Diversity of Methodological Approaches in Social Sciences Example of The Analysis of Media and Online Information Inna Lyubareva Full Chapter

Diversity of Methodological
Approaches in Social Sciences:

Example of the Analysis of Media and
Online Information Inna Lyubareva
Visit to download the full and correct content document:
https://ebookmass.com/product/diversity-of-methodological-approaches-in-social-scie
nces-example-of-the-analysis-of-media-and-online-information-inna-lyubareva/
Diversity of Methodological Approaches in Social Sciences
Modeling Methodologies in Social Sciences Set
coordinated by
Roger Waldeck
Volume 3
Diversity of Methodological
Approaches in Social
Sciences
Example of the Analysis of

Media and Online Information
Edited by
Inna Lyubareva
Roger Waldeck
First published 2023 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced,
stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentioned address:
ISTE Ltd John Wiley & Sons, Inc.

27-37 St George’s Road 111 River Street
London SW19 4EU Hoboken, NJ 07030
UK USA
www.iste.co.uk www.wiley.com
© ISTE Ltd 2023

The rights of Inna Lyubareva and Roger Waldeck to be identified as the authors of this work have been
asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the
author(s), contributor(s) or editor(s) and do not necessarily reflect the views of ISTE Group.
Library of Congress Control Number: 2023936103
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78630-901-3
Contents
Overview: Media, Information Pluralism and Methodological

Choices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Inna LYUBAREVA and Roger WALDECK
Chapter 1. Online Platforms and Analysis of Community

Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Cécile BOTHOREL, Laurent BRISSON and Inna LYUBAREVA
1.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Outline: Ulule and YouTube platforms . . . . . . . . . . . . . . . . . . . 4
1.3. Exploring the social network and communities on the Ulule
platform. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1. Platform, data and construction of the social graph . . . . . . . . . . 6
1.3.2. Algorithmic approaches for the detection of communities
on Ulule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.3. Typology of community forms . . . . . . . . . . . . . . . . . . . . . . 12
1.4. Exploring the social network and communities on the YouTube
platform. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.4.1. YouTube: data and construction of interaction graph . . . . . . . . . 18
1.4.2. Construction of evolving communities . . . . . . . . . . . . . . . . . 21
1.4.3. Community evolution graph . . . . . . . . . . . . . . . . . . . . . . . 24
1.4.4. Description of the evolutionary forms of communities . . . . . . . . 27
1.4.5. Forms of community development on YouTube . . . . . . . . . . . 28
1.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.6. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
vi Diversity of Methodological Approaches in Social Sciences
Chapter 2. Echo Chambers and Opinion Dynamics:

An Agent-Based Modeling Approach . . . . . . . . . . . . . . . . . . . . . 43
Julien MÉSANGEAU and Roger WALDECK
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.2. Agent-based modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.2.1. ABM modeling principles . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.2.2. The modeling cycle and the ODD representation . . . . . . . . . . . 50
2.3. The sociology of echo chambers . . . . . . . . . . . . . . . . . . . . . . . 52
2.3.1. A descriptive definition of the echo chamber . . . . . . . . . . . . . 53
2.3.2. Reducing the echo chamber phenomenon into simple
factors for modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.4. Modeling echo chambers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.4.1. The Netlogo model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.4.2. ODD for the echo chamber model . . . . . . . . . . . . . . . . . . . . 59
2.4.3. A short review on the modeling of opinion dynamics . . . . . . . . 64
2.4.4. Results of the Netlogo model simulations . . . . . . . . . . . . . . . 68
2.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
2.6. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Chapter 3. Diversity of Sources and Pluralism of

Journalistic Framing: A Qualitative–Quantitative
Sociosemiotic Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Dario COMPAGNO and Emmanuel MARTY
3.1. The sources: interfaces between writing and reality . . . . . . . . . . . . 83
3.2. Theoretical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.2.1. Framing theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.2.2. Toward a typology of sources . . . . . . . . . . . . . . . . . . . . . . 89
3.2.3. Citing sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.2.4. Formalizing hypotheses about discourse . . . . . . . . . . . . . . . . 94
3.3. Processing and analysis of the corpus . . . . . . . . . . . . . . . . . . . . 95
3.3.1. Scope and corpus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.3.2. Processing procedures and analysis . . . . . . . . . . . . . . . . . . . 96
3.3.3. Processing and corpus enrichment . . . . . . . . . . . . . . . . . . . . 97
3.3.4. Analysis and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.4. Discussion and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.5. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Contents vii
Chapter 4. Interviews as a Research Method for

Understanding Online Information Pluralism . . . . . . . . . . . . . . . . 115
Pauline AMIEL and Alexandre JOUX
4.1. Introduction: the tools used by qualitative approaches
in journalism studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.2. Semi-structured interviews to gather the views of stakeholders . . . . . 117
4.2.1. Interview methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.2.2. The interview as a means to understanding the dynamics
of online information pluralism . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.4. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Chapter 5. Media and Information Pluralism Through the

Lens of Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Dominique BOUGEROL
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.2. Pluralism in European law . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.2.1. Council of European law . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.2.2. European Union law . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.3. Pluralism in French law . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5.3.1. Pluralism in the constitution . . . . . . . . . . . . . . . . . . . . . . . 147
5.3.2. Pluralism in legislation. . . . . . . . . . . . . . . . . . . . . . . . . . . 151
5.4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.5. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Overview
Media, Information Pluralism

and Methodological Choices
O.1. Introduction to the theme of digital transformation of media

and information pluralism
The debate about information pluralism – a guarantee of a well-

functioning democracy – is not new. It refers to the issues of freedom and
independence of the press, of the sufficient production of varied opinions in
line with the different communities of readers, listeners or viewers (Hiller
et al. 2015). Traditionally focused on issues of media concentration (Napoli
and Gillis 2006; Helberger 2008), the theoretical definition of pluralism
includes different elements (Napoli 2001; Benhamou and Peltier 2006;
Rebillard 2012a; Napoli and Karppinen 2013): sources (the range of content
providers), content (the diversity of the kinds of information or opinions
being issued), exposure of individuals to alternative viewpoints and news
feeds, variety (the number of news topics covered in a given time period),
balance (the distribution of these topics between those that focus attention
on the front page and those that are much more isolated) and disparity (the
differences in journalistic treatment of the same news topic). Recent work
adds other elements, such as the informational richness and added value of
online content (Lyubareva et al. 2020).
Overview written by Inna LYUBAREVA and Roger WALDECK.

The various chapters and works cited in this book are part of the research program known as
Pluralisme de l’information en ligne (“Pluralism of online information”, or PIL) supported by
the French National Research Agency (ANR-17-CE27-0010, 2018-2022).
x Diversity of Methodological Approaches in Social Sciences
The current discussions on pluralism take on a new importance due to the

digitization of media and contemporary media concentration movements and
extend it, a novelty of the digital era, to the intervention of a distinct power:
the GAFAMs (Google, Apple, Facebook, Amazon, Microsoft, etc.). They
implement different ways of controlling information and conditioning
opinions and raise new questions for regulation.
In particular, the development of digital platforms and services has lowered

the barriers to entry for the production and distribution of information because
of the reduced need for capital needed to create and maintain a newspaper, the
increased decentralization of production sources and a sharp reduction in
distribution costs. These factors have facilitated the arrival of new entrants and
the multiplication of digital business models (Lyubareva and Rochelandet
2016). They also favor new informational practices linked to the ease of
accessing, sharing and publishing information online through digital social
networks and aggregators. This movement is thus a priori very favorable to
information pluralism via the diversification of sources and content, the
dissemination of data or the appearance of new forms of journalism.
If such an abundance of content and sources can favor the expression and
formation of plural opinions, it is not however synonymous in itself with quality
information and, consequently, pluralism. There are many examples and
obstacles to quality information, starting with fake news, not to mention the
consequences of transformations in the conditions of production (speed-driven
journalism1), use (snack content2) and the dissemination of information induced
by and subject to the audience data produced by digital platforms and
increasingly filtered by their algorithms, and the massive reproduction of content
by central agencies such as AP, Reuters and AFP (Paterson 2007; Fenton 2009;
Redden and Witschge 2010; Marty et al. 2012; Lyubareva et al. 2020).
This proliferation leads us to rethink the meaning and relevance of

information pluralism. We are moving from a situation of scarcity to an
abundance of news information, making diversity issues more problematic,
which, in a first analysis, would no longer be posed in the same terms
(Lyubareva and Rochelandet 2017). The modalities of news selection have
evolved: traditionally and until recently, journalists, newsrooms and press
1 That is, journalism focused on the speed of information production.

