from their homes, through processing by disaster officials, to tempo-
rary accommodation, and then, assuming all had gone to plan, back to their
homes. Radiant’s work in New Jersey was prefigured by their development
of similar technology for the State of Texas (2011), and other states across
America are currently in the process of developing such systems, too, includ-
ing Colorado, Illinois, Louisiana, Ohio, and Virginia. The use of evacuation
management software, integrated by ICTs, owe their particular origins to
Texas, which we will trace. However, they are also of particular interest and
importance. For, in tandem with other federal programmes, they represent
part of a disparate series of bureaucracy, science, technologies, and surveil-
lance systems from which the ‘digital evacuee’ has evolved.
In this chapter, we investigate evacuation as a particular form or way
of watching over and governing mobility, that requires far more sustained
study. Evacuation is noticeably absent from many of the diverse mobilities
of humans, non-humans, and things that have been opened up to critical
insight through the putative ‘new mobilities turn’ or ‘paradigm’ (although
see Sheller, 2013, for an exceptional study of precisely those who could
not evacuate). Indeed, for the few studies that have turned their attention
to evacuation from this perspective, evacuation has been comprehended
through a broader politics of mobility and meaning which elaborates pre-
cisely the depoliticisation of the term (see, for example, Cresswell, 2006).
In other words, evacuation mobilities have often been approached and are
often represented—as Cresswell shows—in such a way that they are bled
 222 Peter Adey and Philip Kirby
of all social meaning, notably in the testimony of New Orleans’ evacuation
officials. To present evacuation as a purely technical act or engineering solu-
tion might be understood as a political move of closure. That is, mobility is
naturalised as the only possible outcome that removes complicity, guilt, or a
more distasteful politics of racial discrimination from its production. Nota-
bly, too, the involvement of surveillance practices in disaster and emergency
contexts remains underinvestigated within the surveillance studies literature.
This chapter seeks to delve far deeper into the practices and technologies of
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

evacuation and its management.

First, the chapter sets out an approach by drawing together the insights
of mobilities, STS, and conceptualisations of emergency and evacuation. It
then traces a genealogy of evacuation surveillance, registration, and tracking
as it emerged during the Second World War, with the latter’s attendant evac-
uation of mass populations. Second, it explores how the earlier ‘systems’
failed in the chaos of modern emergency management, especially during
Hurricane Katrina’s uneven displacement of New Orleans residents in 2005,
as well as those of Louisiana and the wider region. The chapter considers
how those events provided the impetus for the construction of a nationwide
federal system of evacuation registration and tracking, as well as a state
system, with more parochial, perhaps even divisive, concerns.

EXPLORING EVACUATION MOBILITIES

In some ways, evacuation seems a poor fit for existing and more famil-
iar categories of mobile subjects, such as ‘the homeless,’ ‘refugees,’ and/or
‘migrants,’ even if evacuation certainly touches upon these subjects. In oth-
ers, including the ways that such categories direct state and non-state inter-
ventions toward certain people, it aligns much more precisely. Evacuation
mobilities, for example, have traditionally been some of the most conten-
tious and ‘unknown’ forms of mobility, demonstrated by the fact that, since
the Second World War, numerous systems of governmental surveillance and
registration have sought to manage their inherent plurality and unpredict-
ability. More recently, ‘the evacuee’ has figured on the agendas of authori-
ties pertaining to public health, transportation and highways, policing and
emergency services, security, intelligence, and counter-terrorism, throughout
the scales of local, state, and federal governance and supranational bodies,
NGOs, and the aid sector, as well as a plethora of private companies. The
evacuee is a subject moved away from harm. But the evacuee is also a mobile
subject constituted as a figure by these various skeins of power, knowledge,
technologies, and practices.
It is becoming increasingly important, therefore, to ask just what is at
stake in the surveillance, categorisation, mobilisation, and treatment of the
evacuated as it is a figure that seems to be moving across different grounds,
approaches, institutions, and understandings, at some pace. Central to this
 The Digital Evacuee 223
are the ways in which the evacuee has been subjected to an extensive but
uneven effort to render it as ‘legible’ (Scott, 1999) from its mobile and slip-
pery condition through the administration of different forms of authority,
bureaucracy, and technology. The primary means to do this have been the
application of a series of techniques and technologies that make the evacuee
visible and then sort, order, and manage that subject physically and within
complex and multi-scalar systems of records and databases, accordingly.
Opening up evacuation, then, requires the elaboration of several approaches
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

from interdisciplinary research fields such as science and technology studies

(STS), ‘new mobilities,’ and political-theoretical research on emergency.
First, the new mobilities paradigm shifts attention to take seriously life on
the move orchestrated and experienced through different kinds of mobilities
and infrastructures. Prior approaches to evacuation might tend to fetishize
the locational start and end points of the internally displaced, or refugees,
for example in migration and humanitarian studies. Engineering fields
instead see evacuation as a more abstract technical object, perhaps obeying
certain physical or social laws (Helbing, Farkas, Molnar and Vicsek, 2002).
Approaches from within ‘new mobilities’ encourage the examination of
movement as it is lived and experienced (Hannam, Sheller and Urry, 2006;
Urry, 2007). That said, perspectives from mobilities studies have not tended
to be brought to the study of emergencies or disasters until relatively recently
to explore—in Cresswell—how evacuation mobilities are given particular
meanings in Hurricane Katrina (Bartling, 2006; Tim Cresswell in Bergmann
and Sager, 2008). And in Sheller’s investigation of the humanitarian effort
to deliver aid to Port-au-Prince following the earthquake in Haiti (Sheller,
2013), how evacuations falter through a power-geometry of mobility and
immobility.
The approach taken in this chapter is to build on these insights but to
also focus on the circulation of evacuation and the evacuee as a techni-
cal object, held together by various systems that surveil, categorise, sort,
and manage mobilities (Aas, 2011). Surprisingly, the surveillance studies
literature (Lyon, 2007) is remarkably absent from the fields of disaster man-
agement, becoming far more attuned to the quotidian uses of surveillance
in the everyday, although see Monahan (2007) for an exception on intel-
ligent transportation systems (ITS) and their role in emergency situations.
More recently, Buscher (2013) suggests that the application of big data to
emergency-management systems presents a new task to explore the poten-
tial ethical questions over how that data might be drawn, stored, and used
appropriately and proportionately, especially over vulnerable populations.
Extending these approaches therefore means asking questions about how
the evacuee is shuttled between systems and software and representations,
between practices held within different institutions and organisations, and
is observed by different practices and technologies.
Third, practices of evacuation are closely associated with emergency poli-
tics, which create particularly susceptible conditions for the creep of ‘security’
 224 Peter Adey and Philip Kirby
and previously unpalatable (and maybe non-legal) measures. However, it is
important that calls for the analysis of evacuation under an Agambenian
‘state of exception’ (Agamben, 1998) do not over-determine evacuation.
In other words, and while we will certainly see evacuation emerging from
emergency situations in which the normal rule of law and the separation
of powers has been suspended, reading evacuation only in the context of
exceptional measures and emergency or catastrophe politics forgets its rou-
tine operation inside the normal protections of liberal democratic systems
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

(Lakoff, 2007; Collier and Lakoff, 2008).

These approaches are used to develop arguments and further questions
around what we could call ‘the politics of evacuated mobilities,’ or what
Mimi Sheller describes as an approach through ‘mobilities justice.’ Sheller’s
is a politics that attempts to develop the “capability of all [. . .] [to] access
mobility in order to meet their own basic needs” (Sheller, 2013). What this
particularly recognises is that mobility capabilities are highly and unevenly
distributed across different spaces, subjects, and bodies. Such a distribution
cannot be understood outside of vulnerability to the kinds of events evacua-
tion is deployed to remove a populace from. This chapter explores the imper-
fect systems that record and manage evacuated populations, which have been
made socially, technically, and legally indistinct and that do not enable but
can prevent the very access to mobility that Sheller describes. It questions the
injustice and sometimes violence associated with the process of ‘becoming
evacuee,’ as well as the often halting motions and mobilities of returning. It
unpicks the negotiation and ambiguity of identity as profiles and registration
practices are continuously befuddled, escaped, and frequently gotten wrong.
And it follows the consequences of all this for our mobile lives, made poten-
tially more vulnerable by the very systems intended for our security.

GOVERNING THE EVACUEE

The current efforts we will explore try to turn the evacuee into a far more
addressable subject through satellite systems, mobile technologies, social
media, and interconnected systems. These advances, relying upon certain
trigger moments, tipping points, or ‘somatic markers’ to facilitate political
capital and technological will, have clearly not just happened on their own
but in a broader and longer historical context. In this section we will draw
something of a partial genealogy of the management and (pre)mediation of
evacuee mobilities from the Second World War.

An Administration of Evacuation
Modern and, importantly, state-led evacuation measures developed in West-
ern Europe in preparation for aerial bomber attack. The evacuation or ‘bil-
leting’ of children from vulnerable urban areas and city centres to rural
 The Digital Evacuee 225
communities and the countryside is in many respects unlike the kinds of
evacuation systems this chapter is concerned with. The eventual responsibil-
ity for a state for moving its people out of the way of disaster, however, was
something of a novel thing. Adi Ophir (2007) locates this original interven-
tion in the 1920s, finding that as late as 1900, even a hurricane hitting the
port of Galveston, Texas, did not warrant the resolve of a state to step in in
order to bring relief or defend its citizens from the repercussions of disaster.
As Ophir writes,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

only at the end of the nineteenth century did the practice of saving lives
become a profession, an art, and a science, as it was developed on the
margins, and as an extension, of medicine, public health, the police,
and the army, and it took a few more decades before this growing body
of knowledge, techniques, and practices was integrated into the state
apparatuses.
(2007, 16)

This is not entirely accurate. The history of the European colonial impe-
rial rivalries could well be a story of state and citizen evacuation follow-
ing successions of territorial disputes overturning colonial outposts and
stations. However, the natural disasters that might require evacuation and
other measures of protection saw states beginning to relate to them no lon-
ger as inevitable catastrophes from which it (and its elites) would evacuate
themselves to somewhere safer (as in the plague). Disaster instead became
a providential moment for a state or sovereign to extend its powers to the
protection of the population from harm.
Modern aerial war threatened populations like never before. In Britain,
pre-war imaginaries heavily informed planning through the inter-war Com-
mittee of Imperial Defence, who were confronted with the first realities of
aerial warfare and the difficulties in protecting the daily routines of a work-
force so essential to public life (Adey, 2010) and the continuation of indus-
trial production and services so essential to fighting a war. The evacuation
of children was considered at length; the enormously complex process of
moving almost 3 million people was coordinated by the Ministry of Health
from vulnerable areas to ‘reception’ areas through a range of scheduled and
organised transport to people’s homes and buildings (O’Brien, 1955). Some
of these were camp-like structures built for the purpose, while others were
requisitioned. Some elderly and vulnerable people also followed in the chil-
dren’s footsteps.
In the United States, where much of this chapter focuses, the largest-scale
evacuations took on quite a different and in some ways far more disturbing
form than the population security intended to remove a valued life to safety.
Instead, the Japanese attack on Pearl Harbor saw the increasing paranoia
towards apparent ‘enemy aliens’ and ‘disloyal citizens’ become more inten-
sively directed at the Japanese (of first- and second-generation migrants,
 226 Peter Adey and Philip Kirby
notwithstanding that existing racial tensions already remained and mani-
fested in protests over ‘American’ jobs). 120,000 people or effectively Amer-
icans of Japanese Ancestry under the executive ‘evacuation’ order 9066,
signed by President Roosevelt, were removed from their constitutional and
amendment rights wholesale. Those of Japanese descent were first volun-
tarily and then forcibly moved away from strategic ‘vital’ military exclusion
zones created on America’s West Coast, particularly around LA and San
Francisco, following several very dubious assumptions based on the Census
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Bureau’s detailed demographic data of the location of Japanese communities

around sensitive locations (Ng, 2002).
The evacuees were eventually settled in internment camps in out of the
way remote locations, perhaps most famously at Manzanar, California,
nestled between the Sierra Nevada mountains, and in the Alabama foothills
(Houston and Houston, 2002). Such a process involved a suitably large
scale registration system conducted by the War Relocation Authority and
the Western Defences Command based on initial registration of ‘aliens’ first
by the Justice Department which required photography, fingerprinting, and
relatively detailed registration data. Considerable information sharing also
occurred with the FBI, who investigated suspects and made apprehensions
of other individuals who were highlighted to them. Under categorisation
as Japanese Citizens, Japanese-American Citizens, and Alien Enemies other
than Japanese, the blanket order meant anyone inside these categories was
subjected to evacuation from the designated military areas.
Like the diagrams of zones and stages of evacuation logistics routinely per-
formed by militaries, civilian evacuation appeared a bureaucratic, adminis-
trative, and orderly affair of plans, sequences, and operations represented in
link and node diagrams. Of course, these abstractions tell us much less about
the lived experiences they represented or of their potential violence. In the
wake of Pearl Harbour and other events of the war, outspoken public com-
mentators and influential newssheets encouraged the wider public, as part
of a broader effort, to seek out ‘saboteurs,’ ‘fifth columnists,’ and ‘enemy
aliens’ resident in the United States. For the Japanese American experience,
this meant overt forms of racial profiling through measures coloured by
pseudo-patriotic racial suspicion to, for instance, discriminate between a
person of Chinese and Japanese ancestry in order to ‘shop’ a potential evac-
uee to the authorities for questioning and eventual ‘evacuation’/internment,
first through assembly centers and then to one of the 10 ‘reception’ or intern-
ment camps located away from the military exclusion zone that encircled
many parts of California and the West Coast. Like many other forms of
internment/registration, the evacuees were similarly given a number. Mine
Okubo (1946), an artist who described and sketched her experiences of
internment, titled her book Citizen 13660. The bureaucratic-militaristic and
racially coded rendering of the Japanese ‘evacuees’ was thus a process of
inscription: of representations, of tags and markers that individuated and
abstracted those subjects for evacuation, processing, and internment.
 The Digital Evacuee 227
Cold War Cultures of Simulation
For the period between the Second World War and the hinge point we want
to focus on—that is evacuation management following Hurricane Katrina—
it is impossible to do justice to the nuanced historical, political, and techno-
logical changes regarding evacuation. Evacuations have differed greatly by
cause, by intensity, by temporality, by frequency, by planning (Cutter and
Barnes, 1982). Neither is it adequate to detail the structural changes emer-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

gency evacuation responds to—a changing state system, post-war geopolitical

shifts, and post-colonial moves resulting in mass evacuation movements, for
example, in the formation of Israel and the partition of India that rendered
dispossessed populations on the fault-lines of territory (Aguiar, 2011). What
is clear, though, is that like the Japanese-American experience, populations
have been moved under ‘evacuation’ as an almost euphemism. It is, however,
important to acknowledge the emerging cultures of prediction, experimenta-
tion, and play that brought evacuation into a technical-imaginary domain of
anticipatory simulation science (Ghamari-Tabrizi, 2009).
Geography’s post-war academic complicity in this is becoming clearer,
as historical geographers such as Trevor Barnes and Matthew Farish have
outlined the role of the discipline in the deployment of operational research
methods in the pursuit of evacuation planning, for example, of Bremerton,
Washington, in the mid-1950s. As they argue, “Urban evacuation’s lines
of flight and the systemic grids of emergency response represented a ratio-
nal ideal, but also the refined unreality of abstraction” (Barnes and Farish,
2006, 820). In this Cold War context, evacuation mobilities became the sub-
ject of increasingly complex stochastic models on the one hand—abstracted
into the variables of a burgeoning positivist science—and on the other, the
literal thought experiments of military intellectuals writing scenarios which
inspired the more practical lived abstractions of evacuation war gaming.
Those ‘games’ or ‘exercises’ were built primarily around the management
and coordination of a displaced population leaving in response or in advance
of nuclear attack.
In the late 1950s, however, strategy was famously balanced between shel-
ter and dispersal. In America’s cities, urban analysis was brought to vulner-
ability mapping. While in World War II in Britain as well as America this
urban geography was attentive to understanding the movement of explosive
blast through the channels of the urban landscape (Adey, 2010), post-war
saw far finer-grain analysis of how those channels could permit the move-
ment of evacuating people. The Federal Civil Defence Administration, learn-
ing the lessons of detailed analysis of Tokyo and Nagasaki, drew detailed
maps and overlay tracings which would reimagine the urban landscape as
a network of evacuation escape. These representations would overlay the
identification of likely targets; the cities’ probable distribution of civilians;
and the overall defence pattern of the cities, including engineering depart-
ment designs on traffic control and evacuation assembly (Collier and Lakoff,
 228 Peter Adey and Philip Kirby
2008). Building heights might determine whether they would be likely to
block streets. Evacuation routes could thus be designed on much more
coordinated basis. Even while these kinds of threats now seem far-fetched,
and the threat was predominantly nuclear, the possibility of evacuation also
drew on the lessons of other sorts of events of much less of an order than the
disruption or loss of life predicted in nuclear apocalypse. For example, the
Audrey Hurricane of 1957, which killed at least 500 people in Louisiana,
Texas, and the Bolivar peninsula, brought evacuation to the public’s atten-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tion as a credible measure of broader emergency planning, and response.

Preparing for evacuation composed a general strategy to pre-mediate
emergency events in anticipation of their occurrence through a variety of
representational analytical forms. The Federal Civil Defence Administra-
tion developed the formulaic and statistical modelling of evacuation motion
to develop calculations of survival probability following a thermo-nuclear
detonation. Variables included the speed of evacuation, the positioning of
shelters along the evacuation route, and, notably, the critical relation to
when the evacuation occurred. In this reasoning, an evacuee would be far
safer taking shelter at the time of detonation than they would be travelling.
On the other hand, should evacuees leave before detonation, and depend-
ing on the following radiation fallout, it could have a considerable effect
on population survivability. Neither were these tactics immune from other
cultures of (auto)mobility, despite some of the myths surrounding the auto-
mobile’s centrality in US civil defence measures (Packer, 2008)—particularly
the design of the interstate system. The ideal mobile unit to replace the
individual, the burgeoning evacuation science began to privilege the auto-
mobile as an ideal vector of escape (an assumption that would come to
feature high on the list of criticisms of the events in New Orleans in 2005;
Bartling, 2006; Cresswell, 2006). And insofar as the involvement of figures
such as Bertrand Klass, who conducted psychological surveys of reactions to
bombing, and director of Stanford Research Institute’s Applied Behavioural
Science research division (and even wrote on subliminal advertising), could
have guaranteed the significant study of the psychology of the individual or
family, these approaches generally reduced the evacuee to an aggregation
or as a line of flight. In these early models, the evacuee was simply a line of
movement in a greater conglomeration of fleeing and semi-rational subjects
(Orr, 2006).

Agents
Just as these approaches receded from military authorship following the
height of Cold War nuclear dispersal planning which had found concrete
form in the structural planning of backyard bunkers, large-scale post-war
suburban development, highway planning, and inner-urban degradation,
other fields took on the mantle of evacuation mobility studies in response
to a wider set of threats, such as natural disasters, chemical accidents, and
 Incident
Place
Year

I. Incident.
1. Nature: Natural Accident Exercise
2. Description:
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

3. Build-up: Instant Hours Days

4. Prior Warning: Of Possibility: No Yes When How
Of Approach: No Yes When How
Of Event: No Yes When How
5. Time of Event: Date Hour

II. The Evacuation.

1. Area Evacuated:
2. Reception Area:
3. Number of Evacuees: Total Population
4. Time of Evacuation: Start: Day Hour ; Finish: Day Hour
5. Evacuation Order: By In Name of
6. Spontaneous: No Yes Triggered By
7. Movement Control: No Yes By

III. The Preparation.

1. Plan: No Yes By Date
2. Publicized: No Yes By How
3. Exercise: Government: No Yes Date
Public: No Yes Date
4. previous Evacuations: No Yes Dates

IV. Results:
1. Deaths: No Yes Number
2. Injured: No Yes Number
3. After-Incident Report: No Yes Findings

Recommendations

V. Documentation:

Figure 11.1 Evacuation study sheet

Source: Strope et al., 1975.
 230 Peter Adey and Philip Kirby
fire. The concern of several evacuation scholars at the time was that longer-
term analysis of the effectiveness of evacuation plans was particularly dif-
ficult to achieve given the inadequacy of effective record keeping during
an evacuation and in after-action reports following an emergency (Strope,
Devaney and Nehněvajsa, 1975). Insofar as the evacuee had been made
known within earlier bureaucratic techniques, broader analytical metrics
were not being recorded systematically to permit longer-term analysis across
multiple-type emergencies.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

As improved ways to record evacuation would develop along with tech-

niques to register and process evacuees, advances in computing saw the
new approaches reaching their finest resolution ever in the evolution from
fluid-flow models that averaged out the uncertainty of individual behaviours
(Henderson, 1974) towards the individual itself. The move in rapidly devel-
oping changes in personal computer use and far more sophisticated software
transformed evacuation simulation. The finer granularity of ‘agent-based’
modelling techniques could assign a range of rule-based characteristics to
agents that would simulate individual behaviour and what that might mean
for aggregate crowd behaviour. This meant that for the first time, individuals
could be simulated and visualised in evacuation scenarios. These simulations
would be used especially for fire evacuations, for example, of major build-
ings such as airports and railway stations, and understandably for more
routine forms of mobility, such as traffic flow.
In conclusion, evacuation techniques have evolved through a variety of
bureaucratic means to manage large population groups. The evacuee has
been subjected to increasing layers of abstraction and representation which
has become, perhaps, increasingly distanced from the evacuee itself. So we
see advances in cultures of anticipation and simulation informing the design
of cities, built spaces, and evacuation plans with a sort of shadow evacuee
in their models. In the following section, we explore how and some of the
arguments why this distance is being foreshortened.