2 Short and fragmented information that goes straight to the point, “snack content”, often
adapted to mobile media and social networks.
Overview xi
institutions have played the role of gatekeepers by filtering news according

to professional standards. Although they continue to play this role, with the
development of the Internet they are no longer the only ones, and new actors
contribute to this selection of information in often different ways: through
the provision of search engines, through the use of social media, or through
algorithm-based recommendation systems.
Yet these platforms change the rules of the game in order to maximize the
value extracted from the use of the proposed services (Cardon 2013).
Largely financed by the exploitation of users’ personal data, platforms seem
to conform individuals according to their opinions – by exposing them to
information that matches their profiles – not offering alternative points of
view that are likely to make their opinions evolve and to create debates (see
Eli Pariser’s “filter bubble” or Sunstein’s “echo chambers”). These new
actors are less concerned with information pluralism (Napoli 2011; Vos and
Heinderyckx 2015) than with audience and personal data collection of
readers and digital tools; they guide the choices of journalists themselves as
to which formats and news topics to cover and which to highlight. Pluralism
could thus be affected by the strategies of digital platforms and the
redistribution in communication power (Helberger 2011).
Therefore, it is worth questioning to what extent this increase in content

is not counterbalanced by a decrease in the pluralism of information being
produced, accessed, and consumed (Lyubareva and Rochelandet 2017). A
number of theoretical analyses establish important gaps in this area and the
urgent need to focus attention on the analysis of information in circulation
and its transformative factors (Napoli and Gillis 2006; Karppinen 2009;
Napoli 2011). They point out that the notion of pluralism and its evaluation
need to be redefined, on the one hand, because the assumption that diversity
(as a synonym for quality) is socially desirable needs to be nuanced: it
depends precisely on the definitions one chooses (Karppinen 2018) and the
underlying conception of democracy (Carpentier and Cammaerts 2006).
Approaches to pluralism can thus vary according to the objectives assigned
to journalism: informing different readership, promoting public debate and
the formation of an informed/enlightened public opinion according to the
deliberative democracy model, or fostering the emergence of critical and
power-challenging viewpoints with more radical journalism. On the other
hand, the notion of pluralism must take into account the relationship between
the diversity offered and the diversity consumed (Napoli 2011): does a great
xii Diversity of Methodological Approaches in Social Sciences
diversity of sources mechanically improve the diversity of content, which in

turn, by increasing, would promote the diversity of exposure given that the
readership or audience would have more options? Nothing obvious here.
Moreover, the supposed diachronic relation between a diversity of “news”
offered and a diversity of opinions is not self-evident: more choice of news
does not mechanically stimulate more enlightened and more diverse opinions
and vice versa. Everything depends on the economic models of the
intermediaries of the information market and on the practices of the
individuals who can be locked in bubbles (“filter bubbles”, Pariser 2011) or
echo chambers (Sunstein 2018) through a structuring operated by social
media and personalized recommendation tools contributing to the reduction
of sources, interests and opinions of the individuals in spite of the abundance
of news.
Finally, there is nothing obvious about the link between the conditions of
production (forms of media ownership, journalistic practices, revenue
models, etc.) and the characteristics of the information actually produced.
The massive adoption of digital tools and uses has led to the emergence of
new production practices (new formats, original fact-checking devices,
platformization and search for network effects, etc.) and has allowed for the
entry of new information producers ranging from pure players3 to the readers
themselves, while favoring the entry of new voices (partisan media,
independent or activist journalists, etc.). New alternative actors can
contribute to increasing the diversity of viewpoints while producing news
based on “hearsay” just as traditional reputable news titles can cover a very
wide range of topics while multiplying the identical repetition of central
agency dispatches (Lyubareva et al. 2020). Alternative forms of financing
(crowdfunding, donations) can both contribute to the production of rich and
original information and have a negative impact on the disparity of topics
covered by the media (Cariou et al. 2017)4. In the same vein, the fact that
content is produced by “in-house” journalists does not automatically
guarantee its quality. Therefore, the analysis of the link between the
conditions of production and the characteristics of the information produced
must be part of a detailed empirical study; however, work on this subject
remains very rare (Karppinen 2018).
3 A pure player is an information company that uses exclusively digital media for its
distribution, without a paper edition.
4 Indeed, too much engagement by reader-contributors can create a new form of dependency
for newspaper titles, more than the dependency on advertising revenues.
Overview xiii
In summary, the digitization of the media and the current explosion of

information content on the Internet highlight the reductive nature of the
traditional debate on pluralism, focused mainly on the concentration of
capital in the press, and raise new regulatory issues. As a result of these
changes, some works on the conceptualization of media pluralism in
the current context advocate that studies on this topic should include, beyond
the content itself, the economic, social and regulatory dimensions that have
proven to be insufficiently addressed by the literature (Van Cuilenburg 2007;
Aslama and Napoli 2010; Napoli 2011; Jakubowicz 2015).
For the reasons presented above, the analysis of media and information
pluralism in the digital context takes the form of a complex and evolving
object of study, calling for a systemic analysis of the different socioeconomic
dimensions where the strategies of producers, the informational practices of
consumers and the forms of concentration of the written and audiovisual
media are articulated to give rise to new issues of pluralism in circulation. The
analysis of such an object requires the implementation of appropriate and
necessarily varied theoretical and methodological approaches offering an
explanatory complementarity of the studied phenomena.
O.2. Methodological choices in social science research on digital

transformations
In the social sciences, traditional paradigms are regularly questioned

(Kirman 1989, 1992) and the construction of a research project is often
accompanied, upstream, by a phase of reflection on the theoretical and
methodological frameworks best able to capture new forms of interaction,
and the multi-scale dynamics that underlie them. For example, in economics,
methodological individualism has long constituted the traditional basis of
explanations for industrial phenomena – where any social or economic fact
must be understood from the behavior of individuals – implying a
reductionism whose extreme representation has been the use of
representative agents. In fact, since the Sonnenschein–Mantel–Debreu
theorem, the hypothesis of the existence of equilibrium as a stable and single
state of the economy is not guaranteed, with the implication that the
techniques of comparative statics used by economists are called into
question. A strong version of methodological individualism that constructs
social facts solely on the basis of egoistic individuals acting under purely
xiv Diversity of Methodological Approaches in Social Sciences
material constraints has been particularly criticized because it neglects a

fundamental aspect, that is the structure of interactions (Udehn 2001); it is
not so much the nature of individual rationality that influences the nature of
emergent properties as the structure of the networks of relations (Granovetter
1985) and at the very least, structural individualism must be at the basis for
the understanding of the social facts. Game theory has been an initial
response to the integration of interaction structures, however, it remains
constrained by aspects linked to the computability of equilibria involving,
more often than not, completely random encounters between players, each
player having the same chance of meeting any other player. Finally, a
methodological individualism that integrates institutional constraints (social
norms, legal norms, social preferences, etc.) into the explanation of
individual behavior is certainly more representative of the observed social
behavior (Camerer et al. 2004; Bowles and Gintis 2011).
Today, the massive diffusion of digital technologies and their uses

contributes to the profound transformation of many economic and social
activities and is at the origin of the emergence of new industries and sectors.
It seems even more difficult to apprehend the growing variety of these
mutations from postures based on a strong methodological individualism.
Given the increased connectivity and interdependencies between actors and
industries, the dynamics propagate at different scales (micro, meso and
macro) with multiple feedback loops from micro to macro, and from macro
to micro, through interaction structures that are themselves emerging from
the effect of individual actions.
One of the important consequences of the associated epistemological