THE ‘DIGITAL EVACUEE’

Hurricane Katrina formed over the Bahamas on August 23, 2005, and dis-
sipated across the US’s Great Lakes a week later. As has been well docu-
mented, over these 7 days, it inflicted more than $100 billion worth of
damage, devastated New Orleans and parts of other states in its path, and
claimed nearly 2,000 lives (Knabb, Rhome and Brown, 2006.). The storm
also precipitated one of the largest evacuations in US history, with 1.2 mil-
lion residents along the Gulf Coast subject to either voluntary or man-
datory evacuation orders and nearly 800,000 eventually being displaced
(FEMA, 2006). The effects of the hurricane were so devastating that, in
its aftermath, the name ‘Katrina’ was retired by the World Meteorological
Organization at the behest of the US government. Whether this was out of
 The Digital Evacuee 231
respect for the victims of the disaster or so that the government could avoid,
rhetorically at least, ‘another Katrina’-style catastrophic policy failure is
debatable. Certainly, the failures of emergency management during Katrina
were manifold, with the paucity of effective evacuee tracking one amongst
many other ways that the evacuee failed to be truly grasped by the different
emergency systems.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Grasping Evacuation
By following the evacuee tracking debacle, we can expound the crucial way
Hurricane Katrina has led to sustained yet fragmented production of the
digitally mediated and—especially—tracked evacuee. And through that, we
can unpack how perhaps the failures of Katrina showed the evacuee up in
the terms of what Jenny Edkins (2011) has called the ‘person as such’ as
opposed to the person as a ‘who.’ In other words, the evacuation measures
we will explore, in rendering the evacuee locatable and eventually digital,
have tended to treat the evacuee as abstract coordinates rather than as fleshy,
feeling, and complex mobile subjects.
Of the 1.2 million Americans who were issued with evacuation orders,
only a small proportion were really all that known in any form of record
keeping at the time, with a fuller picture of where this population dispersed
emerging months, even years later (Groen and Polivka, 2008). Whilst, as we
have seen, the US had already tackled certain evacuee issues of ‘tracking’
in earlier decades, the onus was more readily on anticipating and planning
for where they would go and by what means. Being able to account for the
location of evacuees on this scale was unheralded, with many of Katrina’s
victims slipping into the gaps between physical tagging, paper-based records
recorded at ‘rest centres,’ and, more widely, the overlapping jurisdictions
affected by the disaster and, more bluntly, between those who self-evacuated
and those who required assistance and received it.
Paper-based tracking systems given to those who required city and state
aid, the norm at the time, were simply unfit for the purpose (Pate, 2008).
Material tagging like this functioned in theory, but only if persons retained
their paperwork at all times or recalled the information it contained should
the documentation be lost—problems we will examine in more detail below.
Naturally, though, in the confusion after Katrina, what worked well on
paper was less efficacious in practice. Ironically, the monitoring of the hur-
ricane itself and its complex environmental effects was probably far more
successful. Much more was known about the morphology of the hurricane’s
weather system by meteorological radar. The natural disaster itself was far
easier to get a handle on than the social and technological disaster or ‘com-
plex humanitarian disaster’ occurring (and produced) through the imperfect
emergency response (Smith, 2006).
Perhaps the largest single failing pertaining to adequate registration, knowl-
edge, and effective tracking of evacuation occurred during the disaster has
 232 Peter Adey and Philip Kirby
been expressed in the strongest possible terms already. Naomi Klein (see
Klein and Smith, 2008) referred to the ‘ethnic cleansing’ of the city by sepa-
ration, as poor, vulnerable black communities were displaced and separated
not merely by the hurricane, but by the practises that sought to manage the
response.
The lack of effective tagging, record keeping, and information sharing
meant that in many instances families who could not self-evacuate were
divided, leaving parents unable to track down lost children and babies and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

other family members who could not be reunited for days, weeks, and in
some cases months afterwards (Brandenburg, Watkins, Brandenburg and
Schieche, 2007; Lindsay, 2010). Retrospective testimony of the events from
the perspective of the military coordination reveals a certain kind of think-
ing in the ‘heat’ of the emergency moment, which help to explain the limi-
tations of what seems to have been an ‘exceptional’ response, especially
towards the respect of family ties and inter-personal associations, as above
all, life came first (Graham, 2005). Accounts of survivors and local and state
personnel reveal that some of the main problems in evacuating people away
was that social groups, family units, friends, and other group dynamics were
not properly envisaged or provided for in FEMA’s and the city’s emergency
plans. It was as if the evacuee was treated as one of the disembodied agents
in evac simulations or simply a flow to be redirected out of the way, as
opposed to a subject who sustains and is maintained by subtle and vital but
brittle relationships, fractured by the process of evacuation.
When people are evacuated, they often become more vulnerable because
they are separated from formal and informal systems of care. Children sepa-
rated from next of kin, siblings, and parents/guardians remained in a vul-
nerable state in relocation centers like Oklahoma’s Camp Gruber. Those
who were already known as homeless on the National Center for Missing
and Exploited Children (NMEC) needed to be identified, too (Brandenburg
et al., 2007). Patients that were given medical examinations at certain loca-
tions were unable to be treated at others because their diagnoses did not
travel with them (Franco et al., 2006). There is now general consensus that
of the some 1 million people displaced from New Orleans, very few were
able to be effectively tied to their medical records, meaning that treatment,
prescriptions, allergies, and prior medical history could not follow. Some
hospitals taped patients’ personal notes to their forearms before helicopters
evacuated them elsewhere. The Veterans Association was one such excep-
tion, however, because the medical records of their members were electronic,
which meant that those who were moved from Biloxi, Mississippi, and New
Orleans could be effectively tracked as they moved around the country. Fur-
thermore, 10 days after Katrina’s landfall, a system was setup that would
make private pharmacy data (such as those held at a Walmart) about an indi-
vidual’s repeat prescriptions available to doctors through a secure database.
Even those persons in custody within the criminal-justice system, includ-
ing more than 2,000 ‘registered’ sex offenders despite being, perhaps, the
 The Digital Evacuee 233
most visible, were just as problematically treated. Some were notably evac-
uated without pertinent judicial authorities being informed at the point of
reception (Arlikatti, Kendra and Clark, 2012). But the prison system also
presents an interesting case, because it is often held up as an area in which
evacuation plans and technologies are usually the most comprehensive.
Except in Katrina’s case, the example of Orleans Parish prison presented
one of the worst cases of the legal abandonment of evacuation plans, pro-
tocols, and civil rights, as the ACLU’s documentary and written evidence
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

has shown, and illustrated numerous abuses. Not to mention the fact that
the Louisiana courts, their personnel, and the wide-spread mobility of the
Louisiana judiciary during Katrina meant that the state was practically
evacuated of its legal system (Guidry, 2009), the ACLU (2006; Metzger,
2006) have illustrated how forms of abandonment were effectively written
into Louisiana law. All persons in custody of the Parish of New Orleans
are to be released for a period of 10 days upon the declaration of emer-
gency, disaster, or evacuation in the US or State or city of New Orleans,
and asked to make their own way out of the disaster or register for city or
state assistance.1
What we see in Katrina was the moment that the evacuee—as constituted
in the various ways discussed—made its inadequacies so visible. Since World
War II, many of the measures designed to track evacuees have concerned
two contradictory populations: those ‘at risk’ (e.g. medical patients) and
those ‘posing risk’ (felons). Whilst ‘vulnerable’ populations such as medical
patients (to illness and special needs) and felons (to escape and potential
reoffending) were subject to special, albeit faltering concern as we saw ear-
lier during Katrina, the hurricane also presented challenges for a much wider
population and long after the event. The existing systems prior to and during
Katrina were unsophisticated by their imagination of the evacuee, who they
were, what they needed, inadequately assessing their mobilities capabili-
ties. Furthermore, they were performed in the almost complete absence of
software or ICT despite the fact that so much of an evacuee’s identity might
have existed within several different sorts of organisations, bureaucracies,
and sectors, from health to social security in electronic form. More often
than not, it proved impossible to unite the actual evacuee fully with the less-
than-fragmented digital imprint of him- or herself.
Since Katrina, and the ensuing political fallout, congressional inquiries
and subsequent changes to the practice of emergency planning and response
within the United States, the onus has been on the broadening of what we
might call specific ‘technologies of address’ (Thrift, 2004, 585) through
the development of different tracking programs. For Thrift, locatable or
location-aware technologies are producing what he calls a “‘technological
unconscious’ whose content is the bending of bodies-with-environments to
a specific set of addresses without the benefit of any cognitive inputs” first
evolving from the military arts of logistics. In evacuation management, this
has meant seeking to produce legible evacuees so that they might be easily
 234 Peter Adey and Philip Kirby
and automatically locatable in geographic space as well as within wider
systems of records.
Such technologies have their own particular geographies and may result
in the prioritisation of certain bodies. Like the reaction of the federal govern-
ment to Katrina more broadly, we will explore how tracking after the hur-
ricane was enmeshed in the politics of the mobile body, especially its abilities
and disabilities, its material advantages and disadvantages, and perhaps, in
the case of Texas’ evacuee tracking system, even its ‘racial’ characteristics
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

that have been tied into a broader migration politics of welfare provision
and care.

Tracking the Evacuee

Some of the immediate changes following Katrina occurred through the
Department of Transport, particularly as a means to survey, in real time,
the kinds of mobility pathways evacuees could exploit during an emergency.
Dedicated ‘Evaculanes’ form the basis of many city and state emergency
plans, and these are now coupled with so-called ‘smart roads,’ which are
now able to automatically monitor traffic conditions on relatively remote
or rural roads by using radar, cameras, and video servers at strategic loca-
tions. Before Texas introduced a tracking system which is now being applied
elsewhere by the federal government, FEMA began work on a nationwide
effort, the National Mass Evacuation Tracking System (NMETS). The sys-
tem, developed by FEMA’s office of Geospatial Intelligence, was legislated by
the Post Katrina Emergency Reform Act, and was designed “to assist States
in tracking: Transportation-assisted evacuees [i.e. those requiring specialist
transport, such as the disabled], household pets, luggage, medical equip-
ment” (FEMA, 2010). In short, the system was intended to track vulnerable
populations and their accoutrements. Since 2010, the NMETS software and
training has been provided free of charge whilst states are responsible for the
costs of equipment and maintenance.
NMETS demonstrated particularly well the disjuncture between tradi-
tional systems and what even basic technologies could accomplish. States
wishing to adopt NMETS, either in whole or in part, could opt for one of
three options: a stand-alone, paper-based system (Paper-based Evacuation
Support Tool); a simple digital system (Low-tech Evacuation Support Tool);
or a high-tech digital system (Advanced Technology Evacuation Support
Tool; FEMA, 2010, 2012).
The paper-based system can be used alone and also as a backup should
the more advanced systems fail. As a four-ply carbon copy, the form is an
effective bureaucratic record of key information, including the following
categories: name, address, gender, age, contact number, head of household,
medical needs, caregiver contact details, parent/guardian or caregiver infor-
mation, medical equipment, household pets, luggage, as well as transport
mode, carrier, and details. The second option effectively turns this record
 The Digital Evacuee 235
into an entry within a Microsoft Access database, which means reports can
be drawn in real time and shared quickly and easily allowing, for instance,
the manifesting of a group to be evacuated by bus to be planned with
some ease. The more advanced option means both RFID and barcode tags
and more advanced software for RFID tethering, more advanced evacuee
searches, and the export of the data in other formats.
All three systems promise to overcome the particular issue of separated
families through the use of barcode wristbands which are given to evacuees
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and their important possessions. This means that all members of a house-
hold, including their things, are tied together in the system. The bands are
then scanned at each site to record the evacuee’s location and their departure/
arrival times. From this information, responders can use the knowledge of
where evacuees are and where they are likely to go to inform decision mak-
ing. Furthermore, the system is designed to red-flag minors who are unac-
companied in order that they might be reported to the correct authorities
and agencies before or during evacuation. The system also records evacuees
with particular medical needs, recording whether they have sufficient medi-
cation and whether they require a caregiver.
In aggregate, NMETS is composed of both manual and computer-based
elements, functioning through paper records, but also “the Internet, hand-
held scanners, laptop computers, tethered USB scanners and/or RFID read-
ers” (FEMA, 2012). NMETS is a technological assemblage that addresses
the evacuee, their body, identity, and correspondent records in a variety
of ways and through a variety of pinch points. The system ultimately tries
to correct the mistakes of Katrina in a way that imagines and records the
evacuee as a potentially mobile body which is not abstracted from the things
and relations they need.
Although FEMA pioneered this national response to the disaster, pro-
grammes like NMETS were also being implemented at lesser scales, and
perhaps for more dubious reasons. Of all the states that received evacu-
ees from New Orleans and the wider Louisiana area, Texas was foremost
amongst them, receiving more than 250,000 persons, mostly in its largest
city, Houston (Godoy, 2006). But the benevolence of the state and city was
rewarded, at least in the opinion of some, by an increase in crime and anti-
social behaviour, with police officers reporting a spike in emergency calls in
those areas densely populated by evacuees, and the latter victims or suspects
in about 20% of the city’s murders, despite being less than 10% of the (new,
total) population (this was reflected in mainstream national and local news
CNN, 2006). As a result,

Many of the evacuees who stayed on in Houston found the welcome

mat quickly pulled . . . blamed for a rising crime rate, especially mur-
ders, and accused of playing the system—having their housing paid by
the government while making little effort to find a job.
(Buchanan, 2010)
 236 Peter Adey and Philip Kirby
Evacuation was threatening. Far from progressively destabilising many
bounded senses of place, identity, and security, it seemed to reinforce them
(Harvey, 1996), despite the fact that some studies have found really only
very modest increases in crime (Varano, Schafer, Cancino, Decker and
Greene, 2010).
Many evacuees appeared on the centralised federal databases of those
entitled to housing benefit, and so requiring remuneration and welfare
accordingly. Despite some of its own citizens being affected, the principal
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

‘evacuee problem’ encountered by Texas, and Houston, was the manage-

ment of a neighbouring state’s displaced persons. As Godoy (2006) reflected,
a year after Katrina, “So many evacuees came to Texas and stayed that
the state undertook its own accounting.” Part of this ‘accounting’ was the
development of the Special Needs Evacuation Tracking System (SNETS) in
2007, the country’s putative first statewide citizen-evacuation system, which
would use an array of technologies such as RFID, wireless, and mobile data
to allow evacuees to be located in real time, with private-sector support from
Radiant RFID, AT&T, and Intermedix, amongst others.
At first, as the name suggests, it focussed on vulnerable groups only, espe-
cially those with medical needs, but later this was broadened to include
all segments of the population and the re-designation of the System as the
Emergency Tracking Network (State of Texas, n.d.). We are faced, then,
with the perhaps uncomfortable chain of causality that saw Texas house the
majority of Katrina evacuees, followed by reportedly higher crime rates in
cities such as Houston and the designation of ‘problem evacuees,’ followed
by the state’s adoption of an enhanced system for managing and tracking
evacuees in the future. Whether the second on some level inspired the lat-
ter or whether the first link in this chain directly fed the last is a matter of
speculation, but certainly the creation of the ‘digital evacuee’ remains bound
up with issues of the most basic corporeality of what to do with bodies that
require help and welfare.
Evacuation tracking has its own particular geographies and intensities,
with bodies becoming more visible, physically and digitally, at certain nodes
when moving through certain spaces. It is an assemblage that also works
through multiple scales simultaneously, drawing together information from
national, state, and local levels. A huge amount of work is still required to
make systems at each level interoperable, reconciling the data collected from
feeder points, such as hospitals, with centralised, national databases. Cross-
cutting these geographical concerns are therefore issues of public and private
stewardship over data (Lyon, 2003). In the wake of Katrina and the systems
employed by Texas and FEMA, a host of private companies have emerged
into the nascent evacuee tracking industry. In addition to Radiant RFID,
Motorola and AT&T, who engineered the Texan system, Axcess Interna-
tional Inc., Disaster Management Systems, EVACTrak, Elliot Data Systems,
LINSTAR, Med Media, Midwest Card and ID Solutions, Multicard US,
Salamander Technologies, and Transition Works Software are all developing
 The Digital Evacuee 237
systems that will use RFID tags and other, more complex databases to keep
track of evacuees. If this reflects the proliferation of tracking systems, it is
also the addition of another layer in the increasing overlap between private
and public security policies and solutions. But how will such an overlap
be regulated and legislated? Who will maintain and service such systems,
retrieve and store the information that they provide? While in one instance
in Texas a federal judge was brought in to arbitrate the dispute behind a
school’s student tracking system on religious beliefs, such individual atten-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

tion may not be possible as evacuee tracking systems multiply, expand, and
institutionalise themselves into the apparatuses of state and federal gover-
nance over emergencies.
In the wake of these developments, the evolving inter-disciplinary fields
of crisis and disaster informatics are providing other kinds of avenues for
tracking potential evacuees and the displaced through more advanced forms
of the collection and analysis of big data and social media networks. Crowd-
sourcing community sentiment, for example, can provide more effective
early warning of emerging threats or real-time disasters such as an earth-
quake quicker than the seismographs can transmit information (Hughes and
Palen, 2009). They can also prove effective ways to establish where people
are who might in need of evacuation, or, for example, to assess the welfare
and well-being of evacuees once relocated. Unsurprisingly, this is not being
led solely by the state. Civil society and activist groups such as Geeks without
Bounds have proven the value of correlating information gathered through
social media with geographical information system (GIS) –generated maps
in Hurricane Sandy. But the newfound accessibility of people in emergencies
through mobile technology can run up some different sets of questions. For
example, Buscher (2013) has shown that the civilians who could not evacu-
ate Utøya Island in Norway in 2011 became visible to the shooter Anders
Behring Breivik because friends and relatives rang people’s mobile phones
who were hiding. Being locatable, in other words, can have some unlikely
but devastating effects in particular emergencies.
The vulnerability of such bodies may well then be something constructed
by the assemblage of digital technologies, if and when they enter it, as well as
contemporary evacuee tracking systems. After Katrina, many displaced aliens
avoided seeking government assistance, and certainly government tracking
systems, because of their fears that the latter were being yoked together with
the Department of Homeland Security’s immigration databases (see Lindsay,
2011). In this instance, the state was already exercising control over the
mobility and identification of individuals. The state claimed a stake in their
future mobility by arrest, detention, even deportation should they be found
as non-citizens. So we are left with the familiar question (albeit in different
contexts) of what about those people more broadly that, even in emergen-
cies, simply do not wish to be tracked and subjected to the scrutiny that
comes with technologies of address? What if we object to the marking or
‘tagging’ of the body (Aas, 2006) through not only the concerns of privacy
 238 Peter Adey and Philip Kirby
and civil liberties advocates but under religious freedoms. These questions
portend to the issues with which evacuee tracking will need to contend in
the future, especially if more far-reaching and compulsory evacuee tracking
systems are promulgated going forward.
As detailed, the onus on evacuation planning appears to be to harmonise
and join up previously distributed but siloed and fragmented databases and
disparate registration systems and piloted tracking programmes held within
a multitude of private providers, contractors, and local and federal govern-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ment services. In this sense, the digital evacuee, like the passenger, finds itself
mobile and interoperable across numerous systems of records that are being
made to speak to each other in an albeit piecemeal fashion. In the future,
for example, might it be possible for an airline reservation system to serve
as a digital feeder point for a medical evacuee being transported through
the secured, ordered mobilities of the airport? And how might this integrate
with the myriad other non-governmental actors that provide assistance dur-
ing emergencies, like the shelter registration programmes of the American
Red Cross? To date, the privacy implications of such emergency manage-
ment information systems (EMIS) have been given only limited attention,
but as these systems proliferate and the technology behind them evolves,
such concerns promise to move toward the forefront of critical agendas.
For Buscher, such practices may mean the obstruction or rendering “impos-
sible lived practices of privacy boundary management.” We also don’t have
to look to the United States to find examples of the potential implications of
these practices or, conversely, when regulation intended to protect personal
data is misapplied or abused. For instance, some of the very criticisms of the
response effort to the 2005 7/7 attacks on London were the ways in which
the 1998 Data Protection Act was applied in some cases overzealously that
obstructed the efficient sharing of information about individuals between
different emergency responders and organisations (Armstrong, Ashton and
Thomas, 2007; Buscher, 2013).

CONCLUSION

It is important not to get ahead of ourselves. It is not all that clear how many
of these systems are currently active and how many are dormant, awaiting
the federal and/or state contracts that will warrant their implementation.
At present, most evacuation systems, outside FEMA’s and Texas’ develop-
ments, draw on the kinds of hospital and prison technologies outlined previ-
ously, including low-tech tagging, to make an evacuee visible and readable
via wristband and simple paper records to store their pertinent details and
information. The material cultures of address of the past may be slated for
supersession, but they are still required in current tracking methods. Indeed,
the material dimension of tracking will never be completely lost, in that
chips, cards, and other devices that can be read remotely will continue to be
 The Digital Evacuee 239
a part of the evacuee tracking assemblage, even if the records to which such
materials pertain may increasingly become a part of the digital, rather than
physical, landscape.
But even if the digital evacuee can ever perform the automated, interop-
erable, and joined-up dream that seems embedded within the projects of
emergency planners, policy makers, and private industry, the contradiction
always facing these bureaucracies and systems is how the evacuees they are
seeking to address through processes of identification, tracking, and trac-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ing evade the records and representations that stand in for them. It is not
simply because the digital double or fragment of the evacuee is somehow
incomplete. Rather, as Edkins has explored, there is a certain homology
between “systems of registration, identification, and control and a process of
objectification—the production of people as nothing but objects of adminis-
tration” (2011, 7). In other words, for Edkins (2011), and as we’ve seen in
evacuation, the intractable ‘messy’ person, with all their idiosyncrasies, their
choices, their motivations, their friendships, and their complex associations
and loves, ‘gets in the way.’

NOTE

1. Municipal Court of Louisiana, 2007, En Banc Order, ACLU.

REFERENCES

ACLU. (2006) Abandoned and Abused: Orleans Parish Prisoners in the Wake of
Hurricane Katrina. New York: ACLU, National Prisons Project.
AAS, K. F. (2006) “The Body Does Not Lie”: Identity, Risk and Trust in Technocul-
ture. Crime, Media, Culture, 2(2). p. 143–158.
AAS, K. F. (2011) “Crimmigrant” Bodies and Bona Fide Travelers: Surveillance, Citi-
zenship and Global Governance. Theoretical Criminology, 15(3). p. 331–346.
ADEY, P. (2010) Aerial Life: Spaces, Mobilities, Affects. Oxford: Wiley.
AGAMBEN, G. (1998) State of Exception. Chicago: University of Chicago Press.
AGUIAR, M. (2011) Tracking Modernity: India’s Railway and the Culture of Mobil-
ity. Minneapolis: University of Minnesota Press.
ARLIKATTI, S., KENDRA, J. & CLARK, N. A. (2012) Challenges for Multi-sector
Organizations in Tracking and Sheltering Registered Sex Offenders in Disasters.
Journal of Homeland Security and Emergency Management, 9(1). p. 1547–1573.
BARNES, T. J., & FARISH, M. (2006). Between Regions: Science, Militarism, and
American Geography from World War to Cold War. Annals of the Association of
American Geographers, 96 (4). p. 807–826.
BARTLING, H. (2006) Suburbia, Mobility, and Urban Calamities. Space and Cul-
ture, 9(1). p. 60–62.
BRANDENBURG, M. A., WATKINS, S.M, BRANDENBURG, K.L., SCHIECHE, C.
(2007) Operation Child-ID: Reunifying Children with Their Guardians after Hur-
ricane Katrina. Disasters, 31(3). p. 277–287.
BUCHANAN, M. (2010) Katrina Evacuees Found Welcome and Worries in Texas. BBC
News Online. Available from www.bbc.co.uk/news/world-us-canada-11117607
[Accessed 23–11–2014]
 240 Peter Adey and Philip Kirby
BUSCHER, M. (2013) A New Manhattan Project?: Interoperability and Ethics in
Emergency Response Systems of Systems. ISCRAM Conference, Baden-Baden,
Germany, 12–15 May 2013.
COLLIER, S. J. & LAKOFF, A. (2008) Distributed Preparedness: The Spatial Logic
of Domestic Security in The United States. Environment and Planning. D, Society
and Space, 26(1). p. 7.
CRESSWELL, T. (2006) On the Move: Mobility in the Modern Western World. New
York: Routledge.
CRESSWELL, T., BERGMANN, S. & SAGER, T. (2008) Understanding Mobility
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Holistically: The Case of Hurricane Katrina. In Bergmann S. & Sager, T. (eds.)