approach is that it is necessarily multidisciplinary, as opposed to purely
disciplinary research, because it is driven by the object of study, which may
represent a market or an industry, a technology or its appropriation, a form
of social interaction or an economic model, etc. This multidisciplinarity
manifests itself both through the use of varied methodological approaches
and through disciplinary cross-fertilizations providing many different points
of view on a common object of study. By mutually enriching each other,
these cross-disciplinary approaches build interdisciplinary research focused
on a common object of study (Morin 1994; Nicolescu 1996).
At the heart of this approach is a fruitful collaboration between different

social sciences, but also other disciplines, coming notably from the
computational and natural sciences. For example, we can mention the
Overview xv
development, in addition to the hypothetico-deductive approach, which is

classical in the social sciences, of an inductive dimension carried out in
particular by data analysis based on machine learning techniques. It is
therefore important to articulate these theoretical reference frames coming
notably from the social sciences with the reality of the field of observation
for the formulation of the research problem and hypotheses, and to adapt the
methodologies of analysis according to the potential contributions and
strength of each of them (using, for example, the anchored method in
qualitative analyses or data mining in quantitative studies). The contribution
of natural science is particulartly important for the study of the complexity
of social systems. Methodologies from statistical physics and evolutionary
theories from biology have had an impact in different fields of social
sciences, ranging from the study of market dynamics to issues of social and
moral dilemmas (Bowles and Gintis 2011). On the other hand, cognitive
neuroscience has recently permeated the experimental approaches
underlying behavioral economics and psychology in the study of
understanding human action (Glimcher and Fehr 2014).
When we talk about the digital context – characterized by actors

interacting in instantaneous conditions, new forms of many-to-many
communication, an unprecedented pace of innovation affecting new services,
formats, uses and modes of production, as well as emerging industries – the
adoption of a research approach centered around an object of study to
analyze socioeconomic transformations comes up against a certain number
of additional epistemological impediments.
For instance, when it comes to the construction of a research project, one

of the first and foremost steps is the perimeter definition of the object under
analysis, that is, the characteristics and micro-components that structure it,
and the transformations that characterize it. This involves answering the
following questions: How do we establish the observation period and time
measurement? How do we identify the determinants and structuring
elements of phenomena when they may be latent or emergent in nature? A
lack of attention to these choices can greatly limit the validity of the research
results and their generality. The approach can call upon various techniques
to extract this information. In this perspective, the availability of big data
sources or inductive qualitative approaches may provide, in certain contexts,
an advantage for empirically driven approaches over model-driven ones.
Furthermore, the definition of the structuring characteristics of the object of
study will be influenced by the disciplines that come together to define and
xvi Diversity of Methodological Approaches in Social Sciences
understand it, just as these disciplines will have to agree on the validation of
a particular theory or interpretation of data. By way of example, the digital
transformations of media and information can be studied from the economic
perspective with a focus on business models, from the management and
sociological perspectives putting forward new forms of organization and
interaction between actors, or from a legal viewpoint with an interest in
regulatory mechanisms. Each of these disciplines brings a complementary
angle of analysis to the object of study.
In addition, given the extremely rapid pace of innovation and

transformation, it seems important to us to deepen the temporal (and
spatiotemporal) approach in social science research in general, and on
information pluralism in particular. Indeed, in the context of permanent
mutations of the technical social systems, a temporal approach allows us to
identify and describe, for example, transformations in the economic models
of the producers of information, the formation of new public spaces of
interaction, the evolution of regulatory devices, etc., and to understand their
multi-scale dynamics. On this basis, the modeling of a phenomenon can be
possible after the formulation of stylized facts, that is, a “simplified”
representation of the world. This modeling approach is particularly useful for
predicting certain trends and global evolutionary trajectories in a theoretical
manner and for formulating recommendations.
Using the example of media and information pluralism, this book aims to
present a variety of methodological approaches that can be applied to other
objects of study from the social sciences. Through its different chapters, the
book proposes and critically analyses some concrete examples of appropriate
methodologies. Our objective is to identify significant methodological issues
and avenues of thought, some of which are under-utilized in current social
science research.
O.3. Introduction to the chapters in this book
The analysis of media and information pluralism as a complex and

evolving research object may require an interdisciplinary openness in order
to take into account, in addition to the characteristics of journalistic contents,
the new economic constraints, the social interaction mechanisms and forms
of discourse, as well as the functioning and specificities of online practices
and networked technologies. This book aims at presenting and putting into
Overview xvii
perspective different methodological approaches, some of which are well

known in the social sciences as well as others which remain in short supply,
that have proven their relevance to studies on the different dimensions of
media and pluralism. The book is structured into five chapters, each dealing
with an approach associated with the analysis of a particular dimension.
Each chapter is devoted to a particular methodology and aims, beyond a

general presentation of the principles and reference works, to put into
perspective its advantages and limitations for the analysis of the issue of
interest. The book will cover a wide range of methods: qualitative methods,
agent-based modeling, lexicometric content analysis, social network analysis
and the legal approach. Through its different chapters, the book will
highlight that the choice of a method is never neutral, neither for the problem
under analysis, nor for the results (Waldeck 2019). It will address the
question of crossing different methods, as well as the problems that often
accompany an openness to interdisciplinarity (Waldeck 2019). In short, it
will provide access keys to these different methodologies to an audience of
researchers from different social science disciplines. This review of
methodological works is built around three key thematic axes of the digital
transformation of media and information in circulation, namely (1) the role
of online platforms, (2) the new conditions of information production and
(3) the legal issues surrounding information pluralism.
Axis 1: Platforms master the function of infomediation by putting

Internet users in contact with all types of online information and with other
Internet users. An abundant literature has been developed on the functioning
of online groups. This work emphasizes the role of platforms that promote
the emergence of user communities based on repetitive interactions, a
principle of homophily and shared interests (Rheingold 2000; McPherson
et al. 2001; Cohendet et al. 2003; Von Hippel 2005). In these collective
spaces, users come to occupy a hybrid position between consumers,
producers of information and prescribers of opinions. Very often, the
socioeconomic analysis of online communities first focuses on a definition
of their perimeter, for example, from the point of view of the interests or
skills shared by its members (Wenger 1999). Examples include open-source
communities where the group’s perimeter is defined within the framework of
a project, blogs and wikis with their information production communities
and group boundaries defined on the basis of contributions and uses of
specific content, or forums of players who share the same passions and
interests.
xviii Diversity of Methodological Approaches in Social Sciences
The first two chapters of this book focus on the detection and analysis of
new public spaces of interaction within online platforms. In line with the
existing literature, it is assumed that user information preferences and, more
generally, the diversity of information to which individuals are exposed may
be influenced by factors related to sociability, interactivity or a sense of
community in addition to the intrinsic characteristics of the information
produced by the media.
More precisely, Chapter 1 shows how social network analysis tools allow
us to detect a “hidden” aspect of the interactions between platform users,
which takes the form of the formation of latent communities, at the inter-
channel, inter-video or inter-project level, which evolve over time and are
likely to orient the users’ information choices, their behaviors and their
opinions. Compared to the methodological approaches of existing works, the
originality of this approach lies in the detection and analysis of the
evolutionary dynamics of interactions within user groups, which are not
directly observable at the platform level. This analysis calls for a novel
intersection of two methods: social network analysis and community
dynamics using data mining. It also discusses how it can be articulated with
other tools, such as the discourse analysis of informal language and
qualitative methods.
Chapter 2 proposes the agent-based modeling approach to analyze the

formation of echo chambers. An agent-based model is a computer model that
allows simulation of the actions and interactions of autonomous agents,
representing individuals and/or groups of individuals of the real system, in
order to understand the behavior of this system. Using an abductive method
of iterative process of hypotheses testing, this chapter demonstrates how
agent-based simulations permit the identification of factors which have
sufficient explaining power for the phenomena of interest. In particular, for
the analysis of the echo chambers’ emergence, Chapter 2 tests, among other
factors, the convergence of opinions, polarization rate, tolerance of exposure
to other opinions and the formation of friendship links on social networks.
Axis 2: A number of studies have evaluated information pluralism

through the media bias created by various factors related to market structure,
in particular media concentration, with the hypothesis that concentrated
markets would produce a variety of news that is suboptimal in relation to
demand. On the other hand, preserving competition between press titles,
Overview xix
radio stations or television channels would be more conducive to the

production of diversified and quality programs. In the same vein, media
revenue models based on advertising are often considered in the literature as
representing risks for the pluralism of opinions and information (Anderson
and Gabszewicz 2006; Gabszewicz and Sonnac 2006; Garcia Pires 2014).
However, few studies are interested in a detailed analysis of the link
between, on the one hand, the conditions of production (i.e. the forms of
media ownership; their belonging to an institutional category such as
national or regional press, pure players, alternative media, etc.; and the
economic models on and offline) and, on the other hand, the editorial policy
of the media. Such an analysis requires the use of appropriate methods.
Chapters 3 and 4 aim to fill this gap by presenting two different
methodological approaches, respectively, a sociosemiotic approach and a
qualitative analysis by semi-structured interviews.
In order to study the pluralism of sources and journalistic framings in the

articles of different types of media, Chapter 3 develops an original research
combining qualitative and quantitative methods and crossing the textual
analysis of discursive traces with the socioeconomic specificities of the
media (periodicity, socioeconomic models, editorial policy). Starting from
media categories – national and regional press; print and online publication;
daily and periodical periodicity; pure players and news agencies – and various
contents (in order to grasp a plurality of information), the semi-automatic
approach detailed in this chapter allows an understanding of which types of
media often cover events according to the same framings by using the same
types of sources.
Chapter 4 focuses on one of the most widely used tools in journalism and
news media studies: the semi-structured interview. It puts into perspective
the relevance of this classic approach to identify representations of media
independence, particularly from the point of view of the business model,
and of information pluralism within the different media. This chapter
demonstrates how to build a corpus and how to collect relevant discourses
from different types of interlocutors: media group managers (CEOs, editorial
directors, marketing directors, advertising directors, etc.), editors and
journalists. The dynamic approach presented in the chapter begins with the
description of data to understand the general environment that frames the
production of information, then passes through the sorting and classification
stage to structure the corpus, and concludes with the interpretation stage to
give meaning to the observations. It contributes to a better understanding of
xx Diversity of Methodological Approaches in Social Sciences
the objects explored, that is, the perception by the media of pluralism and
quality information and the place of these actors in the current media
ecosystem.
Axis 3: Finally, the last chapter of this book aims to put into perspective
the legal approach to the analysis of media, information and pluralism. The
media and the information they produce play a fundamental role in the
formation of citizen opinions and the proper functioning of democracies.
However, the new practices of economic actors brought about by the advent
of the Internet could radically challenge the traditional forms of State
intervention, calling for a re-evaluation of public policies and even their
redesign through new instruments and the proposal of new forms of
regulation. These adaptations require, upstream, an operational definition of
information, media and pluralism, that is, through the study of legal sources,
the extraction of consensual criteria for the evaluation and regulation of
these key concepts. Therefore, Chapter 5 demonstrates why, despite the
abundant presence of these concepts in legal and regulatory texts, their
characterization and analysis in the field of law are relatively problematic. Is
it possible to approach information pluralism and media pluralism as purely
legal concepts? This is the methodological question discussed in the last
chapter of the book.
O.4. References
Anderson, S.P. and Gabszewicz, J.J. (2006). The media and advertising: A tale of
two-sided markets. Handbook of the Economics of Art and Culture, 1, 567–614.
Aslama, M. and Napoli, P.M. (2010). Diversity 2.0: Rethinking audiences,
participation, and policies. McGannon Center Working Paper Series, 27.
Benghozi, P.J. and Lyubareva, I. (2014). When organizations in the cultural
industries seek new business models: A case study of the French online
press. International Journal of Arts Management, 16(3), 6.
Benhamou, F. and Peltier, S. (2006). Une méthode multicritère d’évaluation de la
diversité culturelle : application à l’édition de livres en France. In Création et
diversité au miroir des industries culturelles, Greffe, X. (ed.). La Documentation
française, Paris.
Bowles, S. and Gintis, H. (2011). A Cooperative Species: Human Reciprocity and
Its Evolution. Princeton University Press, Princeton and Oxford.
Overview xxi
Camerer, C., Loewenstein, G., Rabin, M. (2004). Advances in Behavioral

Economics. Princeton University Press, Princeton.
Cammaerts, B. and Carpentier, N. (2006). Reclaiming the Media: Communication
Rights and Democratic Media Roles. Intellect Books, Bristol.
Cariou, C., Lyubareva, I., Rochelandet, F. (2017). Crowdfunding et qualité de
l’information. Réseaux, 5, 23–56.
Cohendet, P., Créplet, F., Dupouët, O. (2003). Innovation organisationnelle,
communautés de pratique et communautés épistémiques : le cas de Linux. Revue
française de gestion, 5, 99–121.
Dominique, C. (2013). Dans l’esprit du PageRank. Une enquête sur l’algorithme de
Google. Réseaux, 1, 63–95.
Gabszewicz, J. and Sonnac, N. (2006). L’industrie des médias. La Découverte,
Paris.
Garcia Pires, A.J. (2014). Media diversity, advertising, and adaptation of news to
readers’ political preferences. Information Economics and Policy, 28(C), 28–38.
Glimcher, P.W. and Fehr, E. (eds) (2014). Neuroeconomics: Decision Making and
the Brain, 2nd edition. Elsevier, Academic Press, London.
Granovetter, M. (1985). Economic action and social structure: The problem of
embeddedness. American Journal of Sociology, 91(3), 481–510.
Helberger, N. (2008). From eyeball to creator – Toying with audience empowerment
in the audiovisual media service directive. Entertainment Law Review, 9(19),
128–137.
Helberger, N. (2011). Diversity by design. Journal of Information Policy, 1(1),
441–469.
Hiller, R.S., Savage, S.J., Waldman, D.M. (2015). Market structure and media
diversity. Economic Inquiry, 53(2), 872–888.
Jakubowicz, K. (2015). New media ecology: Reconceptualizing media pluralism.
In Media Pluralism and Diversity, Valcke, P., Sükösd, M., Picard, R.G. (eds).
Palgrave Macmillan, London.
Karppinen, K. (2018). Journalism, pluralism, and diversity. In Journalism, Vos, T.
(ed.). De Gruyter Mouton, Berlin, Boston.
Kirman, A. (1989). The intrinsic limits of modern economic theory: The emperor
has no clothes. The Economic Journal, 99(395), 126–139.
Kirman, A. (1991). Epidemics of opinion and speculative bubbles in financial
markets. In Money and Financial Markets, Taylor, M. (ed.). Macmillan, London.
xxii Diversity of Methodological Approaches in Social Sciences
Lyubareva, I. and Rochelandet, F. (2016). L’évolution des modèles d’affaires dans