The Ethics of Mobilities: Rethinking Place, Exclusion, Freedom and Environ-
ment. London: Ashgate, 129–140.
CUTTER, S. & BARNES, K. (1982) Evacuation Behavior and Three Mile Island.
Disasters, 6(2). p. 116–124.
EDKINS, J. (2011) Missing: Persons and Politics. New York: Cornell University
Press.
FEMA. (2006) A Federal Response to Hurricane Katrina: Lessons Learned. Wash-
ington (DC): FEMA
FEMA. (2010) National Mass Evacuation Tracking Systems. Washington (DC): FEMA.
FEMA. (2012) Individual assistance: National Mass Evacuation Tracking System.
Washington (DC): FEMA.
FRANCO, C. et al. (2006) Systemic Collapse: Medical Care in the Aftermath of
Hurricane Katrina. Biosecurity and Bioterrorism: Biodefense Strategy, Practice,
and Science, 4(2). p. 135–146.
GHAMARI-TABRIZI, S. (2009) The Worlds of Herman Kahn: The Intuitive Science
of Thermonuclear War. Cambridge, MA.: Harvard University Press.
GODOY, M. (2006) Katrina: One Year Later: Tracking the Katrina Diaspora: A
Tricky Task NPR News: Available from http://www.npr.org/news/specials/katrina/
oneyearlater/diaspora/ [Accessed 12–11–2013]
GRAHAM, S. (2005) Cities under Siege: Katrina and The Politics of Metropolitan
America. Social Science Research Council: Understanding Katrina: Perspectives from
the Social Sciences. Available from http://forums.ssrc.org/understandingkatrina/
author/sgraham/ [Accessed 12–11–2013]
GROEN, J. A. & POLIVKA, A. E. (2008) Hurricane Katrina Evacuees: Who They
Are, Where They Are, and How They Are Faring. Monthly Labor Review.
33(March). p. 32–51.
GUIDRY, G. G. (2009) Louisiana Judiciary: In the Wake of Destruction. The LA Law
Review, 70. p. 1145.
HANNAM, K., SHELLER, M. & URRY, J. (2006) Editorial: Mobilities, Immobili-
ties and Moorings. Mobilities, 1(1). p. 1–22.
HARVEY, D. (1996) Justice, Nature and the Geography of Difference. London:
Wiley.
HELBING, D., FARKAS, I. J., MOLNAR, P., & VISEK, T. (2002) Simulation of
Pedestrian Crowds in Normal and Evacuation Situations. Pedestrian and Evacu-
ation Dynamics, 21. p. 21–58.
HENDERSON, L. F. (1974) On the Fluid Mechanics of Human Crowd Motion.
Transportation Research, 8(6). p. 509–515.
HOUSTON, J. W. & HOUSTON, J. D. (2002) Farewell to Manzanar: A True Story
of Japanese American Experience During and After the World War II Internment.
New York: Houghton Mifflin.
HUGHES, A. L. & PALEN, L. (2009) Twitter Adoption and Use in Mass Conver-
gence and Emergency Events. International Journal of Emergency Management,
6(3). p. 248–260.
 The Digital Evacuee 241
KLEIN, N. & SMITH, N. (2008) The Shock Doctrine: A Discussion. Environment
and Planning. D, Society and Space, 26(4). p. 582.
KNABB, R. D., RHOME, J. R. & BROWN, D. P. (2006) Tropical Cyclone Report:
Hurricane Katrina. National Hurricane Center.
LAKOFF, A. (2007) Preparing for the Next Emergency. Public Culture, 19(2). p. 247.
LINDSAY, B. R. (2011) Federal Evacuation Policy: Issues for Congress. Washington
(DC): DIANE Publishing.
LYON, D. (2003) Surveillance After September 11. London: Polity.
LYON, D. (2007) Surveillance Studies: An Overview. Cambridge (UK), Malden
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

(MA): Polity.
METZGER, P. R. (2006) Doing Katrina Time. Tul. L. Rev., 81. p. 1175.
MONAHAN, T. (2007) “War Rooms” of the Street: Surveillance Practices in Trans-
portation Control Centers. The Communication Review, 10(4) p. 367–389.
NG, W. L. (2002) Japanese American Internment During World War II: A History
and Reference Guide. Westport (CT): Greenwood Press.
O’BRIEN, T. H. (1955) Civil Defence. London: H. M. Stationery Off.
OKUBO, M. (1946) Citizen 13660. Seattle, Washington: University of Washington
Press.
OPHIR, A. (2007) The Two-State Solution: Providence and Catastrophe. Theoreti-
cal Inquiries in Law, 8(1). p. 117–160.
ORR, J. (2006) Panic Diaries: A Genealogy of Panic Disorder. Durham (NC): Duke
University Press.
PACKER, J. (2008) Mobility without Mayhem: Safety, Cars, and Citizenship. Dur-
ham (NC): Duke University Press.
PATE, B. L. (2008) Identifying and Tracking Disaster Victims: State-of-the-art Tech-
nology Review. Family & Community Health, 31(1). p. 23–34.
SCOTT, J. C. (1999) Seeing Like a State: How Certain Schemes to Improve the
Human Condition Have Failed. Cambridge (MA): Yale University Press.
SHELLER, M. (2013) The Islanding Effect: Post-Disaster Mobility Systems and
Humanitarian Logistics in Haiti. Cultural Geographies, 20(2). p. 185–204.
SMITH, N. (2006) There’s No Such Thing As a Natural Disaster. Understanding
Katrina: Perspectives from the Social Sciences, 11.
STROPE, W. E., DEVANEY, J. F. & NEHNĚVAJSA, J. (1975) Importance of Pre-
paratory Measures in Disaster Evacuations. Stanford (CA): Stanford Research
Institute.
THRIFT, N. (2004) Movement-Space: The Changing Domain of Thinking Resulting
from the Development of New Kinds of Spatial Awareness. Economy and Society,
33(4). p. 582–604.
URRY, J. (2007) Mobilities. London: Polity.
VARANO, S. P., SCHAFER, J. A., CANCINO, J. M., DECKER, S. H., & GREENE,
J. R. (2010) A Tale of Three Cities: Crime and Displacement after Hurricane
Katrina. Journal of Criminal Justice, 38(1). p. 42–50.
 12 The Datafication of Mobility and
Migration Management
The Mediating State and Its Consequences
Dennis Broeders and Huub Dijstelbloem
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

MONITORING AND DATAFYING HUMAN MOBILITY

Modern technologies have increased the possibilities for governments to

gather and process data and this has increased the variety and the depth of
governmental observation and monitoring. This variety includes the data
from new technologies such as radar, infrared, and satellite technology (‘the
eye in the sky’) that allow for different forms of observation and detection,
while depth can be added through technologies such as ICT, biometrics, GIS
technology, and statistical risk calculation. The more recent development
of big data analysis is now also finding its way into public policy mak-
ing. The state’s perception of reality thus becomes more technologically and
statistically mediated and ‘datafied.’ Data of various types and sources are
processed, combined, and connected through networked databases. Even
though policymakers often claim that technology merely does the same job
faster and better, technology also changes both the substance and the nature
of policy. For one thing, it brings new actors to the scene. Baker (2008)
described the work of what he calls a new class of ‘numerati’ that data
mine vast databases for correlations and use these to plan for the future.
He primarily emphasized the commercial brand of this class, but there are
public counterparts in ever larger numbers in, for example, counterterrorism
(Balzacq, 2008; Monahan and Palmer, 2009), youth care (Schinkel, 2011;
Keymolen and Broeders, 2013), international development (Taylor, 2015;
Taylor and Broeders, 2015), and crisis management (see Adey and Kirby,
this volume). For another, the use of ever bigger datasets necessitates policy
thinking in terms of risks and increasingly favours correlations over causali-
ties (Mayer-Schönberger and Cukier, 2013).
The application of new technologies of surveillance and networked data-
base technologies has been very prominent in the field of the management
of migration and international mobility (Amoore, 2006; Broeders, 2007,
Balzacq, 2008; Dijstelbloem and Meijer, 2011; Johnson et al., 2011). All the
keywords apply: large populations in relation to which policy problems are
increasingly defined in terms of risks (risk of international terrorism, trans-
national crime, and violations of immigration law and policy, i.e. irregular
 Mobility and Migration Management 243
border crossing and settlement, multiple asylum claims, etc.) and an ever-
expanding network of surveillance systems and databases aimed at visualis-
ing, registering, mapping, monitoring, and profiling mobile (sub)populations.
The new technologies analysed in this chapter ‘connect’ the monitoring and
management of human mobility in the territorial borderlands of Europe—
the surveillance of the green and especially the blue borders of the EU—with
the territorially detached manifestations of ‘border surveillance’ that work
through the ‘stretched screens’ (Lyon, 2009) that connect an emerging Euro-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

pean network of public border, immigration, and law enforcement officials

and private ‘unofficials.’ In these stretched screens various sources of data
come together, are weighed according to risk profiles and/or correlations,
and are made part of decision trees (Meijer, 2011; Van der Ploeg and Spren-
kels, 2011; Amoore, 2013).
A number of these developments—that will be elaborated on in the
chapter—resonate with insights from literature on ontological perspectives
on the relationships between politics and technology. The central role for
data, calculus and risk assessment in the management of mobility resembles
Latour's concept of the ‘centre of calculation’ (Latour, 1987, 1993). The
integration of various kinds of new technologies and data sources makes the
question what the role of ‘non-human actors’ in this field of policy entails
a very relevant one (e.g. Dijstelbloem and Broeders, 2015). Moreover, these
developments may also cast a different light on the role of the state itself, a
subject to which we will turn our attention in the concluding section. The
main question is whether adding a thicker layer of information and data to
policymaking and the remarkable increase in speed and real-time exchange
of data changes the nature of the state and the state’s agency. Therefore, this
chapter aims to show how new technologies shape the policy perception of
problems of migration and mobility and looks at how the role of the state
itself transforms both on a conceptual as well as on an empirical level.
This chapter is structured in the following way. The first section sketches
a theoretical view on the use of technologies and data applied to the matter
of managing migration and mobility at the national and especially the Euro-
pean level. The following section investigates the way in which new tech-
nologies are applied in European border management—especially through
the Eurosur programme—and concludes that the migrant’s environment and
the migration process have become increasingly datafied: all sorts of data
from a myriad of sources are mashed to create a new ‘situational aware-
ness.’ Next, we turn our attention to the way that migrants and travellers
themselves have become datafied. The individual migrant and traveller are
entered into various biometric databases, with an increasing number of data
points to facilitate identification and, in the aggregate, data mining and pro-
filing. Finally, we discuss the consequences of these developments for the
position of the state and the state’s agency. How can we interpret the process
of datafication in migration and mobility management, and how does it
influence the policy perspectives and implementation? How does it influence
 244 Dennis Broeders and Huub Dijstelbloem
the status of the state as an actor in a technological network? What are the
consequences of the state’s role in collecting, displacing and transforming
information for the materialization of legitimate decisions about citizens,
travellers, and migrants?

HUMAN MOBILITY REGISTERED, CATEGORISED,

AND DATAFIED
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The sheer numbers of international mobility and the fast developments in

modern surveillance technologies have changed the way states register, moni-
tor, and deal with human mobility and migration. The nature of the border
has changed drastically and is now ‘everywhere’ (Lyon, 2005), has become
‘virtual’ (Johnson et al., 2011), or has changed into a ‘border security con-
tinuum’ (Vaughan-Williams, 2010). New technologies of surveillance and the
digitisation of information have drastically enhanced, intensified, and changed
immigration and border policies. Many new technologies of visualisation,
detection, and registration have been added to the policy instruments that
governments and international organizations use to deal with human mobil-
ity in various places and stages of the migratory process (Düvell and Vollmer,
2011). Some technologies only yield direct operational information—think
of infra-red or carbon dioxide scanners—but most (also) generate data on
patterns of mobility and characteristics of the mobile populations. So eventu-
ally, most information, irrespective of its original technological nature ends
up in the great information equalizer that is the database. Data from various
different sources are broken down into bits and bytes and reassembled in
databases that have become a cornerstone of modern migration policy, which
is organized and executed through the ‘stretched screens’ (Lyon, 2009) of the
immigration official. Increasingly the Internet has become the ‘backbone of
backbones’ (Choucri, 2012, 151) that ties the various sources together and
makes them accessible worldwide and in real time. Some of these new tech-
nologies then are very different from earlier forms of border and migration
control. Data management, the calculation of risk factors, and the global use
of biometric databases, profile, and digital watch lists that are able to link
border control posts, embassies, consulates and field officers in refugee situ-
ations with and through the digital back offices of mobility monitoring and
management. Whereas before migration officials primarily relied on their
tacit knowledge—who is taken out of the queue for closer inspection—and
a necessarily very limited ‘watch list of people to look out for’ in their direct
surveillance of border flows, the gathering and analysis of data and the appli-
cation of new technologies in passports and visas—chips and biometrics—
for example, have shifted their attention in part to data flows rather than
the actual moving bodies. More and more data is informing policy making
and feeds into policy execution. Monitoring mobility and migration have
changed drastically in the digital age.
 Mobility and Migration Management 245
As a result, the modern-day problem of immigration authorities lies in the
governance of international mobility. According to Koslowski (2008, 105),
“migration [is] not the ‘new security issue’; it is increasing global mobility,
which is primarily tourism and business travel.” Indeed, the volume of inter-
national mobility is staggering (see Table 12.1). The number of people that
annually pass the borders of the average OECD country is vast and dwarfs
the numbers that are covered under the subheading of ‘immigration.’ More-
over, the risks and problems that haunt policy makers today—international
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

terrorism, illegal migration, and transnational crime—are much more likely

to be enveloped in the flow of human mobility than in the much smaller and
more tightly regulated sub-flow of immigration and settlement.
Mobility has in the past decades become an integral part of modern life.
Cresswell (2010, 27) places mobility at ‘the heart of what it is to be mod-
ern.’ At the same time, mobility is not for everyone: There is a politics of
mobility in which the differential distribution of mobility produces some
of the starkest differences today (Sparke, 2006; Cresswell, 2010). In short,
international mobility is about (international) power relations and about
separating the privileged from the non-privileged (Bauman, 1998). Lyon
(2003, 142) states that surveillance has always been about ‘social sorting’,
about the classification of populations as a precursor to differential treat-
ment. Whether this is done using traditional technologies and techniques—
such as ‘classical’ statistics—or through the use of new technologies and
surveillance systems, the aim remains social sorting. Lyon (2007, 3) also
noted that surveillance always moves somewhere on the continuum between
care and control. Surveillance, visualisation, and categorisation are not a
priori negative or positive for those subjected to them. Their effects are to
a large part dependent on the policy aims that are underlying surveillance,
although it will also be argued in this paper that the use of modern tech-
nologies alters policies themselves too. Technological instruments are, like

Table 12.1 International tourist arrivals 1995–2009 (in millions)

1995 2000 2005 2007 2008 2009

World 533 683 802 901 919 880
• Advanced economies 339 423 451 496 494 470
• Emerging economies 194 260 351 405 425 410
United States 43.5 51.2 49.2 56 57.9 54.9
Europe 309.1 392.2 441 485.4 487.2 459.7
• France 60 77.2 75.9 80.8 79.2 74.2
• Spain 34.9 47.9 55.9 58.7 57.2 52.2
• Italy 31.1 41.2 36.5 43.7 42.7 43.2
• United Kingdom 23.5 25.2 28 30.9 30.1 28

Source: UNWTO 2009 and 2010.

 246 Dennis Broeders and Huub Dijstelbloem
other policy instruments, instruments of power. For states the monitoring
of international mobility aims to ‘filter out’ those that desperately wish to
remain unseen or, at the very least, to remain unidentified: terrorists, irreg-
ular migrants, and criminals. But it also serves to locate, categorize, and
identify those that need to be found, recognized, and categorized in order to
protect and aid populations such as refugees and internally displaced per-
sons. Both categorization—identifying people as part of a group, or at least
ascribing membership to them—and identification are vital for the digital
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

monitoring and management of human mobility.

For one thing, risk, categorisation, and databases determine the options
and treatment of travellers and migrants. According to Rose (2001, 7), risk
can be understood as a “family of ways of thinking and acting, involving
calculations about probable futures in the present followed by interventions
into the present in order to control that potential future.” Defining and cal-
culating risks then determines, literally, how the state will watch its subjects,
migrants, and visitors, what it will see and what it will ‘think’ of that. Risk
categories, Jones (2009, 177) argues on the basis of Foucault, create power
by “obscuring difference by forcing the multiple into manageable units (cat-
egories) with solid separations (boundaries) between them.” Today, with
risk calculation, categorisation, and networked databases taking centre
stage in the monitoring and management of human mobility, the creation
and the operational use of new digitally constructed facts has taken a quan-
tum leap. Surveillance is about tracking flows and in these technologically
mediated times, “the whole process shifts from being focused on persons to
being focused on codes” (Brighenti, 2007 337).
In this respect, the state’s organizations that are responsible for execut-
ing these policies resemble what Latour (1987, 1993) has called ‘centers of
calculation.’ Centers of calculation are specific spots where experiments,
expeditions, and data collection take place and where observations, infor-
mation and knowledge are combined. They can be any site, a laboratory, a
statistical institution, a data bank, and so forth (1999, 304). Characteristic
to these sites is that the objects under study—in the case of border manage-
ment, for example, travellers and migrants, but also routes, intermediaries,
and various contexts—are represented into a sign, an archive, a document,
a piece of paper, a trace (1999, 306). For example, since 2010 the UK oper-
ates a National Border Targeting Centre (as part of its eBorders programme
to analyse passenger data on a number of high-risk routes; see Broeders and
Hampshire, 2013). In the United States the Department of Homeland Secu-
rity has been setting up a network of ‘data fusion centres’ that are primarily
concerned with data flows connected to national security, though this has
been known to be affected by function creep (Monahan and Palmer, 2009).
With new technologies of visualisation and the digitization of migration
policies, the policy process and the implementation of policies change, too.
Digitization does not come without consequences. A new class of policy mak-
ers increasingly makes the most important decisions in terms of migration
 Mobility and Migration Management 247
policy—who fits into what category. Johnson and colleagues (2011, 62)
point out that “the sovereign decisions of the border are as likely to be made
by programmers and mathematicians who write computer code as they are
by uniformed border agents.” In other words, the numerati (Baker, 2008)
that design and construct the digital tools for governments have become
much more important in shaping policy because it falls to them to trans-
late political priorities and decisions into the algorithms and decision trees
that run and organize the data in databases. As a result, the people—Lyon
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

calls them the ‘others’ in the following quote—that are digitally categorized
essentially ebb and flow with changes in algorithms and the settings of the
system. “One difficulty of such others, in current identification regimes, is
that their ranks may expand at will—or whim—through slight statistical
adjustments expressed in the algorithms controlling for entry and eligibility”
(Lyon, 2009, 148). Political and policy decisions about risks and categoriza-
tions and changes therein cause ripples in the mass of data that is collected
and stored about travellers and migrants once they are translated into the
software. Categories do indeed have politics. Aradau and associates (2008,
152; see also Salter, 2008) underscore the consequences of risk-based data-
base systems: “‘Who decides?’ is supplemented by ‘Who gets to imagine the
future?’ The imagination of the future has become one of the main political
stakes.”

MONITORING THE BORDERLANDS OF EUROPE

The different modalities of today’s borders intervene in human mobility in

complex ways. Even though borders are now ‘everywhere,’ there are still
lines on the map that separate countries from each other and are patrolled by
border guards. The—at places—heavily fenced and patrolled border between
Mexico and the US resembles parts of the EU’s external border, both in its
selectiveness of enforcement and in its appearance characterised by mili-
tarised fences and watchtowers. In the ‘Schengen land’ without internal
borders the emphasis has shifted to the communal external borders. The
basic data on these borders are daunting: 27 member states are responsible
for roughly 10,000 kilometres of green borders, 50,000 kilometres of blue
borders, and 1,800 official ports of entry into the EU (Hobbing, 2010).
Unsurprisingly, guarding the border has increasingly become a technological
affair. Much has been invested in new technologies of border surveillance,
driven by both political aspirations to ‘control’ borders and by the ‘home-
land security industry’ to create and expand new markets for surveillance
technology (Hayes, 2009; Broeders, 2011b; Düvell and Vollmer, 2011).
The green Eastern borders of the EU combine a mix of traditional border
patrol and technology (Düvell and Vollmer, 2011). At border crossings heart-
beat and mobile carbon dioxide detectors are used to detect irregular immi-
grants hidden in carriers or cargo. Also, border controls are supplemented
 248 Dennis Broeders and Huub Dijstelbloem
by fixed and mobile document examination systems, CCTV, night-vision
equipment, thermal cameras, and movement detectors. At the blue borders
of the Mediterranean, new technologies of visualisation are set to play an
even greater role and include the use of satellites, drones (UAVs), and the
data mining of various sources of public and operational data (Hayes and
Vermeulen, 2012). Especially through the FRONTEX agency and pro-
grammes such as Eurosur, the EU aims to transform the Mediterranean into
a digital moat shielding the south of the Union against illegal immigration
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

especially. The dangerous water barrier between North Africa and Southern
Europe has attracted many irregular migrants to seek entry to the EU under
very harsh conditions and claimed many of their lives (Carling, 2007; Spi-
jkerboer 2007).
Spain, separated from Africa only by the narrow Straits of Gibraltar, was
one of the first EU countries to turn to new surveillance systems to deal with
(illegal) migration. As early as 2002, Spain began to introduce a maritime
surveillance system, the Integrated External Policing System, for its exter-
nal borders (Sistema Integrado de Vigilancia Exterior, SIVE), to deal with
the rise in illegal immigration (Düvell and Vollmer, 2011, 10). Systems like
the Spanish one were precursors to more encompassing ambitions laid out
by the European Commission in 2008. In that year EU Justice and Home
Affairs Commissioner Franco Frattini launched his so-called ‘border pack-
age,’ a vision for the management of the EU’s external borders (Commission
of the European Communities, 2008). This package consisted of three com-
munications: one on the use of database technology, including the intention
to create the EU–wide Entry/Exit System; one on the future of FRONTEX;
and one Communication examining the establishment of a satellite-based
border surveillance system called Eurosur (Guild et al., 2008). Hobbing con-
siders the border package to be a paradigm shift for the EU:

All that seemed of doubtful value before, such as fully automated border
checks, comprehensive systems of entry-exit control, air passenger sur-
veillance and electronic travel authorisation, hi-tech border installations
including virtual fences, has all of a sudden become part of the EU’s
vision for the 21st century.
(Hobbing, 2010, 68)

After the 2011 Arab Spring and the (in retrospect, small) numbers of
irregular migrants heading for Europe that followed those events, the EU
has pushed forward with the plans for Eurosur and the EU’s smart borders.
In terms of the technological management of migration flows the Eurosur
plan is most interesting, as it aims to create an EU–wide mechanism for the
deployment of surveillance systems ‘at or beyond the geographical exter-
nal borders.’ The plan involves ‘pooling’ all sorts of surveillance technolo-
gies that the member states use at the border and should “ultimately lead
to the establishment of a ‘common monitoring and information sharing
 Mobility and Migration Management 249
environment’” under the direction of FRONTEX (EC quoted in Jeandesboz,
2008, 9). The result should be a ‘system of systems.’ So initially, the aim is to
pool national systems such as Spanish SIVE operated by the Guardia Civil
(Carling and Hernandez-Carretero, 2011), for common use. The main aims
are to reduce illegal migration into the EU, to prevent cross-border crime,
and to enhance search and rescue capacity.
Investments in both the hardware of border surveillance as well as in
software and information management systems are steadily growing. The
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

big companies involved in the migration management and security industry

pushing for the introduction of such new technologies have found a recep-
tive partner in the EU. Hayes (2009) speaks of a ‘security-industrial com-
plex’ and documents the extensive links between the industry and the EU’s
research programme, which funds substantive R&D in this area—not in the
least for technologies for migration surveillance such as robotics, unmanned
aerial surveillance, and satellite tracking technology.
But the investments in the data-analytical or ‘informational’ parts of the
Eurosur programme are also growing. The idea and aim is that FRONTEX
will produce all sorts of risk analyses and tap into all data from the vari-
ous surveillance systems to make ‘intelligence-led’ border patrolling possible
(Léonard, 2010) and coordinate efforts at the border. Risk analysis, accord-
ing to FRONTEX’s website, is the starting point for all FRONTEX activi-
ties, from joint operations through training to research studies, and relies on
all possible sources of data:

In order to identify short—medium—and long-term trends, a wealth of

data needs to be gathered and analysed. For this, FRONTEX monitors
the global security environment, especially those political, economic,
social, technological, legal and environmental factors which could affect
border security.1

FRONTEX’s Eurosur programme will have three layers of data: the

events layer—including migratory flows, an operational information layer
showing what member states are doing, and an analysis layer to determine
how risks are assessed and translated into operational responses.2 This
should create a ‘situational awareness’ and an accurate ‘situational picture’
for border guards to work with. The legal definitions for ‘situational aware-
ness’ and other key concepts in the regulation are simultaneously detailed
as well as wide ranging and defined in the broadest possible terms (Hayes
and Vermeulen, 2012). For example, Article 3 (d) of the Eurosur regulation
reads:

‘situational picture’ means a graphical interface to present near-real-

time data and information received from different authorities, sensors,
platforms and other sources, which is shared across communication
and information channels with other authorities in order to achieve
 250 Dennis Broeders and Huub Dijstelbloem
situational awareness and support the reaction capability along the
external borders and the pre-frontier area.
(European Union, 2013, 13)

In other words, the data gathering, mining, and analysis under the flag of
Eurosur is wide ranging and may draw on many and relatively unspecified
sources. Even though this data is subject to the EU regime of data protection
and the protection of fundamental rights, it is unclear how oversight on this
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

data gathering is organised. In the run-up to the European Day for Border
Guards 2013, there was a call to industry and researchers to showcase their
wares on the ‘technology exhibition’ that was held in parallel. In addition
to calls for various kinds of surveillance hardware, FRONTEX was also
very interested in ideas and products of technology for ‘information sharing
and interoperability’ and ‘information acquisition and fusion.’ Especially
the latter category, calling for applications that can deal with “real time data
mining for processing vast amounts of heterogeneous data” and the use of
“new sources of information (on line news and social media) for intelligence
gathering” (FRONTEX, 2013) shows a new interest in the datafication of
border management. This datafication goes well beyond the data gathered
by public officials in migration management themselves and seems set to
take in all accessible information that may prove useful. It is a datafication
of the migrants’ and travellers’ environment in the broadest sense. The fact
that EUROSUR will also hook up with the EU–wide Common Information
Sharing Environment as ‘its border control function’ (Hayes and Vermeulen,
2012) underlines that the data gathered here will also reach stretched screens
across Europe that are not involved in migration management directly.