les industries créatives : l’exemple de la presse en ligne en France (2004–
2014). Revue d’économie industrielle, 156, 123–157.
Lyubareva, I. and Rochelandet, F. (2017). Modèles économiques, usages et
pluralisme de l’information en ligne. Réseaux, 5, 9–19.
Lyubareva, I., Benghozi, P.J., Fidele, T. (2014). Online business models in creative
industries: Diversity and structure. International Studies of Management &
Organization, 44(4), 43–62.
Lyubareva, I., Rochelandet, F., Etienne, J.-M. (2018). The Evolution of Business
Models in the Online Press: Between Creativity and Imitation. Collection ICCA,
Peter Lang.
Lyubareva, I., Rochelandet, F., Haralambous, Y. (2020). Qualité et différenciation
des biens informationnels. Une étude exploratoire sur l’information
d’actualité. Revue d’économie industrielle, 4, 133–177.
Marty, E., Rebillard, F., Pouchot, S., Lafouge, T. (2012). Diversité et concentration
de l’information sur le web. Réseaux, 6, 27–72.
McPherson, M., Smith-Lovin, L., Cook, J.M. (2001). Birds of a feather: Homophily
in social networks. Annual Review of Sociology, 27(1), 415–444.
Moreau, F. and Peltier, S. (2004). Cultural diversity in the movie industry: A
cross-national study. Journal of Media Economics, 17(2), 123–143.
Morin, E. (1994). Interdisciplinarité et transdisciplinarité. Transversales, science,
culture, 29, 4–8.
Napoli, P.M. (2001). Foundations of Communications Policy: Principles and
Process in the Regulation of Electronic Media. Hampton Press, NJ.
Napoli, P.M. (2011). Exposure diversity reconsidered. Journal of Information
Policy, 1(1), 246–259.
Napoli, P.M. and Gillis, N. (2006). Reassessing the potential contribution of
communications research to communications policy: The case of media
ownership. Journal of Broadcasting & Electronic Media, 50(4), 671–691.
Napoli, P.M. and Karppinen, K. (2013). Translating diversity to Internet
governance. First Monday, 18(12). Available at: https://doi.org/10.5210/fm.
v18i12.4307.
Nicolescu, B. (1996). La transdisciplinarité. Manifeste, Editions du Rocher,
Monaco.
Pariser, E. (2011). The Filter Bubble: What the Internet is Hiding from You.
Penguin, London.
Overview xxiii
Paterson, C. (2007). International news on the internet: Why more is less. Ethical
Space: The International Journal of Communication Ethics, 4(1), 57–66.
Rebillard, F. (2012a). Modèles socioéconomiques du journalisme en ligne et
possibilités d’une information diversifiée. Les Enjeux de l’information et de la
communication, 12(3), 81–95.
Rebillard, F. (2012b). Internet et pluralisme de l’information. Réseaux, 176.
Redden, J. and Witschge, T. (2010). A new news order? Online news content
examined. In New Media, Old News, Fenton, N. (ed.). SAGE Publications,
London.
Rheingold, H. (2000). The Virtual Community: Homesteading on the Electronic
Frontier, Revised edition. MIT Press, Cambridge.
Stirling, A. (1998). On the economics and analysis of diversity. SPRU Electronic
Working Paper Series, No. 28 [Online]. Available at: http://www.uis.unesco.
org/culture/Documents/Stirling.pdf.
Stirling, A. (2007). A general framework for analysing diversity in science,
technology and society. Journal of the Royal Society Interface, 4(15), 707–719.
Sunstein, C.R. (2018). #REPUBLIC. In #Republic: Divided Democracy in the Age
of Social Media, NED-New edition, Sunstein, C. (ed.). Princeton University
Press.
Udehn, L. (2001). Methodological Individualism: Background, History, and
Meaning. Routledge, New York.
Van Cuilenburg, J. (2007). Media diversity, competition and concentration:
Concepts and theories. Media Between Culture and Commerce, 4, 25–54.
Von Hippel, E. (2006). Democratizing Innovation. MIT Press, Cambridge.
Vos, T.P. and Heinderyckx, F. (eds) (2015). Gatekeeping in Transition. Routledge,
New York.
Waldeck, R. (ed.) (2019). Methods and Interdisciplinarity, 1st edition. ISTE Ltd.,
London, and John Wiley & Sons, New York.
Wenger, E. (1999). Communities of Practice: Learning, Meaning, and Identity.
Cambridge University Press, Cambridge.
1
Online Platforms and Analysis

of Community Dynamics
1.1. Introduction
Online platforms, such as digital social networks, structure new ways of

accessing and circulating digital content. These actors master the function of
infomediation via connecting Internet users with any type of online content,
but also with other Internet users (Rebillard and Smyrnaios 2010; Rieder and
Smyrnaios 2012; Smyrnaios 2017; Heo and Park 2014). The result is the
implementation of a “digital” form of audience that marks the shift from an
individual consumer to a user inserted and guided by social networks
(“socially networked user”; Papacharissi 2010). These networks are
structured and evolve within the platform, giving rise to new collective
spaces where users come to occupy a hybrid position between consumers,
producers of information and content, and prescribers.
The analysis of these social networks is not original in itself. Since the
advent of the social web, online community approaches have occupied an
important place in socioeconomic research. The first reason is related to the
role of digital technologies (Rheingold 2000), which, thanks to their
communication functionalities, invite users to build new forms of dialogue
of the “many-to-many” type. The second reason is economic and strategic:
the repetition of interactions within social groups reduces the risks of
opportunistic behavior by economic actors, limits the uncertainties on the
behavior of others and contributes, as a result, “to the regulation of markets
Chapter written by Cécile BOTHOREL, Laurent BRISSON and Inna LYUBAREVA.

2 Diversity of Methodological Approaches in Social Sciences
and networks by structuring their organization and by making them more

efficient” (Benghozi 2006, author’s translation). Finally, on the consumer
side, experience and information sharing helps reduce learning and research
costs, and promotes consumer participation in product and service design
(DiMaggio and Louch 1998; Von Hippel 2005).
The originality of the approach proposed in this chapter lies in the idea
that these interconnected user groups are not always directly observable.
They can be formed at the scale of a platform beyond a specific theme, video
or channel. Indeed, very often, the socioeconomic analysis of online
communities focuses, in an upstream phase, on a definition of their
perimeters, for example, from the point of view of the interest or skills
shared by its members (Wenger 1999). As an example, we can cite the
abundant literature on free software communities, where the perimeter of the
group is defined within the framework of a project; blogs or wikis with their
information production communities, where the boundaries of the groups are
defined on the basis of contributions and uses of specific content; or even
forums wherein players share the same passions and interests.
Our work suggests that instead of imposing an a priori perimeter of

interaction, an important step in the analysis of social interconnections
within digital platforms is to identify the perimeter of these connections. In
other words, we hypothesize that connections between actors may emerge in
unexpected or hard-to-see parts of platforms and that these connections may
have a strong impact on the properties and outcomes of the functioning of
online groups and platforms. As a result, the dynamics of evolution and the
scale of the actions of these community forms may be as important in the
digital context as those of communities and groups previously identified in
social science literature.
This perspective highlights the relevance of social network analysis

(SNA) tools. Focusing on a structural approach to the relationships between
members of an organized social milieu, like a toolbox, SNA makes it
possible to visualize and model social relationships as nodes (social actors,
individuals, groups or organizations) and links (relationships between these
social actors) (Scott 1988; Borgatti et al. 2009; Mercanti-Guérin 2010).
SNA designates a set of methods, notions and concepts based on graph

theory to study relational phenomena. From a global point of view, the aim
is to characterize the whole network (or graph) by quantifying the number of
Online Platforms and Analysis of Community Dynamics 3
nodes (or vertices), relations (or edges in graph theory) or the diameter, that
is, the longest of shortest paths between each pair of nodes, which gives an
indication of the compactness of the graph. The average length of the
shortest paths provides an additional indicator of the overall structure of the
graph: the lower the average length, the more easily information can flow.
From a local point of view, the aim is to identify singular social actors
whose position in the network is remarkable. The measures operated on each
node can express either a local centrality (calculated with respect to the
neighboring vertices or links, such as the degree) or a global centrality
(calculated with respect to the whole graph). The centrality of intermediarity,
for example, consists of finding the set of the shortest paths between any pair
of nodes passing through a given vertex: the vertices which are most
frequently counted in this way are key intermediaries for the circulation of
information.
Finally, from an intermediate point of view, SNA allows for the

exploration of mesoscopic structures. The identification of densely
connected subgroups, called communities, or clusters, is for example
essential in the fight against an epidemic.
Therefore, in SNA, the relationships between individuals (the nodes),

whatever their nature (discussions, information flow or project funding), are
modeled by edges in a graph. In the different contexts of interaction, this
graph modeling allows the social network to emerge and to describe its
properties at the macro-, meso- or microscopic scale. In contrast to
socioeconomic analysis, SNA methods refer to algorithms from machine
learning and more generally from data science, which bring out groups of
individuals that have dense connectivity. Moreover, these methods are of
particular interest to the detection of social interactions that are not directly
observable and to understand the emergence and functioning of online
communities (e.g. Dupouët et al. 2003).
In this chapter, we will show how SNA enriches the analysis of online
platforms via the exploration of interactions taking place within these
interaction spaces. We use two examples: the crowdfunding platform Ulule
and the social media platform YouTube.
After a brief presentation of our field of study – the Ulule and YouTube
platforms – the next two parts focus, respectively, on each of these
platforms. Each part first presents the initial stage of the construction of the
social graph. Its structure is determined by the nature and forms of
interaction between users on each platform. This is crucial for the following
stages of community analysis, both in terms of the choice of analysis
methods and the interpretation of the results.
More specifically, in the case of Ulule we highlight that, based on the

underlying social graph, a multiplicity of algorithmic methods exist to
detect communities. Depending on the choice of an algorithm, the analysis
can lead to varied community forms. This choice subsequently necessitates
further in-depth study, which is often ignored by socioeconomic works,
instead using “ready-to-use” graph tools with certain pre-integrated methods
(such as the Gephi tool1).
In the case of YouTube, where the choice of the community detection

method is justified by the construction of the interaction graph (the graph
containing many connected cliques calling for a particular algorithm
appropriate for such a context), our analysis focuses on another technical and
conceptual lock: the analysis of community dynamics.
In conclusion, we summarize the interest of SNA methods for the

analysis of the different community forms characteristic of online platforms,
suggest some avenues of further research and discuss more generally the
importance of building a strong interface between social and computer
science disciplines in the analysis of digital phenomena.
1.2. Outline: Ulule and YouTube platforms
While very different in terms of their nature and objectives of user

interaction, the two examples presented in this chapter – Ulule and YouTube
– are among the leading platforms in their respective fields. As such, they
represent a particularly interesting field of exploration for us.
Since 2010, Ulule has become one of the leading European crowdfunding
sites with more than 2 million members, 24,000 funded projects and a
success rate of 63% (in 2018). The choice of the Ulule platform is relevant
for our study for two reasons. First, donation crowdfunding, of which Ulule
is a part, is highly developed in France. According to data from the watch
1 Gephi.org.
group Financement Participatif France (FPF), in 2018, among the 33,381

projects funded, 28,474 belonged to this category. Second, through its
interaction features as well as its community space, which allows all users to
exchange experience and advice, Ulule sets as a central objective to organize
Internet users into a true social network of Ululers with its own social capital
(Onnée and Renault 2014). Therefore, we propose to use SNA tools to
identify and describe original relational circles between the funders of
different projects. Importantly, these circles are not necessarily directly
visible to the platform’s managers, researchers, and users.
Like the “hidden” communities of the Ulule platform, the second

example focuses on the interactions of users on the YouTube platform.
Created in 2005, and quickly acquired by Google in 2006, the video hosting
platform allows its users to send, comment, watch and share audiovisual
content. Initially designed for amateur content, YouTube now hosts a wide
range of professional content, including media – news producers. Moreover,
as a support for online information, among the social network platforms,
YouTube comes second to Facebook in France in 2019 (ANR PIL project,
ANR-17-CE27-0010-01). For information producers, it is an important
strategic tool to gain audience. For this reason, the majority of traditional
media, pure players and partisan media now have their own YouTube
channels with teams of journalists dedicated to feeding them content.
YouTube news channels are at the center of our analysis. We suggest that the
interactions between users of this platform around the videos on the different
media channels may contain a latent part that materializes in the formation of
communities both at the inter-channel and inter-video level. This hypothesis is
justified by the structure of the platform’s recommendation algorithm. Indeed,
one of YouTube’s specificities is its algorithmic approach, which is likely to
reinforce the effects of interactions between users. Like some competitors,
YouTube has innovated on this point (Covington et al. 2016): YouTube’s
recommendation algorithm relies on users’ past behaviors to define common
recommendations for those who share a common activity history. It is thus an
algorithm based on “collaborative filtering”. In short, recommendations are
based on shared usage data, with those of User A being used to recommend
User B, and so on.
In both cases – Ulule and YouTube – we highlight that there are groups
of platform users where interactions are dense and intensify over time and
that these groups are not necessarily circumscribed to a theme, a project, a
video or a channel. In other words, in addition to the online communities
traditionally discussed in the literature, it is important to consider original

digital social forms whose framework is not known a priori.
Using two different platforms also allows us to detail step by step the
construction of the analysis of the social phenomena in question by
underlining the critical parts of the research and by proposing possible
solutions. The first step is to model the social graph, which is followed by
the detection and analysis of communities. This chapter highlights the
importance of a close articulation between the socioeconomic and SNA
approaches for the apprehension of the nature and structure of interaction,
specific to each platform, whether for the choice of community detection
algorithms or for the analysis of the trajectories of their evolutions.
1.3. Exploring the social network and communities on the Ulule

platform
1.3.1. Platform, data and construction of the social graph
The projects published on the Ulule platform fall into various thematic
categories, such as video, music, art, education, technology and so on.
Ninety days is the maximum period during which the collection can take
place, and donations can start from 5 euros. Donations can be made with or
without any kind of compensation. The platform accepts projects of bearers
with different status: individual, commercial organization or association.
The specificity of crowdfunding consists of three aspects. First, a social

link between participants in a group (e.g. a project) may exist outside of the
platform and, thus, may not be explicitly observable through online
interactions. Second, interaction between users of a platform does not
necessarily imply direct links between them, unlike blogs and discussion
forums. These links are mainly concretized by the action of financing
common projects. Finally, the perimeter of common interest and,
consequently, of the group is weakly defined: users may have a shared
interest in a thematic category of funded projects, or in a particular type of
project (e.g. projects created by associations), or in a participatory financing
activity in general.
In this context, the identification and analysis of these virtual social

networks requires appropriate tools. We believe that SNA tools, whose
effectiveness has already been discussed in the context of online interaction
analysis (Dupouët et al. 2003; Mercanti-Guérin 2010), are particularly

suitable. We will present below how, starting from raw data (which
contributor funds which project), we can model the community space on a
graph, and then show how the analysis of this graph, mobilizing SNA
concepts, via, among others, the detection of communities at a mesoscopic
level, will allow us to highlight different forms of community.
The data we analyze represents the first 5 years of Ulule’s operation,

from January 2010 through March 2016. After the cleanup, the dataset
includes 19,544 projects on the platform, of which 11,900 were successfully
funded and 7,644 failed. These projects gathered 876,758 contributors, who
contributed a total of 47.75 million euros. Since the Ulule platform does not
allow its users to highlight their friendship or interest links with other users,
we chose to use the only traces of interaction available to us: the
contributions of users to the same projects. We thus build a graph of
co-contributions that will allow us to verify if co-contributions are random
or, as we hypothesize, if there is a community dynamic.
Let there be a set of projects P uploaded by the platform and a set C of

platform users who have contributed to at least one project in P. We thus
define Pu the set of projects in which the contributor u ∈ C has participated.
We define an undirected co-contribution graph G = (V, E) in which each

undirected edge (u, v) signifies that users u ∈ C and v ∈ C have contributed
to at least the same three projects. The set of edges in the graph G is thus
defined by E = {(u, v) : |Pu ∩ Pv| ≥ 3}, and the set of nodes derived from it is
defined by V = {u | ∃ (u, .) ∈ E}. By choosing to create edges only between
two individuals who have co-contributed to at least the same three projects,
we eliminate incidental co-contributions to keep only those that would be
most likely to highlight an interaction between individuals. Such choice must
be supported by the expert’s knowledge, given that at the initial stage of
research, there may be no empirical evidence that particular threshold is the
most appropriate.
The resulting graph has many related components. That is, some parts of
the graph are not connected to each other by an edge. Some algorithms
cannot process several related components at the same time, so it is
necessary, during the analysis phase, to process them separately. However, it
is common that many related components are very small (and as such, of no
interest during the analysis phase), which is indeed the case for our data. We
therefore decided to focus on the largest related component of G which