MOBILITY MANAGEMENT THROUGH ‘STRETCHED

SCREENS’

Where the previous section illustrated how the border environment has
become datafied—the ‘situational awareness’ is pieced together from a myr-
iad of sources—this section turns to the datafication of the individual migrant
and traveller. New technology has made it easier and more convenient for
migration control to move away from the actual border and has partially
become ‘remote control’ (Zolberg, 2003). At the border (see previous sec-
tion), the notion of a border is already stretched by using remote sensing
technology and open sources, but within the procedures regulating mobil-
ity and migration this goes double. The aim is “for borders to increasingly
exist de facto in cyberspace, i.e., become ‘virtual borders’” (Hobbing and
Koslowski, 2009, 14). Mobility management in the digital age is preferably
organised in such a way that decisions on entry and eligibility are taken as
far away from the territorial border as possible. Migrants and travellers are
monitored through data, data mining, profiles, and risk calculation. The vast
 Mobility and Migration Management 251
numbers of the flows of migration and mobility are increasingly matched by
the vast numbers that are registered, categorised, and data mined. Databases
and the ‘screens’ of the terminals of migration officials everywhere shape the
image of migrants and the opportunities they have or lack (Amoore, 2009;
Lyon, 2009).
In recent years many Western countries have been constructing migra-
tion database systems to register and manage especially three categories
of migrants: asylum seekers, migrants on (short-term) visas, and irregular
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

migrants. In the EU this has translated into a new digital border of the EU
consisting initially of the Schengen Information System (SIS) and the Euro-
dac system. In October 2011 the Visa Information System (VIS) became
operational, and in April 2013 the second generation of the Schengen Sys-
tem (SIS II) went live. Plans for an EU–wide Entry/Exit System and smart
border programmes have been on the EU’s agenda since 2008 (Broeders,
2007, 2009, 2011) and gained renewed traction after the migratory move-
ments in the wake of the Arab Spring in 2011 (Hayes and Vermeulen, 2012).
The EU’s investment in these programmes is considerable. The budget for
Home Affairs 2014 to 2020 reserves 822 million euros for existing large
scale IT programmes. The recently created EU Agency for IT in Justice and
Home Affairs and an unknown portion of 4,648 million euros reserved for
internal security funds are earmarked for the development of new large-scale
IT systems (Comission of the European Communities, 2011, 4). These funds
primarily flow towards both privately contracted and the EU’s own techni-
cal experts, developers, and the operators of data systems working with
mobility-related data flows.
These systems all chart and monitor mobility, store vast amounts of per-
sonal data and biometrics and, in the case of the EU systems, can give insight
into the primary and secondary dispersal of migration flows over the mem-
ber states of the EU. For example, the Eurodac system, a central register of
the fingerprints of all asylum applicants in the EU, charts the distribution of
asylum migration across Europe and detects multiple asylum claims in the
EU. Through the registration of multiple claims—someone has filed a claim
for asylum in various member states at various times—the system also gives
some indication of secondary movement. The VIS (Visa Information System)
connects the EU member states’ immigration authorities and consular posts
all over the world. Its central database details the personal and biometric
information of all visa applicants to the EU and the dates on which visas
were applied for, granted, refused, cancelled, withdrawn, or extended. This
means the VIS will generate aggregated information on shifts and trends
in visa-based mobility from certain countries and regions. The Entry/Exit
system, if and when it will be implemented, will be able to generate even
more information on migration processes in the EU as it registers data on
all travellers entering and exiting the EU (however, see Jeandesbosz et al.,
2013, for a critical analysis of the technological and financial feasibility of
the scheme). The main emphasis in all these systems shifts to identification.
 252 Dennis Broeders and Huub Dijstelbloem
All the systems mentioned have become ‘identity technologies’ and revolve
around biometrics—usually fingerprints and/or iris and facial recognition—
as a unique identifier that ties all the data in these systems to one registered
individual (Broeders, 2007). As a result of the use of surveillance technologies
in the adjacent policy terrain of counter-terrorism and mobility, other sys-
tems, data exchanges, and databases such as PNR data exchange, Advance
Passenger Information, and many watch lists, supplement—and sometimes
interact with—the migration surveillance systems (Balzacq, 2008; De Hert
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and Bellanova, 2011). The closer to national security one gets, the bigger
the data sets available, as we have learned from the recent disclosures about
the American PRISM programme (Bigo et al., 2013), and the more various
sources can and will be combined.
With the bulk of data contained in these various systems, categorisation
becomes a prerequisite for a pre-emptive line of sight in migration policy
(Amoore, 2009). Two trends stand out in the development of migration data-
bases. The first is the shift from reactive to pre-emptive databases. Whereas
the SIS, for example, only registered persons who had ‘done’ something
(violated migration law, for example), most other systems register people
because they may do something in the future (people are registered in Euro-
dac because they applied for asylum but also because they may apply in mul-
tiple countries and may become irregular migrants). This explains why new
and proposed systems—such as the EU’s Entry/Exit system in combination
with biometric passports and automated border controls—aim to register
and fingerprint all travellers that will cross the external border: citizens,
visa-free travellers, and those travelling on a visa alike. The digital hay-
stack is collected to pinpoint the needle later. Lyon (2003) calls this ‘dragnet
policies’ in which large quantities of data on as many people as possible are
stored in order to trace an (infinitely smaller) policy-relevant subpopula-
tion. Overall, the pre-emptive governance of mobility through databases
is organised around the logic of blacklisting, green-listing, and grey-listing
the mobile populations that are heading towards the EU (cf. Broeders and
Hampshire, 2013). Those who are blacklisted should be stopped well before
they reach a territorial border, those who are green-listed are fast-tracked for
easy border passage, and those who are grey-listed are labelled ‘suspect’ and
require further scrutiny at the border or, preferably, before embarking on
their journey. The value attributed to the various data points that make up
a category determines the colour and the shades of grey (for a more detailed
analysis, see Broeders and Hampshire, 2013).
The second trend is that categorisation becomes technologically mediated
and datafied. The classic state view on migration and international mobility
is characterised by countries and visa; the citizens of some countries are sus-
pect in terms of the (perceived) risk of illegal migration, refugee flows (per-
ceived to be bogus or real), and forged identity documents. This view was
based on countries and nationalities, and visa policy is the means to regulate.
New technologies build on this classic view and also go beyond it. The EU’s
 Mobility and Migration Management 253
border package, for example, aims to create different flows of people at the
airport on the basis of risk assessment. These flows are organised through
databases and separate systems for registration and passage at the border
itself. The language is important here: The systems and programmes for
fast-track procedures are those for ‘trusted travellers.’ In the original plans
the Commission even used the unfortunately chosen label of ‘bona fide’ trav-
ellers, suggesting that those travelling on a visa were collectively regarded
to be of ‘bad faith’ (Mitsilegas, 2009). The trend in the new systems is to
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

increase the data points on individuals to make more fine-grained analysis

and categorisations possible. The original Schengen Information System
resembled the traditional paper watch lists most closely: The information
that was permitted into this database was limited and almost austere.3 In
comparison the Visa Information System registers a wealth of information
about the applicant himself, his visa-history, and the people or organisa-
tions that support the application. New systems and databases come with
formal policy categorisations but also more informal labels that affect how
groups of travellers and individual travellers are perceived by immigration
officials and those who enact policies ‘on behalf of’ the state such as pri-
vate visa application centres who run the visa procedures for many OECD
countries and carriers such as airlines and shipping companies (Broeders
and Hampshire, 2013). Over the years carrier sanctions have been enacted
and severely tightened in many Western countries (Menz, 2011), increasing
the incentives of airlines to only accept immaculate paperwork and digital
registrations. Just beyond the horizon new technologies for tracking and
tracing migration and mobility are already in play outside the realm of
European migration policy. For example, in situations with refugee flows
and disaster management aid organisations and governments are increas-
ingly using cell phone data (Broeders, 2011b; Taylor, 2015) and other
sources of available data, sometimes even crowd-sourced data (Zook et al.,
2010). Although these innovations will not be integrated into the European
stretched screens anytime soon—though some of these may already be in
the monitoring of the borderlands—the drive to increase the number of
data points in migration policy may in time lead to the integration of new
sources of information.
Mobility management is increasingly operated through the stretched screens
that connect border and immigration officials to a network of national and
international databases that contain data on travellers that have been reg-
istered, categorised, and in some cases ‘tagged’ for further action. In these
databases—and because of the volume of data collected—some categorisa-
tions now move away to some extent from the rather blunt categories of
nationality towards being based on characteristics of (aggregated) individu-
als. De Hert and Bellanova (2011, 1) observe that “Governments once based
the intensity of screening primarily on a traveller’s nationality (‘sorting coun-
tries’), but new sources of information allow them to focus more directly on
personal characteristics (‘sorting individuals’).” Managing migration flows
 254 Dennis Broeders and Huub Dijstelbloem
through the prism of a database then becomes a technologically mediated
combination of individual characteristics and the profiling and data mining
of aggregate individual characteristics. The result is that travellers become
screened in the form of their data doubles (Haggerty and Ericson, 2000),
which are consequently judged to be ‘bona fide’ or ‘mala fide’ and treated
accordingly. Or, as Louise Amoore (2009, 24) writes “the migrant or travel-
ler is both targeted and anticipated—identified via their personal data and
projected forward so that their digital shadow arrives at the border before
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

they do.”

THE DATAFICATION OF MOBILITY AND MIGRATION

MANAGEMENT: CONSEQUENCES FOR THE STATE

An interesting question is whether adding a thicker layer of information and

data to policy making and the remarkable increase in speed and real-time
exchange of data that we have seen in this policy area changes the nature of
the state and the state’s agency. Callon and Latour (1981) have proposed a
certain analogy between science and politics and between the function of the
laboratory and the state. Not unlike laboratories, states consist of various
heterogeneous actors capable of mobilizing each other and forming associa-
tions to execute specific tasks (Latour, 1983). States have to split themselves
up, move towards different locations, and start field projects to test their
own programmes and connect the internal to the external world. As Passoth
and Rowland (2010, 832) conclude, “if the lab is not a place, but a setting
in which problems are scaled up and down, then states are not containers
for political action, but registers of political actors, networks and actions.”
In the migration-control literature similar ideas about the changing nature
of migration governance have been captured under the notion of “remote
control” (Zolberg, 2003) and shifting migration control “upwards, down-
wards and outwards” (Lahav and Guiraudon, 2000).
The analogy with the laboratory envisages the state as a mediator
able to enrol a network of instruments. Here, the state does not mediate
between unhappy spouses; the mediation at hand consists of the creation
and transformation of all kinds of data and information into reliable facts.
These facts are used to make certain risk assessments and support decision-
making processes. This process of mediation has three aspects: place, time,
and meaning.
First, the mediating role of the state is continuously concerned with trans-
ferring data, information, and knowledge from one place to the other. This
gathering of information and coordination between different places does
not take place by simply copying reality and turning it into a representa-
tion. Instead, the “mediation work” (Latour, 1993) can be characterized
as “circulating reference.” As Latour has clarified, “reference” comes from
the Latin referre, “to bring back” (1999, 32). Going out, bringing back
 Mobility and Migration Management 255
and combining all kinds of data is pretty much what the aforementioned
“stretched screens” aim to do. For example, the VIS will connect the EU
member states’ immigration authorities and consular posts all over the
world. Its central database details the personal and biometric information
of all visa applicants to the EU and the dates on which visas were applied for,
granted, refused, cancelled, withdrawn, or extended. In addition, VIS will be
able to generate aggregated information on shifts and trends in visa-based
mobility from certain countries and regions. Moreover, the longstanding
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

and reliable tendency of policy makers to connect various (migration) data-

bases and add more uses for the data—in a process of function or surveil-
lance creep—will make sure that the merging of information from various
sources continues (cf. Prins et al., 2011; Broeders, 2011c). It also means that
new ways have to be found to strengthen what is usually referred to as the
‘interoperability’ of these systems.
Second, the state as consisting of a network of centres of calculation is
increasingly involved with connecting knowledge of the past, the present,
and the future. Risk calculations and categorizations as sketched in the pre-
vious sections link the aggregate past to images of the future in order to deal
with the present. Schinkel (2011, 378) argues that such risk-based registry
systems, “which record and recode the past, are first and foremost relevant
in the present by sketching the contours (under the name of ‘prevention’) of
a statistically constructed future.”
According to him they, in a sense, ‘visualize society’ along lines that
favour governing. In doing so, boundaries become fixed, and categories
become harder and are more and more perceived to be ‘real.’ At the aggre-
gate level, FRONTEX is merging all its surveillance data into a common
monitoring and information-sharing environment and also taps into vari-
ous other sources of (open) data to enrich its information position and add
more and more data to its analyses. The aggregate past and present steer the
present and future operations of the border agency. At the individual level,
many of the migration databases operated by the EU (also) serve to recon-
nect migrants to their migration histories. Unidentified—often irregular—
migrants are traced back to a visa application or asylum claim through the
fingerprints left in the SIS, VIS, or Eurodac in order to help re-identify them
with a view to expulsion (see Broeders, 2009). In these cases the past is liter-
ally catching up with the present.
Third, the mediating role of the state is concerned with transforming
meaning. In order to operate effectively centers of calculation blur the dis-
tinction between the inside and the outside of a laboratory or a state agency
or any other site which gathers and merges information. Laboratories have
consequences outside their walls. The theories that are developed there are
not just deemed applicable in experimental settings but are of influence on
the external world (Latour, 1987). Information collected at different prac-
tices, control posts, and points of entry will have to be standardized and
made communicable to be gathered together and related at specific nodes.
 256 Dennis Broeders and Huub Dijstelbloem
As such, the parts of the outside world that are under surveillance have
to be transformed into a kind of ‘proto-laboratory’ to make information
extracted from different settings compatible.
These three characteristics of mediation processes have severe conse-
quences for the position of the state and the justification of the state’s role
in the datafication of mobility and migration management. Crucial to the
technologies that are deployed is that they aim to deliver data, information,
and facts on the basis of which decisions are to be taken. These are, in some
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

cases, the sovereign decisions of the state on entry and non-entry, eligibility
and non-eligibility. As Latour and Woolgar (1979) have shown, central to
the “construction of scientific facts” is the erasing of traces. Facts are con-
structed in long and messy processes, but the trajectory that leads to their
institutionalization is hardly referred to in scientific publications. Similarly,
Brighenti (2007, 338) notes that “the legibility of social phenomena is often
achieved at the expense of the recognition of their richness.” Once consid-
ered a fact, it is presented as a ‘fait accompli.’ Policy decisions require a
similar kind of unequivocal categorization, not in the least in migration pol-
icy, which is ultimately about inclusion and exclusion. ‘Faits accomplis’ are
crucial to socio-technological practices in which decisions have to be taken.
On a more general level, the erasing of traces can be understood as a process
of ‘purification.’ Modernity, according to Latour (1993), is overwhelmingly
inhibited by ‘hybrids’—marriages between subjects and objects, humans and
humans—as is the case in the socio-technological systems which are used for
mobility management and border control. However, hybrids are likely to be
denied a legitimate place in the modern order. Instead, they are brought back
to a binary position. They are either a subject or an object, not anything in
between. The implication is that the erasing of traces serves to mask the
constructed nature of what are supposed to be clear-cut subjects or objects.
A fact can only speak for itself once it is declared independent of human
actions.
In a similar vein, the state’s information management is aimed at the cre-
ation of undisputable facts. Interestingly, the process of datafication starts
off by vastly increasing the hybridization of the available data and informa-
tion. The size of the intake of data from increasingly varied sources is grow-
ing fast—there is a reason the debate is about big data—making the resulting
data set a fuzzy mashup. However, information networks and ‘centers of
calculation’ then transform bits of information by selecting and connect-
ing them and build a workable image of reality, similar to what Deleuze
described as the creation of virtual date doubles, some of which are consid-
ered ‘green,’ some ‘black,’ and some ‘grey,’ with the latter category being
an transit category—that is, being on its way to become at least temporarily
one of the former. However, there is no natural correspondence between
these representations and the outside world. To make them correspond, the
intermediary steps that connect them must be made invisible in the end. If
messy and hybrid, the constructed nature of fact would come to light and
 Mobility and Migration Management 257
undermine the rational justification of decisions. The mediating role of the
state, it can be concluded, works out in two directions. On the one hand, it
relates, collects, and transforms all kinds of data and information to gather
them together and turn into robust facts. On the other hand, it evades the
traces of this process to purify the outcome and to present it as an objective,
undisputable ground for decision making.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

NOTES

1. See FRONTEX website under the heading of Risk analysis: http://frontex.

europa.eu/intelligence/risk-analysis. Accessed 17 March 2014.
2. Klaus Roesler, Director of Operations division, FRONTEX. Report of Panel
Discussion III—Eurosur and the future of Border Management. European Day
for Border Guards, 23 May 2013, Warsaw, Poland.
3. The data included in SIS were: first name and surname, known aliases, first
initial of the second or middle name, date and place of birth, distinguishing
physical features, gender, nationality, whether the person may be considered to
be armed and/or dangerous, the reasons for the SIS report, and the measures
that need to be taken.

REFERENCES

AMOORE, L. (2006) Biometric Borders: Governing Mobilities in the War on Terror.

Political Geography, 25(3). p. 336–351.
AMOORE, L. (2009) Lines of Sight: On the Visualization of Unknown Futures.
Citizenship Studies, 13(1). p. 17–30.
AMOORE, L. (2013) The Politics of Possibility. Risk and Security Beyond Prob-
ability. Durham (NC) and London: Duke University Press.
ARADAU, C., LOBO-GEURRO, L. & VAN MUNSTER, R. (2008) Security, Tech-
nologies of Risk, and the Political: Guest Editors’ Introduction. Security Dia-
logue, 29(2–3). p. 147–154.
BAKER, S. (2008) The Numerati. Boston/New York: Houghton Mifflin Company.
BALZACQ, T. (2008) The Policy Tools of Securitization. Exchange, EU Foreign and
Interior Policies. Journal of Common Market Studies, 46(1). p. 75–100.
BAUMAN, Z. (1998) Globalization. The Human Consequences. Cambridge: Polity
Press.
BIGO, D. et al. (2013) Open Season for Data Fishing on the Web. The Challenges of
the US PRISM Programme for the EU. CEPS Policy Brief, 293, June 2013
BRIGHENTI, A. (2007) Visibility: A Category for the Social Sciences. Current Soci-
ology, 55(3). p. 323–342.
BROEDERS, D. (2007) The New Digital Borders of Europe: EU Databases and the
Surveillance of Irregular Migrants. International Sociology, 22(1). p. 71–92.
BROEDERS, D. (2009) Breaking Down Anonymity: Digital Surveillance of Irregu-
lar Migrants in Germany and the Netherlands. Amsterdam: Amsterdam Univer-
sity Press.
BROEDERS, D. (2011) A European ‘Border’ Surveillance System under Construc-
tion. In Dijstelbloem, H. & Meijer, A. (eds.) Migration and the New Technologi-
cal Borders of Europe. Houndmills, Basingstoke, and Hampshire (UK): Palgrave,
40–67.
 258 Dennis Broeders and Huub Dijstelbloem
BROEDERS, D. & HAMPSHIRE, J. (2013) Dreaming of Seamless Borders: ICTs
and the Pre-Emptive Governance of Mobility in Europe. Journal of Ethnic and
Migration Studies, 39(8). p. 1201–1218.
CALLON, M. & LATOUR, B. (1981) Unscrewing the Big Leviathan: How Actors
Macro-structure Reality and How Sociologists Help Them To Do So. In Knorr
Cetina, K. & Cicourel, A. V. (eds.) Advances in Social Theory and Methodology.
Toward an Integration of Micro- and Macro-Sociologies. London: Routledge &
Kegan Paul, 277–303.
CARLING, J. (2007) Migration Control and Migrant Fatalities at the Spanish-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

African Borders. International Migration Review, 41(2). p. 316–343.

CARLING, J. & HERNÁNDEZ-CARRETERO, M. (2011) Protecting Europe and
Protecting Migrants? Strategies for Managing Unauthorised Migration from
Africa. The British Journal of Politics and International Relations, 13. p. 42–58.
CHOUCRI, N. (2012) Cyberpolitics in International Relations. Cambridge: The
MIT Press.
COMMISSION OF THE EUROPEAN COMMUNITIES. (2008) Communication
from the Commission to the European Parliament, the Council, the European
Economic and Social Committee and the Committee of the Regions. Preparing
the Next Steps in the Border Management in the European Union. Com (2008)
69 final. Brussels, 13.2.2008.
COMMISSION OF THE EUROPEAN COMMUNITIES. (2011) Building an Open
and Secure Europe: The Home Affairs Budget for 2014–2020. COM(2011) 749
final, Brussels, 15.11.2011.
CRESSWELL, T. (2010) Towards a Politics of Mobility. Environment and Planning
D: Society and Space, 28. p. 17–31.
DE HERT, P. & BELLANOVA, R. (2011) Transatlantic Cooperation on Traveler’s
Data Processing: From Sorting Countries to Sorting Individuals. Washington
(DC): Migration Policy Institute.
DIJSTELBLOEM, H. & BROEDERS, D. (2015) Border Surveillance, Mobility Man-
agement and the Shaping of Non-Publics in Europe. European Journal of Social
Theory. 18(1), p. 21–38.
DIJSTELBLOEM, H. & MEIJER, A. (eds.) (2011) Migration and the New Tech-
nological Borders of Europe. Houndmills, Basingstoke, and Hampshire (UK):
Palgrave.
DÜVELL, F. & VOLLMER, B. (2011) European Security Challenges. Background
paper ‘Improving EU and US Immigration Systems,’ 2011/01, Robert Schuman
Centre for Advanced Studies, San Domenico di Fiesole (FI): European University
Institute.
EUROPEAN UNION. (2013) Regulation of the European Parliament and of the
Council Establishing the European Border Surveillance System (EUROSUR).
PE-CONS 56/13. Brussels, 11.10.2013
FRONTEX. (2013) Invitation to Industry/Researchers to Showcase during the Euro-
pean Day for Border Guards, 28.1.2013. Available from: www.frontex.europa.
eu/news/invitation-to-industry-researchers-to-showcase-during-the-european-
day-for-border-guards-m4VuuC. [Accessed 19.8.2013]
GUILD, E., CARRERA, S. & GEYER, F. (2008) The Commission’s New Border
Package. Does It Take Us One Step Closer to a ‘Cyber-Fortress Europe?’ CEPS
Policy Brief, nr. 154, March 2008.
HAGGERTY, K. & ERICSON, R. (2000) The Surveillant Assemblage. British Jour-
nal of Sociology, 51(4). p. 605–622.
HAYES, B. (2009) Neoconopticon: The EU Security-Industrial Complex. Transna-
tional Institute/Statewatch.
HAYES, B. & VERMEULEN, M. (2012) Borderline: The EU’s New Border Surveil-
lance Initiatives. Assessing the Costs and Fundamental Rights Implications of
EUROSUR and the ‘Smart Borders’ Proposals. Berlin: Heinrich Böll Foundation.
 Mobility and Migration Management 259
HOBBING, P. (2010) The Management of the EU’s External Borders. From the Cus-
toms Union to FRONTEX and E-Borders. In Guild, E., Carrera, S. & Eggenschwiler,
A. (eds.) The Area of Freedom, Security and Justice Ten Years On. Successes and
Future Challenges under the Stockholm Programme. Brussels: CEPS, 63–72.
HOBBING, P. & KOSLOWSKI, R. (2009) The Tools Called to Support the ‘Deliv-
ery’ of Freedom, Security and Justice: A Comparison of Border Security System
in the EU and in the US. Ad Hoc Briefing Paper. Brussels: European Parliament,
Directorate-General Internal Policies, Policy Department C, Citizens’ Rights and
Constitutional Affairs, Committee on Civil Liberties, Justice and Home Affairs,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

PE 410.681, February 2009.

JEANDESBOZ, J. (2008) Reinforcing the Surveillance of EU Borders. The Future
Development of FRONTEX and EUROSUR. Challenge Research Paper 11.
Brussels: CEPS.
JEANDESBOZ, J. et al. (2013) The Commission’s Legislative Proposals on Smart
Borders: Their Feasibility and Costs. A Study for: Directorate General for Inter-
nal Policies. Policy Department C: Citizens’ Rights and Constitutional Affairs.
Brussels: EU.
JONES, R. (2009) Categories, Borders and Boundaries, Progress in Human Geog-
raphy, 33(2). p. 174–189.
JOHNSON, C. et al. (2011) Interventions on Rethinking “The Border” in Border
Studies. Political Geography, 30(2). p. 61–69.
KEYMOLEN, E. & BROEDERS, D. (2013) Innocence Lost? Care and Control in
Dutch Digital Youth Care. The British Journal of Social Work, 43(1). p. 41–63.
KOSLOWSKI, R. (2008) Global Mobility and the Quest for an International Migra-
tion Regime. In Chamie, J. & Dall’Oglio, L. (eds.) International Migration and
Development: Continuing the Dialogue: Legal and Policy Perspectives. Geneva:
International Organization for Migration, 103–143.
LAHAV, G. & GUIRAUDON, V. (2000) Comparative Perspectives on Border Con-
trol: Away from The Border and outside The State. In Andreas, P. & Snyder, T.
(eds.) The Wall around the West. State Borders and Immigration Controls in North
America and Europe. Lanham (MD): Rowman and Littlefield Publishers, 55–77.
LATOUR, B. (1983) Give me a Laboratory and I will Raise the World. In Knorr-
Cetina, K. & Mulkay, M. (eds.) Science Observed. Perspectives on the Social
Study of Science. London: SAGE, 141–170.
LATOUR, B. (1987) Science in Action. Cambridge (MA): Harvard University Press.
LATOUR, B. (1993) We Have Never Been Modern. Cambridge (MA): Harvard Uni-
versity Press.
LATOUR, B. and WOOLGAR, S. (1979) Laboratory Life: The Construction of Sci-
entific Facts. Beverly Hills: Sage.
LÉONARD, S. (2010) EU Border Security and Migration into the European Union:
FRONTEX and Securitisation through Practices. European Security, 19(2).
p. 231–254.
LYON, D. (2003) Surveillance after September 11. Cambridge: Polity Press.
LYON, D. (2005) The Border Is Everywhere: ID Cards, Surveillance and the Other.
In Zureik, E. & Salter, M. (eds.) Global Surveillance and Policing. Borders, Secu-
rity, Identity. Cullompton (UK): Willan Publishing, pp. 66–82.
LYON, D. (2007) Surveillance Studies. An Overview. Cambridge: Polity Press.
LYON, D. (2009) Identifying Citizens. ID Cards as Surveillance. Cambridge: Polity
Press.
MAYER-SCHÖNBERGER, V. & CUKIER, K. (2013) Big Data. A Revolution That
Will Transform How We Live, Work and Think. Boston/New York: Houghton
Mifflin Harcourt.
MEIJER, A. (2011) Migration Technology and Public Responsibility. In Dijstel-
bloem, H. & Meijer, A. (eds.) Migration and the New Technological Borders of
Europe. Houndmills, Basingstoke and Hampshire (UK): Palgrave, 105–133.
 260 Dennis Broeders and Huub Dijstelbloem
MENZ, G. (2011) Neo-liberalism, Privatization and the Outsourcing of Migration
Management: A Five-Country Comparison. Competition and Change, 15(2).
p. 116–135.
MITSILEGAS, V. (2009) The Borders Paradox. The Surveillance of Movement in a
Union without Internal Frontiers. In Lindahl, H. (ed.) A Right to Inclusion and
Exclusion? Normative Faultlines of the EU’s Area of Freedom, Security and Jus-
tice. Oxford: Hart Publishing, 33–64.
MONAHAN, T. & PALMER, N. (2009) The Emerging Politics of DHS Fusion Cen-
ters. Security Dialogue, 40(6). p. 617–636.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

PASSOT, J. H. & ROWLAND, N. J. (2010) Actor-Network State: Integrating Actor-

Network Theory and State Theory. International Sociology, 25(6). p. 818–841.
PRINS, C. et al. (2011) iGovernment. Amsterdam: Amsterdam University Press.
ROSE, N. (2001) The Politics of Life Itself. Theory, Culture and Society, 18(6).
p. 1–30.
SALTER, M. (2008) Imagining Numbers: Risk, Quantification and Aviation Secu-
rity. Security Dialogue, 39(2–3). p. 243–266.
SCHINKEL, W. (2011) Prepression: The Actuarial Archive and New Technologies
of Security. Theoretical Criminology, 15(4). p. 365–380.
SPARKE, M. (2006) A Neoliberal Nexus: Economy, Security, and the Biopolitics of
Citizenship on the Border. Political Geography, 25(4). p. 151–180.
SPIJKERBOER, T. (2007) The Human Costs of Border Control. European Journal
of Migration and Law, 9(1). p. 127–139.
TAYLOR, L. (2015) No Place to Hide? The Ethics and Analytics of Tracking Mobil-
ity Using African Mobile Phone Data. Environment and Planning D: Society and
Space. Available from www.academia.edu/4785050/No_place_to_hide_The_ethics_
and_analytics_of_tracking_mobility_using_African_mobile_phone_data.
[Accessed 25–8–2013]
TAYLOR, L. & BROEDERS, D. (2015) In the Name of Development: Power,
Profit and the Datafication of the Global South. Geoforum, 64(August 2015),
p. 229–237.
UNWTO. (2009) Tourism Highlights 2009 Edition. Madrid: UNWTO.
UNWTO. (2010) Tourism Highlights 2010 Edition. Madrid: UNWTO.
VAUGHN-WILLIAMS, N. (2010) The UK Border Security Continuum: Virtual Bio-
politics and the Simulation of the Sovereign Ban. Environment and Planning D:
Society and Space, 28(1). p. 1071–1083.
VAN DER PLOEG, I. & SPRENKELS, I. (2011) Migration and the Machine-
Readable Body: Identification and Biometrics. In Dijstelbloem, H. & Meijer, A.
(eds.) Migration and the New Technological Borders of Europe. Houndmills,
Basingstoke and Hampshire (UK): Palgrave, 68–104.
ZOLBERG, A. (2003) The Archaeology of ‘Remote Control. In Fahrmeir, A., Faron,
O. & Weil, P. (eds.) Migration Control in the North Atlantic World: The Evolu-
tion of State Practices in Europe and the United States from the French Revolu-
tion to the Inter-War Period. Oxford: Berghahn, 195–222.
ZOOK, M. et al. (2010) Volunteered Geographic Information and Crowdsourc-
ing Disaster Relief: A Case Study of the Haitian Earthquake. World Medical &
Health Policy, 2(2). p. 7–33.
 13 Migrants at/as Risk
Identity Verification and Risk-Assessment
Technologies in the Netherlands
Karolina La Fors-Owczynik and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Irma van der Ploeg

INTRODUCTION

European policy discourse on migration shows a tendency to prioritize the

security of EU citizens over securing migrants (Tsianos and Kuster, 2010).
Practices geared toward the sorting out of entitlements to access and resi-
dence of migrants mobilize a discourse that increasingly frames them as
posing risks to the societies of their host countries (Ministerie van Binnen-
landse Zaken en Koninkrijksrelaties, 2011, 2012a, 2012b) notwithstanding
the fact that only a small minority of migrants are associated with criminal
offences, for example, organized crime, illegal entry, human trafficking, or
terrorism. Generally, migrants with refugee or asylum status, as well as those
seeking this designation, arrive as a consequence of wars, natural disas-
ters, political or economic unrest, poverty, human trafficking, international
crime, or a combination thereof. These immigrants can be regarded as being
at risk. Although technological systems of migration and border control are
generally installed to prevent threats to a country’s inhabitants, including
threats associated with migrants, countries also have a human rights obli-
gation to assist and protect vulnerable migrants. Today, the perspective on
migrants as vulnerable and at risk, rendered acute with the current refugee
crises in the Middle East and on the Mediterranean Sea, seems increasingly
at odds with the perspective dominating border-management policies and
technologies (European Commission, 2006; Council of the European Union,
2012; European Parliament, 2012).
This chapter discusses how migrants and travellers are increasingly
framed as posing a risk to the Netherlands. More specifically, we analyse the
way that modes of prevention (Bigo and Guild, 2005; Broeders, 2009) and
their associated risk assessment systems have come to focus on translating
specific problems around migrants into problems of identity fraud, illegal
entry, and illegal residence. The ways in which these problems are defined
and managed has directed attention towards improving the establishment
and the management of ‘identity.’1 These identification practices have ethi-
cal implications because they may be instrumental in discriminative exclu-
sion and inclusion. For this reason, we focus in particular on identification
 262 K. La Fors-Owczynik and I. van der Ploeg
and risk-assessment practices performed by border control, migration, and
law enforcement agencies.
In the Netherlands, the constellation of border and immigration agen-
cies is known as the ‘alien chain,’ and that of law enforcement agencies as
the ‘criminal justice chain.’ Establishing migrants’ and travellers’ identities
and organizing fraud-proof identity management systems have high prior-
ity within both chains. This is pursued by elaborate risk-profiling practices,
identity-management systems, and sharing of information along and between
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

the range of agencies within the chains, such as the Alien Police, the Royal
Military Police, and the Immigration and Naturalization Service (IND).
Whereas terms such as ‘identity,’ ‘prevention,’ and ‘risk’ concerning migrants
suggest a certain self-evidence at first sight, these terms derive their mean-
ing from complex arrangements of local, national, and international institu-
tions, digital technologies, legal settings, and professional practices. This is
also evident in the international commitments. The Netherlands has a broad
range of information exchange tasks, including information on migrants and
travellers. As part of the European Union, the Netherlands is, for instance,
committed to performing policing tasks through using the national database
of the Schengen Information System (SIS2) and exchanging data within the
pan-European law enforcement organization, Europol.3 The Netherlands is
also obliged to enforce the Dublin convention,4 which prescribes member
states to determine whether an asylum seeker has submitted a prior applica-
tion in another member state. As part of these assessments, the fingerprint
database EURODAC5 is consulted. Another obligation for the Dutch govern-
ment is the exchange of visa data on migrants through the Visa Information
System (VIS6). Given that the Dutch sea border also constitutes part of the
EU’s external border, the Netherlands actively participates within the EU’s
border management organization FRONTEX.7 For example, the Dutch also
aggregate and share data on illegal entry through the EU’s border surveillance
system, EUROSUR.8 Beyond these European organizations and systems, a
large variety of national and local organizations exist in the Netherlands that
deploy local border control, immigration, and law enforcement procedures.
The two identification systems, as well as the risk-profiling system analysed
in this chapter, are embedded within these European and Dutch regulatory
and technological settings, yet operated solely in the Netherlands.
These three systems are the following: first, the identification and veri-
fication console of the Dutch Immigration and Naturalisation Service, the
INS console, second the console of the information provision program (in
Dutch: Programma Informatievoorziening Strafrechtsketen) within the Dutch
criminal law chain, the PROGIS console, and the Advanced Passenger
Information-System, the API system. Although they differ significantly, as
they are situated in different practices and operated for different purposes,
this chapter aims to articulate some shared characteristics and implications
of their use. We suggest that all three systems contribute to a shift that
increasingly frames migrants as posing a risk, as opposed to potentially
 Migrants at/as Risk 263
being at risk. Moreover, we aim to show how the quest for ‘accurate iden-
tity’ and the technologies used to approach this goal transform the problem
in specific ways and, in the process, create new uncertainties and risks for
those subjected to them. We argue that this happens because of the particu-
lar way problems of identity fraud and illegal entry are translated (Latour,
1987) into specific technological solutions.
To illustrate this dynamic, the next section first introduces the systems
under study and briefly describes how identity verification of certain catego-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ries of migrants and risk profiling of travellers is enabled by these systems.

The third section discusses how the ‘accuracy of identity’ and an accurate
‘risk profile’ are produced by focusing on the development of ‘risk indi-
cators.’ The fourth section looks a bit deeper into the functioning of the
systems in practice and identifies a number of pitfalls in the socio-technical
production of ‘identity’ and ‘risk.’ The analysis is based on empirical data
from interviews with professional users of these systems within the alien and
law enforcement chains and guided by actor-network theory (Akrich, 1992;
Latour, 1987) and a material semiotics–informed approach (Law, 2008).
The final section draws the arguments together and concludes that the tech-
nological quest for ‘accurate identification’ and risk prevention introduces a
range of new risks for migrants.

PREVENTION SYSTEMS IN DUTCH IMMIGRATION, LAW

ENFORCEMENT, AND BORDER MANAGEMENT

‘Risk prevention’ with regard to identity fraud, illegal entry, and interna-
tional crime has a high priority in Dutch immigration policy (Ministerie
van Binnenlandse Zaken en Koninkrijksrelaties, 2012a). In the discourse
on improving this prevention, visual metaphors dominate: Migrants and
their associated risks need to be made “more visible” (Prins et al., 2011),
and there is a need for more ‘complete and accurate pictures’ of migrants.
A report by the Dutch Advisory Committee on Alien Affairs, for instance,
describes a need for acquiring a “better view on asylum seekers” (Tweede
Kamer, 2005), and another report from the Ministry for Foreign Affairs sim-
ilarly emphasizes the need for a more “comprehensive picture” of migrants
(Zijderveld et al., 2013). The need for improving this ‘visibility’ is invari-
ably addressed through the introduction of advanced digital technologies
for identity-verification and risk-assessment purposes. The INS console, the
PROGIS console, and the API system exemplify such technologies.

The INS Console

The INS console is a biometric fingerprint scanning and facial photograph-
ing system employed by the Immigration and Naturalization Service (INS,
in Dutch: Immigratie en Naturalisatie Dienst) of the Dutch Ministry of
 264 K. La Fors-Owczynik and I. van der Ploeg
Interior Affairs. The European Council Regulation (EC) No. 380/2008 pre-
scribes that all EU member states, including the Netherlands, register “a
photograph [. . .] and two fingerprints taken flat and digitally captured” in
each residence permit issued to a third-country national. ‘INS consoles’ have
been installed specifically to meet this regulation by producing facial photo-
graphs and ‘high-quality’ digitally scanned fingerprints. These data are then
incorporated into the residence permits of non–EU migrants (Immigratie
en Naturalisatiedienst, 2013). First, it was only allowed to register scanned
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

fingerprints and personal photographs on the residence permits. But as of

1 March 2014, a legal change allows for the registration of the mentioned
biometric features in a central database, in the Basic Facility for Aliens.9
During every subsequent verification process, the residence permit holder
places his/her fingers on a reader, and the images are compared with the fin-
gerprints stored on the permit. Residence permits furnished with these bio-
metric features are considered to increase certainty about the permit holder’s
identity and that the document truly belongs to the immigrant to whom it
has been issued (Teeven, 2013). Enrolment in the system will eventually
include all third-country nationals acquiring a residence permit. However,
when the interviews used in this chapter took place, the INS had only begun
enrolling asylum seekers. Our analysis in subsequent sections of the ways in
which ‘identity’ and ‘risk’ are translated and performed within this system,
thereby reinforcing the trend of shifting perceptions from seeing them being
at risk to seeing them as risk, therefore pertains only to this particular sub-
group of migrants.
The digital fingerprint scanning by the INS console is considered an
improvement upon previous enrolment procedures for asylum seekers. That
procedure involved rolling the applicant’s 10 fingers in ink and producing
prints on paper, the so-called dactyslips. All original dactyslips are archived
in the Dutch National Investigation Information Service (DNRI). Copies
of these sheets are subsequently transferred to a central European system
called Eurodac, where they are checked against those already registered. In
case of a ‘no-hit,’ the fingerprints are registered, and the application will be
processed. In case of a ‘hit,’ this is taken as evidence of ‘asylum shopping,’
or at least as evidence that an earlier application has been submitted by the
same person in another EU country. Eurodac is the main tool for the execu-
tion of the Dublin Convention (Council Regulation (EC) No 2725/2000)
that determined that one cannot apply for asylum in more than one of the
EU member states and that the country of first application remains respon-
sible for that person and their application (Van der Ploeg, 1999). In addi-
tion to checking in Eurodac, copies of applicants’ dactyslips are registered
in the Basic Facility for Aliens (Basis Voorziening Vreemdelingen), an IND
database containing data on all migrants in the Netherlands, and in the
law enforcement database HAVANK (Het Automatisch Vingerafdrukken-
systeem Nederlandse Kollektie, or Automated Fingerprint system Dutch
Collection).
 Migrants at/as Risk 265
PROGIS System
Prevention of identity fraud by convicted criminals is a key target within
the Dutch law enforcement chain. Although this particular issue is part of
a much wider concern, it has recently been pushed to the forefront after
several prisoners were found to be others than those who were actually sen-
tenced. Therefore, the Dutch Act on Identification of suspects, convicts, and
witnesses (Wet identiteitsvaststelling verdachten, veroordeelden en getuigen,
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

2009) prescribes law enforcement authorities to ‘establish the identity’ of

suspects, convicts, and witnesses to crimes by taking their digital fingerprints
and facial photographs.
In line with this law, the police and other judicial authorities introduced
a new10 biometric system, the PROGIS console (PROGIS), to guarantee that
‘the right person’ is punished. At the end of an identity-verification and enrol-
ment process, which involves taking a facial photograph and scanning 10
fingers, a so-called ‘ID state’ is produced. This ID state meets the objective of
the Dutch Ministry of Security and Justice in acquiring a “more integrated,
criminal person image” (in Dutch: integraal strafrechtelijk persoonsbeeld) in
that it is claimed to ensure a person’s identity during criminal investigations.
The enrolment of suspects and witnesses in PROGIS constitutes the entry
point to the Dutch criminal justice system. From that point on, the ‘estab-
lished identity’ on the ID state becomes the reference point against which all
fingerprints and facial images taken at later moments in the procedure are
cross-checked and verified. In order to use the same person’s file between
criminal justice agencies, each PROGIS ID state is linked to a so-called
criminal justice chain number (CJCN, in Dutch: strafrechtsketennummer—
SKN) and stored in a central database, the Criminal Justice Chain Data-
base (CJCD). The CJCN is meant to harmonize the administration of cases
among agencies. In case of recidivism, this database allows for updating a
criminal record. This occurs when separate criminal cases become associ-
ated with the same person through one CJCN.
Remarkably, it has been the Dutch Alien police that was tasked with oper-
ating PROGIS for the criminal justice chain, primarily because of the expe-
rience and expertise they acquired through conducting identity research11
on migrants within the alien chain. However, biometric data produced by
PROGIS is only used by partners within the criminal justice chain to verify
whether a person’s data is already registered in the CJCD. Therefore, if an
immigrant offends or becomes a criminal suspect, they would be enrolled
in PROGIS the same as anyone else in the Netherlands. We are less inter-
ested here in the specific framing of an immigrant as posing a risk in the
sense of being a suspect in the context of a particular crime investigation.12
Our concern here regards the particular framing as risk that is performed
with the very enrolment—either as a suspect or a witness—in PROGIS. The
likelihood for a migrant to become suspected of identity fraud, for instance,
is higher than for ‘regular’ residents, since the identification process for
 266 K. La Fors-Owczynik and I. van der Ploeg
immigrants is in general far more complicated than that for Dutch citizens
and therefore more prone to error. Along this line, we highlight below how
the very use of PROGIS can increase the chances for a migrant to be framed
as identity fraudster.