contains 2,081 nodes and 4,749 edges (Figure 1.1).
The existence of this graph highlights the fact that there is a social
network specific to the Ulule platform and transverse to its various projects.
The density of the graph is low (0.002), with a majority of nodes having a
low clustering coefficient. The average clustering coefficient is indeed 0.26,
but a significant number (25%) of nodes are involved in cliques where all
their neighbors are themselves connected to each other, which means that
they have also co-contributed to at least three common projects. Concerning
the degree, we find a power-law distribution of degrees, classical in online
social networks (scale-free property of real complex networks). The average
degree is 4.56 and only 25% of the nodes have a degree higher than 4, the
maximum degree being 199. So, there are Ululers who co-fund projects with
many different contributors (24 of them have more than 50 neighbors in the
graph). The average length of the shortest paths is 3.97, with a diameter of
13 (the length of the longest shortest path). This sheds light on the small
world network, which, in addition to the scale-free property, refers to a
classical social network (Barabási 2003).
Figure 1.1. Graph showing the largest connected

component of co-contributions on Ulule
1.3.2. Algorithmic approaches for the detection of communities

on Ulule
By applying machine learning algorithms on the resulting graph, we can

show how Ululers, by grouping together, bring out different community
organizations. The study of online communities has been an important part
of socioeconomic research since the arrival of the social web. In this
research, the communities studied are often delimited and known a priori. In
network science, on the other hand, the notion of communities refers to a
machine learning method, dedicated to graphs, which makes groups of
individuals emerge on the basis of their connectivity.
Yet based on mathematical foundations of graph theory and encoded in

algorithms, the definition of a community is not consensual. There is no
universal definition or precise formula for the type of objects to be searched
(Fortunato and Hric 2016; Arifin et al. 2017). The definition of community
most frequently found in the network science literature is derived from the
mechanism of preferential attachment. It implies that a community is a group
of nodes (a subgraph) where there must be more edges connecting them than
edges connecting the community to the rest of the graph (Radicchi et al.
2004; Fortunato 2010). In this movement, a community is defined as a group
of vertices with a higher-than-average edge density (Newman 2006).
Faced with this lack of precise definition, in practice, communities are

defined by the algorithm that detects them, that is, they are the end products
of the mechanisms governing these algorithms (Fortunato 2010).
There are many algorithmic approaches to community detection. If they

are generally based on the notion of link density, their internal mechanisms
are often very different. We insist on the fact that the choice of the
algorithm, often ignored as a problem in socioeconomic works mobilizing
SNA tools, is far from being neutral: the resulting partitions, that is, the set
of communities, are indeed very different from one method to another.
In this work, we propose a methodology that helps select the community

detection algorithm. The idea is to characterize the communities resulting
from different methods via the integration of metrics describing their
topology, but also, if available, “business” metrics, for example,
socioeconomic, related to the analyzed problem. In the case of the Ulule
platform, the introduction of these additional metrics for the production of a
community typology allowed us to solve, at least partially, the problem of

choosing a specific algorithm.
To illustrate our points, we have chosen three well-known methods that

are representative of SNA: Edge betweenness, Louvain and Walktrap.
These methods are very different from one other. Edge betweenness
(Girvan and Newman 2002) identifies inter-community links by selecting the
links with high betweenness centrality, that is, links that are most frequently
found on the shortest paths between each pair of nodes in the graph
(Freeman 1977).
This method starts by computing the betweenness centrality of all links

and orders them in descending order. Then, in an iterative way, it removes
the most intermediate link, thus progressively disconnecting the graph until
it has disconnected nodes (i.e. as many related components as nodes). In the
vein of hierarchical top-down classification algorithms, the method produces
a dendrogram. In most uses of this method, a partition is obtained by cutting
this dendrogram at the level that maximizes the modularity score Q
(Newman 2004). This modularity measure is often used to evaluate a
partitioning. It compares the fraction of intra-community edges with the
fraction of inter-community edges. Variations of this method do not use
modularity and seek to minimize conductance (another quality metric of a
partition, non-global as it is relative to a community, which measures the
fraction of outgoing edges over the total number of edges linked to nodes in
the community) (Leskovec et al. 2010).
While the Edge betweenness method is based on the detection of central

edges to divide a network into communities, the Louvain method will act in
a bottom-up manner, by agglomerating nodes during successive iterations
and seeking, again, to optimize modularity (Blondel et al. 2008). This is a
hierarchical bottom-up clustering method that maximizes modularity locally
by modifying the composition of the communities little by little. The
algorithm acts in two steps: first, by moving each node to neighboring
communities in order to maximize its contribution to increasing modularity;
then, in a second step, a meta graph is created producing a community graph.
These two steps are repeated as long as they generate an improvement in
modularity.
The Walktrap method (Pons and Latapy 2005) is based on a completely

different principle. In this stochastic approach, if two nodes are in the same
community, the probability that a random walker will go from one to the
other in only a few moves is very high (the notion of a trap) and
consequently the distance is low. Once the distances between nodes have
been computed using this principle, a hierarchical bottom-up approach of
successive agglomerations will merge (as in Louvain) the nodes and then the
communities, this time minimizing the increase in Euclidean distance
(caused by these mergers).
(a) Louvain (b) Edge betweenness
(c) Walktrap
Figure 1.2. Segmentation into communities according to the three tested algorithms.
Only the eight largest communities of each partition are colored, the others remain
gray. For a color version of this figure, see www.iste.co.uk/lyubareva/diversity.zip
These three well-known methods produce quite different communities

(see Figure 1.2) and each has its own intrinsic advantages and problems
(Dao et al. 2020). For example, the Louvain method generally generates a
few fairly large communities, which may be problematic for some studies,
since the method fails to highlight small communities, which may also
provide important information about the overall organization of a social
network (Fortunato and Barthelemy 2007). The other two methods reveal
finer partitions, with smaller groups (Edge betweenness), or even very small
ones (Walktrap).
As expected, Table 1.1 shows that the Louvain algorithm partitions the
Ulule network into fewer communities than the other two methods.
However, their large size in terms of member numbers has relatively little
influence on the degree, which averages 3.63 for communities that gather 90
members on average versus 2.29 for 12 members on average for Walktrap.
The average proximity centrality and the average density remain stable
regardless of community size or the algorithm used, contrary to the
clustering coefficient which slightly varies: some clusters obtained by
Walktrap are indeed very small, probably with nodes at the periphery of the
graph which do not form cliques.
Number of Clustering
Members Degree Intermediarity Proximity
communities coefficient
Louvain 22 90.48 3.63 0.26 2,871.45 0.24
Edge Bet. 71 28.90 2.96 0.20 2,234.84 0.23
Walktrap 167 12.39 2.29 0.15 1,695.03 0.22
Table 1.1. Communities generated by three state-of-the-art methods. The indicators

characterizing the nodes (last four columns) are computed for each member of each
community, then an average is computed for each community. The summary shows
here the averages of the indicators for all communities by method
1.3.3. Typology of community forms
Based on the three community detection methods and the resulting

partitions, we seek to develop a typology of Ulule communities. In order to
do so, we propose to add two types of metrics – relational, derived from the
graph, and socioeconomic, derived from data on the behavior and activity of
the platform’s users (Tables 1.2 and 1.3) – in order to identify the profiles
of community members. This manipulation allows us to describe the
communities by vectors of presence rates of different contributor profiles as
a criterion for comparing the partition results of three community detection
methods.
Definition
Degree: number of individuals having co-contributed to at least three
Degree
same projects.
Local clustering coefficient: measures how connected a node’s
Clustering neighborhood is. The higher the coefficient, the more the
coefficient neighborhood tends to be a clique, all the neighbors
are themselves connected to each other.
Centrality of intermediarity: number of times the node is on the
Betweenness
shortest path between two nodes of the graph.
Proximity centrality: defines how central a node is (i.e. it has the least
Closeness
distance to all other nodes).
Table 1.2. Centrality measures of contributors in the social graph
Socioeconomic attributes of a contributor