The Advanced Passenger Information System (API)

Within the context of border control, travellers from non–EU states are
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

routinely assessed against risks of illegal entry (e.g. fraudulent visa or

passport), international criminal activity, or threats to national security
(e.g. terrorism). All these problems have contributed to the development
of more orchestrated risk profiling practices of passengers internationally.
Problems of international crime and terrorism generated new modes of data
exchange, as EU agreements on the exchange of Passenger Name Records
(PNR) of travellers with the US (2012), Canada (2006), and Australia
(2012) exemplify. In fact, airports transformed into ever-more-advanced
security checkpoints (Schouten, 2014), where, beyond facilitating the
mobility of travellers, the filtering out of ‘untrustworthy’ passengers grew
into a top priority.
The Advanced Passenger Information project (henceforth API) is situ-
ated at Schiphol Airport and is part of a larger programme called Sus-
tained Border Management13 (in Dutch: Verbeterd Grensmanagement) in
the Netherlands. API has been developed to meet EU (EC, 2004/82/EG)
and national regulations for the sole purpose of preventing illegal entry
into the Netherlands and is used by Dutch border control and immigra-
tion authorities to check all travellers moving through Schiphol Airport
against ‘API profiles.’ A wide range of passenger attributes, including
behavioural characteristics, air travel histories, modes of plane ticket
purchase, and classifications of personal belongings, are assessed against
these API profiles.
In 2010, the International Civil Aviation Organization (ICAO), the
World Customs Organization (WCO), and the International Air Trans-
port Association (IATA) revised the guidelines for API. According to
these guidelines, part of the flight information from the ‘departure con-
trol system’ (DCS) of airlines and individual passengers’ data from the
machine-readable zone of the travel document and passengers’ seat and
baggage information need to be included in the API dataset. This data-
set can be extended to include a maximum of 39 items (depending on
country-specific legislation). The EU also follows the international API
guidelines. Carriers of incoming and outbound flights in the EU are obliged
by Directive 2004/82/EC to submit information on passengers to border-
control authorities prior to their departure, which also applies to the Neth-
erlands. Such data may be kept by these authorities for a maximum of
24 hours. The Minister for Immigration suggests in a policy analysis about
the work of the Dutch Royal Military Police, Dutch Customs Services,
 Migrants at/as Risk 267
General Intelligence and Security Service, and the Immigration and Natu-
ralisation Service, as well as test results of the API project, that if these
agencies could gain access to passengers’ travel document data and check
that data against API risk profiles before travellers cross the border, this
would foster security and more efficient processes by agencies, as well
as smoother border crossing of travellers (Ministerie van Binnenlandse
Zaken en Koninkrijksrelaties, 2011).
API data and profiling are perceived as pivotal elements in apprehend-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ing migrants involved in human trafficking. While on the one hand being
construed as useful in the protection of potential victims, by identifying
those ‘at risk’ of this and other, related crimes (e.g. prostitution or exploi-
tation of illegal workers), API data are, on the other hand, also considered
useful in identifying those posing a risk, in particular regarding illegal
entry. Our analysis below will indicate how both uses of API profiles can,
in fact, merge into each other, thus casting doubt on the proclaimed affec-
tivity of each.

DOING RISK: BUILDING ‘INDICATORS’

The prevention of potential problems associated with migrants hinges in

part upon the sharing of information among practitioners, organizations,
and technological systems. Both the prevention of identity fraud through
identity documents and of illegal entry through risk-profiling methods
are shaped by the ways in which classifications, standards, watch lists,
and other tools are implemented. This ‘operationalisation,’ in turn, takes
the form of marking specific personal and other characteristics as ‘indica-
tors.’ Although such indicators are seemingly straightforward character-
istics, they are in fact the result of a considerable amount of work (La
Fors-Owczynik and Valkenburg, this volume). Moreover, framing risks,
that is, a possible future occurrence, as something detectable in the pres-
ent, requires an expectation to be translated into observable categories,
which, at the same time, renders these risks more ‘real.’ In addition, and
despite their apparent simplicity, risk indicators need to be continually
interpreted by professionals. The efforts to determine and interpret risk
indicators involve an opaque set of problem transformations by different
actors that introduce a range of ambiguities concerning the functioning of
the technologies.

Establishing ‘Proof’ of Identity: Revisiting the ‘Entry-Point

Paradox’
Ambiguities around what constitutes a risk of identity fraud also emerge
in the ways in which ‘identity’ is established within the configuration of
PROGIS. Although this system is intended to provide more accurate ways
 268 K. La Fors-Owczynik and I. van der Ploeg
of identifying persons by biometric and other technologies, thus to prevent
the risk of identity fraud, enrolment in the system requires prior ‘proof’ of
identity:

PROGIS-consoles do not check the authenticity of ID documents before

a person becomes enrolled. [. . .] so, professionals must know IDs. New
European documents are all made of polycarbonate, such as Polish IDs,
for instance. If you drop one, you hear a tinny sound. Therefore, while
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

being busy with someone, policemen often let documents ‘accidentally’

fall. If a document sounds dull, that indicates fakeness.
(PROGIS professional, city A)

This quote exemplifies two points. First, ID documents are not checked for
authenticity by PROGIS; instead, professionals must decide whether an ID
document used at enrolment is genuine or fake. Second, the potential problem
of identity fraud is transformed into observable indicators, such as, in this
particular example, the sound of a falling ID card. Thus, a general problem
of identification, one that Roger Clarke (1994) labelled the ‘entry-point para-
dox,’ resurfaces here: the accuracy of each digital identification or verification
system, however sophisticated, is only as strong as its weakest link, which
often lies right at the start of the whole procedure, the authenticity of ‘breeder
documents’ (e.g. IDs, birth certificates). This clearly is an issue in PROGIS,
where its claimed technologically enhanced accuracy, in practice, may depend
upon an improvised trick that is anything but high tech or accurate.
In a similar vein, the INS console is claimed to provide an ‘accurate iden-
tity’ for migrants while also requiring prior ‘identity proof’ from them. A
professional identifying so-called ‘invited refugees’14 via a mobile INS con-
sole in refugee camps explains:

I: How can you be sure that the [undocumented] person you want to
identify with the INS-console is the one he/she claims to be?
P: That you never know. I can give you a basic answer: I can only hope
that the person who said something about him/herself to me, said the
same thing to our UNHCR15 colleagues. It can be two times a lie, but
since I cannot figure this out, I must be satisfied with the consistency
of the refugee’s story.
(INS professional, city D)

The above excerpt indicates once more the ‘clay feet’ of many high-tech
identification and verification systems today. Sometimes all one has to go on
at enrolment is a minimum level of consistency in a claimed identity, life,
or flight story. In the final analysis, it is still the experienced worker ‘in the
field’ whose judgement about the credibility of someone’s personal account
carries weight. Whether this is to be seen as a strength or a weakness—some
may find this remaining human element reassuring—it is a key element in
 Migrants at/as Risk 269
the functioning of such systems that is usually left out of the accounts of
technological accuracy and their role in combating identity fraud.

Turning Attributes into Indicators for ‘Illegal Entry’

Building risk profiles from indicators visible in the Advanced Passenger
Information system shows similar shifts in localization of the problem. The
problem here is ‘illegal entry,’ and the localization may be quite literal, in
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

the sense that geographical locations and travelling routes may become part
of a particular risk profile:

Ultimately, you want to focus on every flight, but you need to build this
up gradually. Where does our first priority lie? [. . .] if other partners
also have problems with flights from a given country, than there must
be something wrong with those flights. If 5 professionals from different
organizations say the same about 3 places, than there is something risky
with those flights.
(API professional, city A)

The same professional explains how a range or a combination of rather

generic attributes and characteristics of passengers on flagged flights are
turned into risk indicators.

You can discover that [. . .] ‘facts’ about persons traveling from those
risk flagged places are also important, these help you learn the type
of persons you look for. [. . .] you often search for particular men or
women of a certain age, behaviour, clothing or travel companions. I
often check passengers against such indicators.

As indicated earlier, API profiles are also used to prevent illegal entry
connected to international crime like human trafficking. An IND official,
who also manages API, describes how particular instances of this crime are
sometimes spotted through the API profiling:

We did quite well with that profile and stopped many women who were
glad and grateful for what we did. We were also happy we could protect
them from becoming prostitutes. [. . .] The funny thing is that a month
after we started to screen flights against the profile, trafficked women no
longer travelled directly to Madrid, but took the route: Amsterdam—
Paris—Barcelona—Madrid. The change in the route assured us that the
profile was good, and that criminal organizations are well informed.
The API profile [. . .] did well. This allowed us to indicate for the min-
ister the usefulness of API data and the importance of binding several
issues together in one profile.
(API manager, city A)
 270 K. La Fors-Owczynik and I. van der Ploeg
Interestingly, this quote illustrates as well how the API system–based pro-
filing may contribute to the convergence, if not conflation, of protecting
migrants at risk (of trafficking and sexual exploitation) with prevention of
travellers posing a risk (of illegal entry). There is no need to doubt the inten-
tions of border guards in any way in order to legitimately raise the question
whether stopping someone at the border casts them as victim or as perpetra-
tor, as it is unclear whether this is followed by sending them back, prosecut-
ing them, or really helping them. The fact of traffickers changing their routes
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

is taken as a sign of success of the profile and could indicate that it is above
all the mere prevention of illegal entry that counts.
It is also clear that professionals do not merely follow standardized pro-
tocols or automated decision processes. Seeing certain attributes of travel-
lers as risk indicators can also be based on tacit knowledge and acquired
experience:

If I only consider what is registered in the system, I would only follow

the system. The moment I look at my quantitative data analysis and
indicators suggest that persons who smuggle others always have a red
cap on, I also start ‘qualitative research.’ If I only focus on persons with
red caps I would not see if someone is smuggling people in yellow rain
boots, for instance. You have to rely on your senses [. . .] and adjust
your profile.
(API professional, city A)

We repeatedly heard about professionals’ ‘gut feelings’ being the reason

for modifying risk profiles and introducing new indicators. Thus a shared
or hybrid agency exists between human actors and automated systems that
makes the process leading up to the decision who will be assigned a risk
qualification and who will be stopped for further inspection rather obscure.
The constant need to amend API profiles and the flexible conception of indi-
cators renders the process even more elusive.
A major part of the work in identifying risk-associated migrants is
about fixing the ambiguities, uncertainties, and (im)probabilities concern-
ing what, in practice, constitutes such risk. Yet the whole idea of detecting
and preventing risk in this context requires the presumption that risks are
somehow observable, a presumption that lies at the basis of the efforts to
develop indicators for them. This section, however, gave examples sug-
gesting that the particular indicators used within the operation of the INS
console, the PROGIS console, and the API system sometimes increase ambi-
guities and uncertainties about risks. Moreover, as the next section aims to
make clear, the very use of these three systems generates a proliferation of
such ‘indicators’ and requires workarounds to make them function, that,
together, have the ironic effect of increasing the perception rather than the
prevention of risk.
 Migrants at/as Risk 271
PITFALLS OF MAKING RISKS ‘VISIBLE’

To distinguish those entitled to enter the country, obtain employment, or

receive government support from those who pose a security risk has always
been a politically laden process. Pivotal in this is the conceptualization of
identity as something fixed and verifiable. As Caplan and Torpey (2001)
argue, identification processes transform the ‘who’ question into ‘what kind’
of a person someone is. This shifts the search for identity towards the clas-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

sification of personal characteristics into categories. Similarly, Btihaj argues

that “this collapse of the ‘who’ into the ‘what’ [. . .] indicates their [identi-
fication modes] inherent limitations in capturing the ambiguity of identity
and the complexity of the lived experience” (Btihaj, 2010, 6).
Technological innovations like machine-readable passports, e-IDs, and
biometric identification systems exemplify efforts for producing certainty
about a person’s identity (Holowitz and Noiriel, 1992; Caplan and Torpey,
2001; Bigo and Guild, 2005; Lyon, 2009). However, this certainty rests on
the acceptance of a series of problem translations that remain unexamined
by those operating the systems but that nonetheless each introduce uncer-
tainty and possibility of error (Van der Ploeg and Sprenkels, 2011).
Continuing this line of thought, in this section we argue that efforts to
make the risks of identity fraud and illegal entry visible requires a particular
type of visibility, one that always could have been otherwise. Moreover,
we show how the particular visualization of risk produced by the INS and
PROGIS consoles and the API system inscribes these risks onto material
objects such as clothing or personal belongings and body parts.
The practices within which these systems are used exemplify that ‘iden-
tity’ and ‘risk’ are increasingly constitutive of each other: To identify immi-
grants presupposes risk assessments, and to assess immigrants against risks
involves practices of identification and identity verification. As much as
identification practices are increasingly about extensive categorization and
registration of personal attributes and modes to verify these, risk assess-
ments of immigrants are increasingly about cataloguing a growing number
of personal attributes as risk indicators. That is to say, if an immigrant’s
‘identity’ does not conform to a norm during a verification process by the
INS or PROGIS console, or if certain attributes of an immigrant match with
an API risk profile, that identifies them as a ‘risky migrant.’
In the following we highlight how operating the systems discussed cre-
ates additional ‘risks.’ The systems all have their weaknesses and fallibilities
that may be exacerbated when agencies exchange information and repeat
procedures in different systems. The next section demonstrates how this
occurs in the field and provides further examples of how problems of iden-
tity fraud, illegal entry, or related crimes by immigrants become translated
into other issues within the configurations of the INS, PROGIS console, and
API system.
 272 K. La Fors-Owczynik and I. van der Ploeg
False Positives and False Negatives
The INS console was intended to provide ‘improved’ biometric data for
verification by skipping a translation step within the fingerprinting processes
(see above). In the new system, fingerprints are directly digitally scanned, so
the scanning of the ink-based dactyslips is no longer necessary.

P: The problem was that if you take a dactyslip [. . .] a fingerprint con-

Downloaded by [University of California, San Diego] at 19:23 09 January 2017

sists of lines. If you put here a magnifying glass, then you see that the
lines consist of stripes and points, that’s a print of a scanned finger-
print. What happens now with the Dactyslip is that the sheets [of the
inked fingerprints] become digitalized and the lines of your fingerprint
are transformed into stripes and points. So, you have a large quality
loss. The moment you scan fingerprints by the INS-console, the qual-
ity becomes in any case higher. A bad digital fingerprint scan is prob-
ably still better than the best photocopy of an original dactyslip.
(INS console operator, city A)

However, the improved quality and matching rates turn out not to be as
straightforward as this. In order to attain the expected accuracy of verifica-
tion, significant workarounds are needed that show how risk assessment
and identity verifications still rely on assessments independent of the system:

I: How about the lady with whom it went wrong?

P: [. . .] the system says ‘no match’ for her fingers [. . .] Are we going to
refuse that lady a residence permit, because the prints might not be
hers? [. . .] The lady was a Somali national, and although Somalis do
that, she had not mutilated her fingerprints [. . .], I checked her fin-
gers myself. Then I took her prints and got a 40 % result. I was like:
what is this? [. . .] if you clean your fingers, then the prints are worse,
if they are greasy, you gain better prints. What are the oily spots on
your skin? It is here [points to side of nose] and on your forehead. So,
to get good prints we ask people to rub these spots [. . .]. Yet, these
techniques were not helpful. Finally, the project manager said, put
her four fingers on the scanner and fold them a little around the table,
then you have the right pressure. It was not easy, but we improved the
fingerprint quality. We achieved 80%; that was good quality.
(INS officer, city A)

What we see in this excerpt is how the accuracy of the system is in fact
something that takes a lot of effort and additional ‘techniques’ to achieve.
More specifically, the system is made to perform well because the actual risk
assessment is done differently and turned out negative (‘no risk’): When her
fingerprints fail to give a match, an asylum-seeking woman from Somalia
is first suspected of perhaps trying to sabotage the verification process by
 Migrants at/as Risk 273
mutilating her fingertips because she is Somali (“Somalis do that”). The
specific risk indicator for that type of sabotage (having mutilated fingertips)
is subsequently checked visually by the IND officer (“I checked her fingers
myself”). After satisfying himself this way that no fraud is attempted, the
negative verdict of the system is not believed, and all effort goes into the
production of a better fingerprint scan. The belief in the first assessment is so
strong that even if all this fails, the project manager is asked to step in and
help out to achieve a positive verification.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

The other side of this is that any reliance on such a system introduces
significant new risks for those whose fate depends on being believed and
whose identity claim of being at risk. That is, a recognised refugee with right
of stay is ‘verified’ this way. Instead of the promised accuracy and certainty
of digitalized biometric identity verification, a number of highly contingent
factors determine whether a person passes this verification test, including
the assessment by the operator, the manager, and the availability of the right
levels of pressure and greasiness of the fingers to be scanned.
But often enough, if a person’s fingerprints turn out to be unreadable
within the configuration of a digital biometric system, this is perceived as an
indicator of potential identity fraud. Consequently, the owner of the finger-
print becomes suspect and changes from a person claiming to be at risk into
one that is perceived as risk. This is the case even though it is known that
there are more problems with the fingerprinting systems.
For example, the PROGIS console, although equally intended to improve
accuracy, is acknowledged to have an in-built risk of failure, stemming from
a weaker predecessor system:

Webfit is the predecessor of PROGIS, and was the first program using
fingerprints that were taken digitally. In Ter Apel, the Alien Police said:
we tried out Webfit, but it did not work for us and it took too much
time to work with it. The program of Webfit was actually installed in
such a way that it always delivered negative results for asylum seekers
when verifying their fingerprints. Webfit turned out to have only been
tested on Dutch fingers [. . .]; despite updates, the basis for PROGIS is
still Webfit.
(INS professional about PROGIS)

Ironically, a system intended for the highly diverse population of inter-

national refugees and asylum seekers had been tested on a relatively homo-
geneous population (‘Dutch fingers’), thus rendering it basically unsuitable
for its purpose.
Here we see the rather paradoxical conception of the body underlying the
very idea of biometric technologies as automated identification and authen-
tication tools playing out (Van der Ploeg, 2011) on the one hand; biometrics
is based on the biological fact that every individual is physically unique. No
two fingerprints are identical, everybody’s irises are different from those of
 274 K. La Fors-Owczynik and I. van der Ploeg
another, the same way that no two faces, voices, or retinas are exactly the
same. This is the condition of possibility of the very idea of biometric tech-
nologies as identification tools.
On the other hand, however, there is a simultaneous assumption of simi-
larity: every human person is assumed to have a clearly audible voice, a set
of 10 fingerprints, two irises, a recognizable face, and so on. Though hardly
ever mentioned as such, this assumption of similarity is as crucial to the
functioning of biometric systems as is the assumption of uniqueness.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

With respect to the human bodily features used in biometrics, this means
that there is an assumption of normality that is defined as a range of varia-
tions that constitute ‘the normal.’ Such notions of normality are built into
the equipment: hand scanners have particular shapes and sizes, with des-
ignated places to put the fingers; fingerprint systems are designed for the
registration and comparison of a particular number of fingerprints (most
systems use 1, 2, or 10) and a particular range of ridge definition; cameras
to scan faces are directed at a specific height, and the accompanying face-
recognition software often works best for a particular shade range of skin
colour, and so on. In the case of PROGIS and its underlying legacy software
Webfit, it is this assumed similarity that turns out to be unwarranted and
causing trouble.
The question is, however, for whom it causes trouble. The quote shows
operators disqualifying the system: it does not work for them, and it takes
too much time to operate. The other side of the coin, however, is that an
unsuccessful enrolment or verification may equally well be attributed to the
asylum seeker, casting suspicion on them. In a context in which policy aims
at limiting admittance of refugees, and prevailing distrust of migrants’ iden-
tity claims is the very reason for introducing technologies that come with
the promise of accuracy and certainty, one may wonder who will get the
benefit of doubt in any concrete instance of failed recognition on the work
floor. This is why we suggest that these systems introduce new risks for the
people subjected to them.
In a different way, and for a different set of travellers and migrants, the
API system–based profiles constitute new risks as well. As described earlier,
a traveller can become framed as ‘risk’ if a match emerges between the attri-
butes of this passenger and an API profile:

We were informed by Madrid that a gang is trafficking women from

South- and Central America for prostitution. [. . .] We learnt that these
women use Amsterdam as a transit; therefore we built a profile on flights
from South and Central-America. We wanted to see all the women, in
the age group of 18–24 years, travelling through Amsterdam to Madrid.
If a match emerged on these four criteria, then the system signalled that
passenger. Since not all female passengers were being trafficked, you had
to ask what travel purpose a passenger had. [. . .] there were often no
reasons for a stopping, but these are indicators you base your profile on.
 Migrants at/as Risk 275
What the above illustrates, however, is how a particular combination
of rather general attributes renders a person suspect if a match with the
API profile occurs. The very generality of the attributes mentioned—a
certain age range or gender—demonstrates the limitations of risk assess-
ments by API. False positives are certain to be many. This highlights that
an API profile, and the indicators it is built on, do not really ‘capture’ risks
but rather mediate and displace the ambiguities and inherent uncertainties
of risk.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Organizational Boundaries and Information Exchange

Visualization of risk is also done by information exchange between differ-
ent agencies within the alien and law enforcement chains. This informa-
tion exchange implies a certain “partiality of professional knowledge” (La
Fors-Owczynik and Valkenburg, this volume) as each profession focuses on
different things, and digital technologies are perceived to overcome inter-
agency boundaries among immigration, border control, and law enforcement
practices. The following interview excerpt demonstrates that during these
communications, professional and organizational boundaries are actively
maintained. This is in part a consequence of the ways in which knowledge
exchange brings along overlap of tasks between separate local practices.
The different identification tasks of the Alien Police for law enforcement
and immigration demonstrate that intensive efforts—or ‘boundary work’
(Gieryn, 1983)—are undertaken to protect the integrity of the immigration
and law enforcement chains. Yet these same efforts make cooperation pos-
sible among these professions:

P: INS is a distant chain partner of the Aliens Police, but not one that
can access the criminal justice database [database for PROGIS-data].
That is separate.
K: Yes, but the INS also does identification.
P: Yes, [. . .] the INS very often contacts the Aliens Police, and explains:
“Listen, we have someone in detention, but we would like to present
him/her [at the embassy].” Then we [Alien Police] start an identity
investigation, but not at the PROGIS-console, but according to our
identity investigation task on aliens within the alien chain. So we
put the whole PROGIS issue aside. As you see cooperation [with the
INS] is there.
(PROGIS professional, city B)

The distinct functions of the Alien Police in establishing persons’ identi-

ties within the law enforcement chain via PROGIS and within the alien chain
by comprehensive investigations on ‘undocumented’ migrants draw bound-
aries between the law enforcement and immigration chains. Although these
boundaries serve to maintain legally underpinned distributions of work,
 276 K. La Fors-Owczynik and I. van der Ploeg
authority, and responsibility, they also frustrate the INS and PROGIS con-
sole operators:

We both have the same goal to prevent identity fraud, and one-time reg-
istration, and multiple use. And then, of course, legislation should also
cooperate in this, a new bill is in progress [. . .]. But now there are two
parallel trajectories and I hope they will meet in the best interest of the
alien. Because currently assessments are done double, and why? Why
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

don’t you look for cooperation?

(INS console operator, city A)

As “assessments are done double” through enrolment in the PROGIS and

INS consoles sometimes, the risk of something going wrong and someone’s
identity thereby becoming suspect increases. Risks are “situated” entities in
the bodies of knowledge upon which they are based (Suchman, 2009) and
contingent upon the different socio-technical network within which they
emerge. The introduction of new systems, the double use of systems, and
the information exchange between them and the organizations that operate
them thus all generate additional risks of error, the burden of which will fall
on the people subjected to them. The new legislation currently underway
referred to in the quote is presented to reduce the occurrence of this problem
by proposing to have one central database for all biometric data used in the
alien chain, the BVV mentioned above; at the same time, however, the use of
biometrics throughout the various IND and border management procedures
is significantly extended, as are the number of agencies that will be legally
entitled to enter data or access this database. The Privacy Impact Assessment
conducted on this new legislation, conducted by the Ministry of the Interior,
identifies a range of ‘risks’ connected to the new network of systems, such as,
for instance, a risk of ‘data overload,’ and emphasises the need for a range of
risk-reduction measures (Ministerie van Veiligheid en Justitie, 2013).

CONCLUSION: TECHNOLOGY SHIFTS AND

PROLIFERATES ‘RISKS’

Dutch efforts to reduce identity fraud and illegal entry in the context of
migration have come to concentrate on prevention in recent years. Policy
makers were calling for a ‘better, more complete picture’ of migrants in
order to enable effective risk assessment of migrants and travelers entering
the country. This led to an increased reliance on information systems and
digital identification. This chapter has focused on three recently introduced
systems for identifying and verifying those posing a risk: the Advanced Pas-
senger Information system for profiling travelers and the biometrics-based
identity verification systems within the alien and law enforcement chains,
the INS console and the PROGIS console.
 Migrants at/as Risk 277
Our analysis has highlighted how none of the socio-technical configura-
tions of these technologies delivers the flawless accuracy hoped for. While
in some respects improvements regarding the prevention of identity fraud
and illegal entry by travelers and migrants may have been achieved, a closer
look into the way these systems work in practice has shown how a range
of potential system fallibilities lead to a shift in localization of the problems
rather than a solution.
With the use of the systems, new risk indicators are developed that
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

amount to new norms concerning what a valid identity is and, hence, what
a trustworthy migrant is. Conforming to these new norms is a challenge
also for legitimate migrants and travelers involving aspects trivial outside
the context of digital identification: the ‘machine readability’ and greasi-
ness of one’s fingertips or the age at which one travels a particular route at
a particular moment. The (in)ability to produce readable fingerprints or a
particular combination of rather general attributes such as country of ori-
gin and age/gender may come to inspire distrust in a particular traveller’s
identity claim. This happens despite the fact that most of these systems, to
some extent at least, are built on clay feet: To enter one of the systems for
identity verification, one needs to provide a primary proof of identity, the
validity of which cannot be checked by the system itself. Hence, a range of
tricks, experiential knowledge, and trust comes into play in order to assess
the reference identity proofs on which all later translations and verifications
come to rely.
The consequences of this are twofold. On the one hand, it seriously
undermines the trust put into the technology and the belief that the latest
technology delivers what it promises: accurate and strong verification and
identification of those migrants posing a risk of fraud or becoming illegal
residents. In this sense, these technologies may foster a false sense of control,
security, and successful ‘prevention.’
On the other hand, however, these ‘clay feet’ may cause a failure of pro-
tection of those most in need of it: From being a refugee fleeing from risk
for life and limbs, an asylum seeker confronted with the systems set up to
keep out those deemed non-eligible for the protection offered by asylum
status may find themselves being recast as posing a risk; a failure to enrol
may be met with suspicion of a deliberate attempt to avoid successful reg-
istration; a false negative during subsequent verification may cast them as
identity fraudster. They may chance upon operators, who, for some reason,
and unlike the ones interviewed for this chapter, are little inclined to guide
them carefully through either the enrolment or the later verification pro-
cesses with the range of workarounds required to make the system perform.
In this sense, these systems and the belief in them actually pose new risks for
those coming to our borders for protection against risk.
The Privacy Impact Assessment on the new legislation mentioned above,
with its extensive enumeration of ‘risks’ to migrants’ rights posed by the
increased reliance on and extended use of biometrics throughout the alien
 278 K. La Fors-Owczynik and I. van der Ploeg
chain proposed in the new bill within the Netherlands’ legislature, underscores
this view. Even though it includes a range of risk-mitigating measures—for
this issue of false negatives, that is, a failure of the systems to match two sets
of fingerprints from the same person, for example, it proposes a mandatory
check by a dactyloscopic expert in case of mismatch—it remains to be seen
how effective and feasible these will turn out to be. The proliferation of ‘risks’
when using the INS and PROGIS consoles and the API profiles underlines
the pivotal need for all involved to remain reflective about the pitfalls the use
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

of these systems can bring along for the lives of persons assessed by them.

NOTES

1. In opposition to the identity-management discourse, where ‘digital identity’ is

often equated with a collection of personal data, this chapter regards identity
as a ‘multi-layered concept,’ as something “interactive, mediated, relational,
and dynamic, as opposed to something pre-given to be ‘expressed’ or ‘regis-
tered’” (Van der Ploeg, 1999).
2. SIS—http://ec.europa.eu/dgs/home-affairs/what-we-do/policies/borders-
and-visas/schengen-information-system/index_en.htm
3. Europol—www.europol.europa.eu/content/page/about-us
4. Dublin Convention—http://europa.eu/legislation_summaries/justice_freedom_
security/free_movement_of_persons_asylum_immigration/l33153_en.htm
5. EURODAC—http://ec.europa.eu/dgs/home-affairs/what-we-do/policies/
asylum/identification-of-applicants/index_en.htm
6. VIS—https://secure.edps.europa.eu/EDPSWEB/edps/Supervision/VIS
7. FRONTEX—http://frontex.europa.eu/
8. EUROSUR—http://frontex.europa.eu/intelligence/eurosur
9. The new bill also enables a significant extension of use of biometrics through-
out the alien chain (Ministerie van Veiligheid en Jusitie, 2014).
10. The digital fingerprint scanning via PROGIS is new compared to the classi-
cal dactyloscopy routine of law enforcement practices. This entails the ink-
based fingerprinting of suspects in remand and convicted criminals, and the
production of dactyslips. Dactyslips are digitally copied and saved in two
major police databases: in AFDC and in the Facility for Verification and Iden-
tification (FVI, in Dutch: Voorziening voor Verificatie en Identificatie, VVI;
Ministerie. van Veiligheid en Justitie, 2011). (The original sheets are submit-
ted to and archived by the Dutch National Research Information Service/
dNRI/.) The fingerprints taken by PROGIS during the identification process
are regularly compared with fingerprints stored in AFDC. If fingerprints have
not been registered yet, dactyslips are produced and their copies transferred
to AFDC. In case a suspect is an alien, fingerprints are also compared with
the BFA.
11. The Dutch Alien Police operates PROGIS from the back office, including an
extra control function. Back-office tasks were granted to the Dutch Alien
Police based on the already-accumulated experience of this unit in conducting
identity research on immigrants within the alien chain. These tasks include
establishing the identity of suspects by PROGIS, which can involve ‘identity
research’ on immigrants (Ministerie van Binnenlandse Zaken en Koninkrijk-
srelaties, 2011).
 Migrants at/as Risk 279
12. Immigrants who overstay their visa, who are not part of an asylum-seeking
process, or who have simply no identity documents all are enrolled in PROGIS
as illegal residents.
13. The Dutch Sustained Border Management programme is coordinated by the
Department of Identity Management and Immigration of the Ministry of Inte-
rior Affairs. This includes four sub-projects: Passenger Related Data Exchange
(PARDEX), Advanced Passenger Information (API), no queue (No-Q), and a
Registered Traveller (RT) programme. PARDEX is a central system intended
for sharing passengers’ information (acquired prior to their travel) with law
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

enforcement agencies. The No-Q and RT are aimed at fostering automated

border passage for all EU passengers based on machine readability of travel
documents.
14. Most countries have signed agreements about inviting a given number of refu-
gees. The Netherlands invites a maximum 500 refugees yearly. INS profession-
als select these refugees by visiting camps and conducting research about who
shall become entitled to an invitation.
15. Refugee camps are generally run by UNHCR and local authorities.

BIBLIOGRAPHY

AKRICH, M. (1992) The De-scription of Technical Objects. In Bijker, W. & Law,

J. (Eds.) Shaping Technology/Building Society: Studies in Sociotechnical Change.
Cambridge (MA): MIT Press, 205–224.
BIGO, D. & GUILD, E. (Eds.) (2005) Controlling Frontiers: Free Movement into
and within Europe. London: Ashgate.
BROEDERS, D. (2009) Breaking Down Anonymity. Digital Surveillance on Irregu-
lar Migrants in Germany and the Netherlands. Amsterdam: Amsterdam Univer-
sity Press.
BTIHAJ, A. (2010) Recombinant Identities: Biometrics and Narrative Bioethics.
Journal of Bioethical Inquiry, 7(2). p. 237–258.
CAPLAN, J. & TORPEY, J. (Eds.) (2001) Documenting Individual Identity: The
Development of State Practices in the Modern World. Princeton (NJ): Princeton
University Press.
CLARKE, R. (1994) Human Identification in Information Systems: Management
Challenges and Public Policy Issues. The Information Society, 7(4). p. 6–37.
COUNCIL OF THE EUROPEAN UNION. (2000) Council Regulation (EC) No
2725/2000 Concerning the Establishment of “Eurodac” for the Comparison of
Fingerprints for the Effective Application of the Dublin Convention. Official
Journal of the European Union, 43(L316). p. 1–10.
COUNCIL OF THE EUROPEAN UNION. (2004) Council Directive 2004/82/EC
on the Obligation of Carriers to Communicate Passenger Data. Official Journal
of the European Union, 47(L261). p. 24–27.
COUNCIL OF THE EUROPEAN UNION. (2008) Regulation (EC) No. 380/2008
of 18 April 2008 Amending Regulation (EC) No 1030/2002 Laying Down a Uni-
form Format for Residence Permits for Third-Country Nationals. Official Journal
of the European Union, 51(L115). p. 1–7.
COUNCIL OF THE EUROPEAN UNION. (2012) Council Decision of 22 Sep-
tember 2011 on the Signing, on Behalf of the Union, of the Agreement between
the European Union and Australia on the Processing and Transfer of Passenger
Name Record (PNR) Data by Air Carriers to the Australian Customs and Border
Protection Service. Official Journal of the European Union, 55(L186). p. 1–12.
 280 K. La Fors-Owczynik and I. van der Ploeg
EUROPEAN COMMISSION. (2006) Commission Decision of 6 September 2005
on the Adequate Protection of Personal Data Contained in the Passenger Name
Record of Air Passengers Transferred to the Canada Border Services Agency. Offi-
cial Journal of the European Union, 49(L91). p. 49–60.
EUROPEAN PARLIAMENT. (2012) Agreement between the United States of Amer-
ica and the European Union on the Use and Transfer of Passenger Name Records
to the United States Department of Homeland Security. Official Journal of the
European Union, 55(L215). p. 5–14.
GIERYN, T. F. (1983) Boundary-Work and the Demarcation of Science from Non-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Science: Strains and Interests in Professional Ideologies of Scientists. American

Sociological Review, 48(6). p. 781–795.
HOLOWITZ, D. & NOIRIEL, G. (1992) Immigrants in Two Democracies: French
and American Experiences. New York: NYU Press.
IMMIGRATIE EN NATURALISATIEDIENST. (2013) Vingerafdrukken op
Verblijfsdocumenten—Geen Fotokaart Meer. Online Newsletter IND. Available
from www.ind.nl/Nieuws/Pages/Vingerafdrukkenopverblijfsdocumenten.aspx.
[Accessed 28–11–2013]
LATOUR, B. (1987) Science in Action: How to Follow Scientists and Engineers
through Society. Cambridge (MA): Harvard University Press.
LAW, J. (2008) Actor Network Theory and Material Semiotics. In Turner, B. S. (ed.)
The New Blackwell Companion to Social Theory. Oxford: Blackwell, 141–158.
LYON, D. (2009) Identifying Citizens: ID Cards as Surveillance. Cambridge, Mal-
den: Polity Press.
MINISTERIE VAN BINNENLANDSE ZAKEN EN KONINKRIJKSRELATIES
(BZK) (2011) Kamerbrief over de Aanpak van Illegaal Verblijf in Reactie op
het WODC-Rapport Illegalenschatting 2009. Kamerstukken, II 2010–2011, 31
594(10). p. 1–11.
MINISTERIE VAN BINNENLANDSE ZAKEN EN KONINKRIJKSRELATIES.
(BZK) (2012a) Het Gebruik van Passagiergegevens in het Grensbeheer, Den
Haag, Kamerbrief 2012–0000143248, p. 1–8.
MINISTERIE VAN BINNENLANDSE ZAKEN EN KONINKRIJKSRELATIES
(BZK) (2012b) Notitie inzake Openbare Orde Bevoegdheid Burgermeester. Den
Haag, April 26th.
MINISTERIE VAN VEILIGHEID EN JUSTITIE. (2011a) Het Fundament Progis.
Available from www.vka.nl/sites/default/files/downloads/Boekje Progis Het Fun-
dament.pdf. [Accessed 13–11–2013]
MINISTERIE VAN VEILIGHEID EN JUSTITIE. (2011b) Protocol Identiteitsvast-
stelling Strafrechtsketen (pp. 1–36). Available from www.rijksoverheid.nl/doc
umenten-en-publicaties/brochures/2011/12/22/protocol-identiteitsvaststelling.
html. [Accessed 25–04–2013]
MINISTERIE VAN VEILIGHEID EN JUSTITIE. (2013) Privacy Impact Assessment
met betrekking tot de Wijziging van de Vreemdelingenwet 2000. The Hague,
p. 1–15.
MINISTERIE VAN VEILIGHEID EN JUSTITIE. (2014) Besluit van 21 Januari 2014
tot Wijziging van het Vreemdelingenbesluit 2000 in Verband met de Uitbreiding
van het Gebruik van Biometrische Kenmerken in de Vreemdelingenketen. Staats-
blad van het Koninkrijk der Nederlanden, 44, pp. 1–23.
PRINS, C. et al. (2011) iOverheid. Amsterdam: Amsterdam University Press/WRR.
SCHOUTEN, P. (2014) Security as Controversy: Reassembling Security at Amster-
dam Airport. Security Dialogue, 45(1). p. 23–42.
SUCHMAN, L. (2009) Embodied Practices of Engineering Work. Mind, Culture,
and Activity, 7(10). p. 4–18.
TEEVEN, F. (2013) Reactie op de brief van de Commissie Meijers van 24 januari
2013 over het EURODAC-voorstel. The Hague: JBZ Raad, 1–6.
 Migrants at/as Risk 281
TSIANOS, V. & KUSTER, B. (2010) Transnational Digital Networks, Migration
and Gender, Mig@net Project Deliverable No. 6: Thematic Report “Border
Crossings.” Available from https://mignetproject.eu. [Accessed 12–03–2015]
TWEEDE KAMER DER STATEN-GENERAAL. (2005) Terugkeerbeleid. Verslag
van een Algemeen Overleg. Kamerstuk, 29344 (47). Available from www.zoek.
officielebekendmakingen.nl. [Accessed 22–05–2013]
VAN DER PLOEG, I. (1999) The Illegal Body: “Eurodac” and the Politics of Biomet-
ric Identification. Ethics and Information Technology, 1(4). p. 295–302.
VAN DER PLOEG, I. (2011) Normative Assumptions in Biometrics: On Bodily Dif-
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

ferences and Automated Classifications. In Van der Hof, S. & Groothuis, M.

(eds.) Innovating Government—Normative, Policy and Technological Dimen-
sions of Modern Government. The Hague: IT & Law Series, T.M.C. Asser Press,
Springer, 29–40.
VAN DER PLOEG, I. & SPRENKELS, I. (2011) Migration and the Machine-
Readable Body. Dijstelbloem, H. & Meijer, A. (eds.) Migration and the New
Technological Borders of Europe. London: Palgrave Macmillan, 68–104.
ZIJDERVELD, M., RIDDERHOF, W. & BRATTINGA, M. (2013) Basis Start Archi-
tectuur van de Vreemdelingenketen: Kennis Delen, Informatie Gebruiken, Samen
Doen. Den Haag: Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

This page intentionally left blank

 Contributors
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Irma van der Ploeg is a Senior Research Fellow at UNU-MERIT, Maastricht

University, where she coordinates the research programme on Social
and Ethical Aspects of ICT-based Innovation. She holds a degree in phi-
losophy from the University of Groningen and received her PhD from
Maastricht University. She has published extensively on philosophical,
normative, social, and ethical aspects of medical technologies and ICTs.
She has been involved in a range of research projects for several Dutch
ministries, the Netherlands Organization for Scientific Research, the
Advisory Council for Science, Technology and Innovation, the Rathenau
Institute, and the European Commission (FP6, FP7). She was principal
investigator of the DigIDeas project ‘Social and Ethical Aspects of Digital
Identities. Towards a Value Sensitive Identity Management’, for which
she received a Starting Grant from the European Research Council. Her
current research interests include STS approaches to social and ethical
aspects of ICTs, in particular security, surveillance, and identification
technologies; responsible innovation through ICTs; and the informatisa-
tion of the body.

Jason Pridmore is an Assistant Professor in the Department of Media and

Communication at Erasmus University Rotterdam. His research interests
are focused primarily on surveillance and consumption, specifically new
media and marketing and the performativity of markets. His research
further examines how these connect with personal identity and practices
of identification and their implications for issues of security. His previous
work has focused on marketing techniques such as the use of loyalty cards
and social media integration as forms of collaborative surveillance. He
is the author of numerous texts on the subject of consumer surveillance,
including the entry on consumer surveillance in the Routledge Handbook
of Surveillance Studies (2012) and the expert report on the surveillance
of consumers and consumption, part of the report on the surveillance
society commissioned by the British Information Commissioner in 2006.
Jason received his PhD from the Department of Sociology at Queen’s
University, Canada, in 2008. Before working at Erasmus University, he
 284 Contributors
was Senior Researcher on the DigIDeas project based in Maastricht, the
Netherlands, and a Post-Doctoral fellow for the New Transparency Proj-
ect within the Surveillance Studies Centre at Queen’s University.

Peter Adey is professor of geography at Royal Holloway University of Lon-

don. His work lies at the intersection between space, security, and mobility
and the blurring boundaries between cultural and political geography. He
has worked at Royal Holloway since 2012, leading the geopolitics and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

security Masters of Science programme with politics and international

relations. He is also chair of the Social and Cultural Geography Research
Group, one of the largest research groups of the Royal Geographical
Society. In 2011, he was awarded a Philip Leverhulme Prize for his con-
tributions to human geography, and his research focuses on the putative
‘new mobilities paradigm.’ He is the author of the Mobility (2009) and
co-editor of the Handbook of Mobilities (2014).

Anders Albrechtslund is an Associate Professor at Aarhus University, Den-

mark. He received his PhD from Aalborg University (2008) and is currently
working on a book about online social networking and self-surveillance.
Recent publications are on topics such as surveillance, technologies, and
Internet practices.

Anne-Mette Albrechtslund is an Assistant Professor at Aalborg University,

Denmark. Her research interests include online communication, digital
identities, and fiction and culture in everyday life. She has published
work about these interests, including the PhD dissertation Storytelling
and Identity in Online Gaming Communities: Exploring Online Culture
and Communication as Narrative Practices (2011).

Dennis Broeders is a Senior Research Fellow at the Dutch Scientific Council

for Government Policy (WRR), an advisory body to the Dutch govern-
ment located within the prime minister’s department, and professor of
technology and society at the department of Sociology of the Erasmus
University Rotterdam. His research broadly focuses on the interaction
between technology and policy, with specific areas of interest in cyber-
security governance and Internet governance as well as surveillance,
databases, and migration control. At the council, he is currently heading
a project on ‘Big Data, Privacy and Security.’ He was previously project
coordinator of the council’s report on the future of government digiti-
sation (iGovernment, Amsterdam University Press, 2011) and recently
published The Public Core of the Internet: Towards a New International
Agenda for Internet Governance (2015, Amsterdam University Press).
He has held visiting research fellowships at the Social Science Research
Centre Berlin (WZB) in 2008 and at the Oxford Internet Institute at
the University of Oxford in 2011. He has published articles in various
 Contributors 285
internationally leading journals in the fields of sociology, political science,
and criminology.

Huub Dijstelbloem is Professor of philosophy of science and politics at the

University of Amsterdam (UvA) and Senior Research Fellow at the Neth-
erlands Scientific Council for Government Policy (WRR) in The Hague.
Bringing together the fields of philosophy of science and political philoso-
phy, his research employs a combination of conceptual, normative, and
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

empirical approaches. He has published on issues relating to border con-

trol and surveillance technologies, on the science–democracy debate, and
on philosophical pragmatism. He is a member of the Advisory Board of
the UvA’s Institute for Interdisciplinary Studies (IIS) and a board member
of the Netherlands Graduate Research School of Science, Technology and
Modern Culture (WTMC). He was previously also a member of the edi-
torial board of Krisis, an open-access, peer-reviewed journal for current
issues in philosophy. He has been affiliated to University of California
San Diego and the Eduardo Mondlane University in Maputo, Centre for
Policy Analysis, Mozambique. He is one of the initiators of Science in
Transition, a movement that aims to reflect on the organization of quality
assessment and social impact of the sciences and on science policy.

Francisca Grommé is a post-doctoral researcher at the Department of Sociology,

Goldsmiths, University of London. Her area of interest is the introduc-
tion of technologies for collecting data about citizens and consumers in a
variety of governmental practices. As a researcher within the ERC–funded
project Peopling Europe: How Data Make a People (ARITHMUS), she is
currently studying changes in the production of official statistics by national
statistical institutes, mainly the census. She is also completing her doctoral
dissertation on the introduction of surveillance technologies in Dutch crime
control at the University of Amsterdam, Departments of Anthropology and
Political Science.

Philip Kirby completed his PhD on US homeland security policy at the

Geography Department at Royal Holloway, University of London, before
taking a research position at the Geography Department at the University
of Exeter. He is currently Research Fellow at the Sutton Trust, a think
tank in London, which seeks to improve social mobility through educa-
tion. His research interests include US homeland security policy, critical
geopolitics, and the politics of UK education.

Karolina La Fors-Owczynik is an external PhD researcher at Tilburg Insti-

tute for Law, Technology and Society (TILT) at Tilburg University in the
Netherlands. She earned her MPhil and LLM degrees from the Faculty
of Law at Tilburg University and her MA in international relations from
Corvinus University in Budapest, Hungary. Her work deals with the con-
troversies inherent in the construction and daily use of digital risk profiles
 286 Contributors
or ‘risk identities’ that are implemented for a variety of preventative
assessment purposes concerning children and migrants in the Nether-
lands. Previously, Karolina has been involved as a junior researcher and
later PhD candidate with the ERC–funded DigIDeas research project
Social and Ethical Aspects of Digital Identities Towards a Value-Sensitive
Design. Furthermore, she participated in the FP7 research projects HIDE,
FIDIS, and STORK, all of which dealt with the ethical, legal, and societal
implications of identity management technologies.
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Peter Lauritsen is an Associate Professor in information studies at Aarhus

University, Denmark. Amongst his research areas are the use of CCTV
by the Danish police force as well as the use of civic registration numbers
in surveillance practices. He was the head of the project Surveillance in
Denmark (2010–2015), which investigated various aspects of the Danish
surveillance culture, drawing on theoretical concepts from actor-network
theory and post-phenomenology. He is currently working on a new proj-
ect focused on how surveillance and social welfare are intertwined. This
includes research on privacy, public perceptions on surveillance, and sur-
veillance as infrastructure. Peter Lauritsen has written several articles on
surveillance as well as the book Big Brother 2.0 (2011). Peter Lauritsen
is an active participant in the public debate on surveillance in Denmark
and serves as columnist for the newspaper Information.

Vlad Niculescu-Dinca is a lecturer in the media and communication depart-

ment of the Erasmus School of History, Culture & Communication in
Rotterdam, the Netherlands. Previously, he participated as a PhD student
in the ERC–funded DigIDeas project Social and Ethical Aspects of Digital
Identities. Towards a Value Sensitive Identity Management. In this con-
text, he examines contemporary technologically mediated policing (social
media surveillance, sensing technologies, mobile communications, geo-
graphic information systems) with an STS–inspired approach. He draws
from ethnographic research, which he previously performed at several
police organizations within the EU. Vlad’s research interests include qual-
itative research approaches to surveillance, privacy, and suspicion and
social and ethical aspects of new information technologies’ design and
use. Vlad holds a master’s degree from the University of Twente in phi-
losophy of technology and society, where he investigated the structure of
ethical debates around new and emerging technologies. He holds degrees
in software engineering from the Technical Universities of Eindhoven
(PDeng) and Bucharest (BSc).

Isolde Sprenkels studied philosophy at Tilburg University and is currently

an external PhD candidate at Maastricht University. Her work concerns
socio-technical configurations in which children’s identities as consumers
in a digitizing society are enacted. In addition, she is a lecturer at Avans
 Contributors 287
University of Applied Sciences, Academy of Marketing and Business Man-
agement, where she also participates in the Centre of Expertise Sustainable
Business. Previously, Isolde worked as a teacher in secondary school and
as a junior researcher at the Infonomics and New Media Research Cen-
tre at Zuyd University of Applied Sciences. She conducted research on
digital identities and identity management systems and co-developed and
taught a minor on digital identity and on marketing and new media. She
contributed to publications on social and ethical aspects of identification
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

technologies such as biometrics used in migration and border manage-

ment. Furthermore, she participated as a PhD student in the ERC–funded
DigIDeas project Social and Ethical Aspects of Digital Identities: Towards
a Value Sensitive Identity Management.

Valerie Steeves, BA, JD, PhD, is an Associate Professor in the Department

of Criminology at the University of Ottawa in Ottawa, Canada. She is
the lead researcher on MediaSmart’s Young Canadians in a Wired World
project (YCWW), which has been tracking young people’s use of new
media since 1999. With Jane Bailey, she co-leads the eGirls Project,
an examination of the performance of gender on social media. She is
also a co-editor of Transparent Lives: Surveillance in Canada, a 2014
multi-disciplinary report that maps out seven main trends in emerging
surveillance practices. Professor Steeves received her JD from the Univer-
sity of Toronto and was called to the Bar of Ontario in 1984.

Tsjalling Swierstra studied philosophy and political science and received his
PhD in philosophy in Groningen; he worked at the technical university of
Twente until being appointed full professor in Maastricht, where he now
heads the philosophy department. He is a member of the Maastricht Uni-
versity Science and Technology Studies (MUSTS) research programme.
He is director of the centre for Ethics and Politics of Emerging Technolo-
gies, a member of the Advisory Committee on Health Research, and a
member of the Program Committee of the Responsible Innovation pro-
gram funded by the Dutch Research Council NWO. He has published
widely on the ethics of new and emerging science and technology (NEST
ethics) and on the mutual shaping of science and technology and mor-
als (technomoral change) in numerous books and in journals like Krisis,
Technology and Human Values, Technology in Society, Nano-Ethics, and
Futures.

Daniel Trottier is an Assistant Professor in the Department of Media and

Communication at Erasmus University Rotterdam. His current research
considers the use of social media by police and intelligence agencies, as
well as other forms of policing that occur on these platforms. As part of
this research, he has participated in two European Commission projects
on security, privacy, and digital media. Daniel has authored several articles
 288 Contributors
in peer-reviewed journals on this and other topics, as well as Social Media
as Surveillance (Ashgate, 2012), Identity Problems in the Facebook Era
(Routledge, 2013), and Social Media, Politics and the State (co-edited
with Christian Fuchs; Routledge, 2014). Daniel previously held post-
doctoral fellowships at the Communication and Media Research Institute
(CAMRI), University of Westminster, the Department of Informatics and
Media at Uppsala University Sweden, and the Department of Sociology at
the University of Alberta, Canada. Daniel completed a PhD in sociology
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

at Queen’s University in Kingston, Canada, as part of the Surveillance

Studies Research Centre.

Govert Valkenburg is currently a Postdoctoral Research Fellow at the Maas-

tricht University Science, Technology and Society (MUSTS) research
programme. His work has concerned social and political aspects of a
wide range of technologies. These technologies have ranged from
human genomics to privacy and security technologies and energy and
sustainability technologies. He has analysed these technologies through
political-theoretical lenses such as public reason, citizenship, and notions
of the political, as well as through lenses from constructivist science and
technology studies, such as actor-network theory and critical theories of
technology. He is currently working on the conceptualization of modes
of governance that can be used in the societal management of sustainable
transitions. Additionally, the mobilization of expertise in the democratic
governance of technological societies is one of his research topics. His
teaching has included social, cultural, and political theory. He is currently
one of the programme coordinators of the Dutch Graduate School of Sci-
ence, Technology and Modern Society WTMC.

Sally Wyatt is Professor of digital cultures in development, Maastricht Univer-

sity, and programme leader of the eHumanities Group, Royal Netherlands
Academy of Arts and Sciences (KNAW). She originally studied economics
(BA McGill, 1976; MA Sussex, 1979) but later did a PhD in science and
technology studies (Maastricht, 1998), which focused on different ways
of transmitting data over networks. She has more than thirty years of
experience in teaching and research about technology policy and about
the relationship between technological and social change, focusing par-
ticularly on issues of social exclusion and inequality. Her current research
interests include digital media in the production of knowledge in the
humanities and the social sciences and the ways in which people incor-
porate the Internet into their practices for finding health information.
She is one of the co-editors (together with Paul Wouters, Anne Beaulieu,
and Andrea Scharnhorst) of Virtual Knowledge. Experimenting in the
Humanities and the Social Sciences (MIT Press, 2013).
 Index
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

Actor Network Theory (ANT) 11, 16, border management 2, 8, 16, 243, 246,
40, 155, 187, 191, 207, 215, 250, 261–3, 266, 276
265 boundaries 21, 57, 75, 89, 107, 112–13,
Adey, Peter 16, 221, 225, 227, 242 131, 141, 150, 156, 200, 255,
Advanced Passenger Information 275
System (API) 262, 266, 269, 276 boundary object 14, 114, 141, 150,
advergames 86–90, 92–4 152, 156
advertising 40, 54, 61, 85, 90–2, 228 boundary work 275
agency 11, 14, 39, 41, 72–4, 83, 191, Bowker, Geoff 11, 17, 78–9, 82, 96,
215, 243, 254, 270 112, 123, 181, 187, 191–2, 204
Akrich, Madeleine 11, 17, 41, 58, 79, Broeders, Dennis 16, 242–3, 252, 257
94, 115, 123, 187, 191, 200, Brubaker, Rogers 7, 17
203, 263, 279 Butler, Judith 9, 17, 31, 35
Albrechtslund, Anders 12, 21, 22, 30,
32, 34 Callon, Michel 53, 59, 77–8, 94, 152,
Albrechtslund, Anne-Mette 12, 21–2, 34 157–8, 191, 204, 254, 258
algorithm 2, 55, 107, 164–5, 174, 182, camera 15, 22–8, 206–15, 234, 248,
247 274
Amoore, Louise 181–2, 242–3, 251–2, care of virtual self 57, 60, 63
254, 256 categories 2, 40–50, 53–6, 78, 79–82,
Andrejevic, Mark 63, 75, 127, 138 92, 107–8, 164, 172, 179–81,
application programming interface 226, 246–7, 251–5, 267
(API) 13, 37–9, 41–2, 53 categorization 12–13, 41, 57, 106,
apps 13, 37 246–7, 255–6, 271
ascription 41 centres of calculation 243, 255
assemblage 40, 110, 235–9 children 13–4, 61, 77–94, 224–5,
asylum 243, 251–2, 256, 262; seeker 232; child abuse 118, 121; as
251, 262–4, 272–7 consumers 79, 83; protection of
authentication 1, 46, 273 104, 184, 188
circulating references 55, 254
Barthes, Roland 23–4, 29–30, 35 citizenship 12, 143, 145, 151
Bauman, Zygmunt 51, 56–8, 245, 257 Clarke, Roger 2, 17, 196, 204, 268,
Bentham, Jeremy 207, 216 279
Big Brother 286, 214 classification 184–8, 190–2, 194, 198,
big data 10, 223, 237, 242, 256 201
biometrics 2, 7, 242–4, 251–2, 273–4, close-circuit television (CCTV) 10,
276–7 206–17
border control 244, 247, 250–2, 256, co-creation 31, 84, 86, 93
261–3, 266, 275 cognitive development 78–80, 82, 90–1
 290 Index
Cold War 227–8 European Border Surveillance System
constructivism: construction 3, 6, 11, (EUROSUR) 243, 248–50
31, 40, 51–2, 62, 82, 92, 109, European Commission 141, 144, 203,
120–1, 126, 155, 215, 256 248, 251, 261
consumers 19, 40, 77–94, 140–2, 179; European Council 261, 264
identity of 77–94 European Dactyloscopy (EURODAC)
consumption 39–40, 51, 52, 56–7, 251–2, 255, 262, 264
78–9, 86, 163 European Parliament 144, 159, 261
contingency 141, 156 European Union 3, 143, 203, 250, 262
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

contingent 6, 11–2, 38–9, 82–3, 92–3, Europol 262

122, 273, 276 evacuation 16, 221–39
Cooley, Charles Horton 8, 17, 51 evacuee: digital 16, 221, 230, 236,
crime: displacement 195–6, 201; 238–9
investigation 206, 265; evidence 61, 109, 264; regimes of 164,
prevention 196; youth 163, 177
169–70, 176, 180 exchange practices 78–9, 83, 85–6,
cumulative exposure 60 89–90, 93–4
cyberspace 21, 250
Facebook 3, 13, 22, 27, 30–3, 38–9,
data: big 10, 223, 237, 242, 256; 42–56, 125–9, 134, 185–6,
corporate 179; mining 83, 199–200
163–72, 180–1, 243, 248, 250, fairness 90–2
254; personal 1, 3, 6, 13, 50–7, false negative 272, 277–8
163, 238, 251, 278; protection false positive 121–2, 272, 275
2, 203, 238, 250 FEMA 230, 232, 234–6
database 1, 2, 12, 39, 103–7, 182–4, feminist 4
189–93, 175–6, 214–15, 223, fieldwork 69–180
242–8, 251–5, 262–4 Fine, Cordelia 90, 92, 98
datafication 242, 254 fingerprint 264, 278
decision making 177, 235, 257 folk theory 15, 206–8, 214
design 2, 12, 14, 16, 30–1, 41, 68, 115, Foucauldian 9–10, 57
142, 153, 155–7, 191, 199–200, Foucault, Michel 9–10, 57, 63, 71, 182,
202–3, 228–30, 247 192, 207, 246
digital persona 2 Foursquare 22, 30, 34, 42
Dijstelbloem, Huub 16, 242–3, 257, FRONTEX 248–50, 255, 262
260, 281 fundamental rights 2, 250
discourse 4–11, 31, 77–94, 105, 140,
142, 166, 207, 261, 263, 278 games 13, 22, 38, 52, 86–90, 140, 155,
domestication 63–4, 74 227
Dublin Convention 262, 264 gaming 86, 88–9, 125, 137, 196, 227
geographic information system (GIS)
e-commerce 7 184, 185, 189, 192–3
e-governance 2 Giddens, Anthony 8, 17, 51, 54
emergency management 16, 222–3, Goffman, Erving 9, 17, 34, 51, 58, 126
231, 238 Goodwin, Charles 167, 170–1, 177,
empowerment 83, 140–3, 154–6 164
enactment 6, 34, 93–4, 148 Grommé, Francisca 14, 163
enrolment 11, 121, 264–9, 274, 276–7
Entry-Exit system 248 Haraway, Donna 11, 17, 167
entry-point paradox 267–8 Helgesson, Claes-Frederik 77–8
ethics 3, 35, 90, 180–1 Homeland Security 221, 237, 246–7
ethnography 83, 165, 185, 206, 214
Europe 2, 14, 16, 142, 144, 147, 224–5, ID-card 5, 268
243–5, 247–8, 251–3, 261–2, 268 identification 7, 278
 Index 291
identity 1, 3–17, 94, 125–6, 129, Livingstone, Sonia 90–1, 130–1, 140–3,
137, 141, 150, 155, 174, 180, 153, 155
224, 235–6, 261–2, 264–5, Lyon, David 6, 10, 52, 62, 63, 75, 182,
271; accuracy of 263, 268–77; 223, 236, 243–5, 247, 251–2,
attributes 50, 52, 53, 198; cards 271
5, 51, 268; consumer 77–94;
digital 6–7, 22, 34, 38–9, 55–6; machine-readable; travel documents
documents 38, 252, 267; fraud 271; zone 266
63, 261–9, 271–3, 276–7; market 39, 40, 50, 77–8, 147–9;
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

identity management 1, 4, 7–8, devices 13, 37–41; practices 78,

14, 52, 112, 126, 262, 278; 90–3; practitioners 77–94
proof of 267–8, 277; provider 8, marketing 2, 13, 54–6, 77–85, 87–94,
52; verification of 261 155, 165–8, 176–81
immigration 16, 243–5, 248, 253, Markham, Annette 21
262–3, 266, 275 material semiotics 164, 263
Immigration and Nationalization Mead, George Herbert 8, 51, 58, 126,
Service (INS) 262 129
immutable mobile 113 media design 142, 153, 155–7
impression management 67 media education 140, 142–3, 145, 151
information and communication media literacy 141–6, 153, 155
technologies (ICTs) 2, 10, 51, mediation 21, 112, 221, 224,
64, 143, 221 254–6; mediating state 242;
information society 3, 149 technological 185–7, 191–3
inscription 79, 94, 226 media users 38–9, 41, 141–2, 148, 152,
Instagram 3, 33, 50 154–7, 202
institutionalization 142, 149, 152, 156, media wisdom 140–157
256 migrants 15–6, 219–25, 243–6, 250,
International Civil Aviation 255, 261–8, 270–8; irregular
Organization (ICAO) 266 246–48, 251–55; undocumented
275
job: market 61; search 70 migration 2, 8, 16, 223, 234, 242–57,
261, 276
Katrina 16, 222–3, 227, 230–7 mobility 15–6, 228–30, 233–4, 237,
Kirby, Philip 16, 221, 242 242–57, 266
Kjellberg, Hans 77–8 Mol, Annemarie 11, 164, 166, 191
knowledge 142–9, 163–7, 172–8; monitoring 6, 8, 10, 15–16, 62–3,
administrative 172; local 113, 66–8, 75, 93, 119, 126, 129–37,
164, 172; professional 104, 184–202, 242–4, 246–8, 253–5
107–8, 180; situated 166–7; tacit multiplicity 141, 154, 156
113, 244, 270
Nairn, Agnes 90, 92, 95
laboratory 246, 254–5 network 11, 15, 22, 26–8, 31–4, 276;
La Fors-Owczynik, Karolina 103, 261 actor see actor network theory;
Latour, Bruno 11–12, 51, 55, 77–8, social 12, 22, 27, 30–3, 71, 77,
113, 144, 187, 191, 194, 200, 125–6, 129, 136–7, 151, 170
207, 214–15, 243, 246, 254, New Orleans 222, 228, 330–5
256, 263 Niculescu-Dinca, Vlad 15, 184
Lauritsen, Peter 206 normalizing practices 78–9, 85, 91–2, 94
Law, John 11, 78, 84, 112, 187, 191–2 normativity 3, 41, 191, 196
law enforcement 2, 5, 16, 63, 75, 104, Norris, Clive 10, 210
111, 120, 243, 262–5, 275–6
legitimization 79, 83, 85, 93, 195 obligatory passage point 152, 156–7,
leisure 26, 78, 145 194
LinkedIn 3, 37, 50, 56 oligopticon 15, 207, 214–16
 292 Index
Orwell, George 208, 209 ProKid 104, 106–107, 109, 111, 115,
Orwellian 209 118
prosumption 39–40, 52
panopticon 207, 214, 215 public 6, 28, 32, 43, 46, 50, 60, 68–70,
parents 31, 66, 71, 87, 106, 108, 115, 73, 88, 200, 225, 236, 242;
117, 126, 129–34, 137, 148, 232 spaces 10, 29, 130, 209
passenger 238, 246; advanced publicity 64, 69
passenger information system
(API) 259, 262, 266–7, 269, queer theory 4
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

274; passenger name record

(PNR) 259 Reference Index 104–7, 114, 119
passports 4, 8, 12, 191, 244, 252, 271 refugees 15, 222–3, 246, 268, 273–4,
peer 14, 22, 60, 63, 67, 74, 81, 125, 279
126–9, 133; pressure 80 registered traveler program (RTP) 279
performance, performativity 9, 11, 22, representational practices 78–9, 83,
30–3, 59, 78, 84, 93, 112, 126, 85–6, 89–90, 93–4
133, 187, 192–5, 198, 201, 214 reputation 63–4, 71–2, 74, 127–9
personal 1–3, 30, 39, 47, 57, 60, 62, responsibility 64, 72, 74, 141–2, 145,
68–69, 70–1, 74, 79, 130, 199, 146, 147, 156–7, 168, 180
254, 267, 271 Ricoeur, Paul 5, 31
Pfitzmann, Andreas 2, 7 risk: assessment 13–4, 16, 103–5,
photograph, photography 23, 25, 26–7, 108–122, 243, 253–4, 261–3,
33, 62, 65, 71, 127, 128, 264, 271–2, 275–6; indicators 2, 14,
265 103, 106, 108, 111, 114–16,
Ploeg, Irma van der 1, 13, 15, 16, 77, 199–22
78, 92, 116, 140, 155, 177, 184,
243, 261, 264, 271, 273, 278 Schengen 247; Schengen Information
police, policing 104, 105, 111, 116, System (SIS) 25, 262
168, 176, 179, 184–7, 191, 195, science and technology studies (STS)
200–1, 206, 210, 214 10–12, 141, 154, 191, 222
policy 103, 105, 141, 144, 148, 155, school 64, 82
163, 165, 169, 172, 175, 178, screens 167, 180, 192; stretched 243–4,
181, 192, 201, 202, 239, 242, 250, 253, 255
245, 253, 256, 274, 276 script 11, 79, 115, 125, 126, 134–6,
postcard 27–8, 32, 33 145, 147, 152, 153, 168
post-colonialism 4, 227 security 199, 216, 222, 223, 224, 236,
practice-based approach 78, 94 244, 245, 249, 252, 266, 267,
pre-emption 252 271
prevention 14, 104, 105, 112, 120, self 5, 6, 51, 52, 54, 60, 63, 125,
196, 261, 263, 270, 277 126, 131, 136, 140; digital 13;
Pridmore, Jason 1, 13, 37, 56, 140, 163 censorship 71; monitoring 8;
privacy 2, 68, 70, 113, 127, 131, care (see care of the self)
136–137, 199, 238; violation 69, Sharing 3, 12, 22, 26, 29, 30, 32, 33,
72, 74, 186, 199, 202 41, 47, 60, 68, 89, 107, 129,
Privacy Impact Assessment (PIA) 276, 131, 133, 135, 149, 196
277 shortlisting method 184, 188, 190,
private 6, 28, 29, 32, 69, 74 196–8, 201–2
profile 6, 30, 38, 42, 46, 48, 49, 50, socialization 91
52–3, 57–8, 62, 64, 72, 89, 163, sociality 13, 22, 32, 33, 63, 65
165, 178, 179, 200 deactivation social media 22, 33–4, 37–9, 41,
of 65; profiling 226, 243, 254, 50–53, 54–58, 62, 68, 73–5,
262, 276; risk profile 107, 114, 132, 237; monitoring of 63, 185,
116, 118, 263, 267, 269, 270 190, 194, 197–8, 201–2
PROGIS-console 262, 265, 268, 275 social sciences 163, 173, 178, 181, 192
 Index 293
social sorting 12, 192, 245 translation 11, 24, 94, 113, 115,
social theory 3, 8–9 141–2, 148, 151, 156, 187,
Sontag, Susan 27, 33 200–1, 271, 277; chain of 78
Sprenkels, Isolde 77, 140, 271 Turkle, Sherry 21, 51
standardizing, standardization 40, 46, Twitter 30, 43, 46–50, 89, 165, 185,
55, 56, 149, 151–2, 181, 255, 195, 197, 213
270
Star, Susan Leigh 11, 78, 82, 112, 114, United States 142, 225, 226, 233, 238,
141, 150, 152, 187, 194 245, 246
Downloaded by [University of California, San Diego] at 19:23 09 January 2017

state: mediating state 242, 243, 244, Urry, John 23, 24, 25, 28, 223
246, 252, 254–7 users 11, 21–2, 30–1, 38–41, 45, 50,
Steeves, Valerie 77, 87, 125, 127, 130, 52, 60–3, 67, 70, 72, 141–2,
135, 197 148, 152, 154–7, 200
Strathern, Marilyn 164, 166
surveillance 6, 52, 61, 62–3, 75, 89, Valkenburg, Govert 103
126, 129, 132, 134, 137, 167, Valkenburg, Patti 9, 126, 131, 133
185–7, 197–8, 201–2, 206, 214, verification 5, 43, 46, 54, 263–5, 271,
216, 243, 245, 262; cameras 272–4, 277
206, 209, 213; studies 9–10, visa 244, 251–3, 255; Visa Information
207, 222, 223 System (VIS) 251, 255, 262
suspect 168, 171, 173, 176–7, 235, 275 visibility 13, 57, 60, 62, 65, 68, 74–5,
suspicion 15, 107, 122, 201, 202, 210, 126, 136, 263, 271
211, 226, 252, 265; categorical 10 visualization 166, 192, 230, 243–5,
Swierstra, Tsjalling 184 255, 271, 275

technological determinism 11, 141, Wyatt, Sally 141, 155

209–10
technology 3, 10, 22, 26, 34, 38, 40, youth 68, 80, 83–5, 143, 148, 163–4,
51–2, 54, 64, 65, 83, 92, 121, 175; youth care 103–5, 111,
126, 141–2, 155–6, 163, 166, 113, 120–1, 164, 242; youth
186, 190–1, 201, 215, 223, 238, groups 184–5, 190–3, 200–2;
242–4, 249, 256, 275 problem youth 169–70, 172–4,
tourism 23, 24, 26, 28, 33, 245 179–80, 184–5, 190–3, 200–2