Objective Average in euros of the objective of the funded projects.
Contribution Average amount of contributions.
Funded projects Number of projects funded.
Thematic specialization rate: ratio of projects in the most funded
Specialization
theme to the total number of funded projects.
Contribution time Median status of projects at the time of contribution.
Proportion of funded projects for which the median progress of
Early contributions
funded projects is lower than that of its neighbors.
Homophily attributes
Neighbor Average of the thematic specialization rate of all the neighbors of
specialization rate the contributor.
Neighbors’
Average median progress of projects throughout neighbor
contribution
contributions.
moment
Thematic similarity Rank correlation in the ranking of thematic categories funded by the
with neighbors contributor and those funded by its neighborhood.
Table 1.3. Socioeconomic and behavioral

attributes of a contributor in the social graph
The reader may notice that in addition to the metrics usually mobilized in
the literature to describe crowdfunding contributors, we create additional
variables related to (i) a level of thematic specialization of a contributor and

its neighbors in the graph; and (ii) the moment of arrival to the funded
projects of the contributor and its neighbors in the graph. This information
was introduced in keeping with the literature on crowdfunding and allows us
to analyze the social and behavioral proximity between contributors and to
model the homophily principle.
Hierarchical clustering, combined with principal component analysis and

applied to the set of relational and socioeconomic metrics, revealed five groups
(clusters) of Ululers within the platform’s social network, corresponding to
different profiles of users: sponsors, followers, forerunners, collaborative
specialists and specialists.
The first cluster (18 individuals) gathers the contributors who have
central positions in the graph, that is, they make the link between the
different contributors of the platform and facilitate the diffusion of
information. They contribute on average to 140 different projects and are
connected, via these projects, to more than 100 different contributors (the
average degree in this cluster is 105.4). These funders are not necessarily
interested in a specific category of projects and their neighbors are weakly
connected to one another (the average clustering coefficient in this cluster is
0.05 against 0.26 on average in the graph).
We propose to refer to them as sponsors since their number is very small

(only 18 individuals) and their support is substantial for several Ulule
projects of various natures.
The second cluster (653 individuals), which we propose to call followers,

is characterized by a long delay in the arrival to the projects, on average after
60% of the time elapsed since the opening of the campaign. This type of
contributor is also characterized by a number of projects funded close to the
average. They chose to contribute to very big projects in terms of funding
goal (average goal is over 17,000 euros).
The third cluster (538 individuals), which we propose to call Forerunners,

is characterized by an early arrival in projects. Notably these contributors
come to the projects long before all their neighbors.
The last two clusters are distinguished by a very high rate of

specialization of the funders themselves, and their neighbors, hence their
name: specialists. These two profiles focus on certain thematic categories of
the platform’s projects. The specialists of cluster 4 (368 individuals) have a
very high clustering coefficient (0.9 vs. 0.26 on average in the graph),
indicating a strong cohesion of links between neighbors and their solidarity
in the choice of projects to support. An important characteristic of this
cluster 4 is the amount of the contribution per funder, which is higher than
the average in the graph (47.6 euros vs. 43 euros in the graph). This result
highlights the link between the volume of contributions and social
involvement. On the other hand, the specialists of cluster 5 (504 individuals),
less connected to each other, are characterized by a strong proximity of
interest with their neighbors, who specialize on exactly the same themes
(Kendall’s tau = 0.4 in cluster 5 vs. 0.14 in the graph, p-value < 0.05). Their
specialization rate is relatively stronger (0.8 vs. 0.7 in cluster 4 and 0.58 for
the graph on average). This corresponds to an even more restricted choice of
funding themes than in cluster 4. We thus propose to distinguish the
collaborative specialists of cluster 4 from the specialists of cluster 5.
Collaborative specialists and specialists do not necessarily occupy the most
central positions in the graph. They do not contribute to a very large number
of projects and are not attracted by the size of the projects.
In order to identify families of communities and to produce a typology of

their forms, we describe the communities by vectors of presence rates of
different contributor profiles. To generate different families of communities,
for each partition (i.e. from either the Louvain, Walktrap or Edge
betweenness algorithm), we apply a clustering method (unsupervised
learning). From a methodological point of view, without any preconceived
ideas on the types of communities we wanted to obtain, nor on the number of
clusters, we chose to apply and compare four classic methods: PCA-AHC,
a dimension reduction by principal component analysis (PCA; two
dimensions) followed by an agglomerative hierarchical clustering
(Euclidean distance, Ward’s method of variance minimization); FA-AHC, a
dimension reduction by factor analysis (two dimensions, more robust to
noise than PCA) also followed by an agglomerative hierarchical clustering
(Euclidean distance, Ward’s method of variance minimization); the K-means
method; and finally, the decision tree that shows discriminating variables.
Applied to the partition from the Edge Betweenness method, the decision
tree produces families almost identical to the K-means method
(completeness = 0.964 and adjusted Rand index [ARI] = 0.854). The
proximity with the clusters produced by the PCA-ACH (or FA-ACH) is less
obvious if we consider these measures (completeness = 0.513 and ARI =
0.398) but the characteristics of the families found remain close. Only the
number of communities in each cluster changes a little. Indeed, each of these
techniques, whatever the partition, allows us to identify three families of
communities, which can be summarized as follows:
– Family 1 (sponsors, followers, forerunners): balanced communities
composed of all profiles, including sponsors, with a strong domination of
followers. These communities are very large.
– Family 2 (specialists): communities very clearly dominated by
specialists; very few sponsors, forerunners and followers. These
communities vary greatly in size.
– Family 3 (followers, forerunners): communities that are dominated by
forerunners and to a lesser extent followers; there are no sponsors and the
specialists are very rare. The communities in this family have a much
smaller disparity in size, concentrating in small or even micro-communities.
Concerning the choice of the community detection method, we find an

instance where, in spite of the disparities of the partitions, they seem
equivalent from the point of view of our analysis. Indeed, they allow us,
beyond the differences linked to the algorithms, to highlight the families of
communities, no matter the choice of the underlying algorithm. The
extension of this analysis can include, for example, a comparison of the
different families of communities from the point of view of their economic
efficiency, where efficiency in the context of crowdfunding may refer to the
success of the fundraising campaigns of the projects supported by the
detected communities. Our explorations in this area (Lyubareva et al. 2020;
Bothorel et al. 2021) show that the most efficient communities, which carry
the maximum number of funded projects, come from families 2 and 3. They
are small, with 3–13 members. These are quite specialized and on average
fund projects across four different themes. The projects are very varied in
terms of funding goals (all types of amounts are represented), but less than
25% of the projects have a larger than average scope, so we have rather
modest projects. An interesting observation is that the least successful
communities also belong to families 2 and 3. A more detailed analysis

reveals that what clearly differentiates the most successful communities is
the clustering coefficient and the number of comments, which are
significantly higher. In other words, the thematic specialization and the
founding principle of homophily that characterize family 2 and partially
family 3 do not in themselves guarantee the success of crowdfunding
projects. These must be combined with a strong social involvement and
cohesion among its membership in order to give rise to a significant
economic performance of fundraising campaigns.
For that reason, contrary to some existing work (McPherson et al. 2001;
Bisgin et al. 2010), these results highlight that the principle of homophily
and thematic proximity is therefore not the only determinant of online
community formation and success. Thematic diversity can also be a driver
for the development of online comunities. The role of diversity and network
externalities in online communities has also been demonstrated in Wang and
Kraut (2012). On the other hand, thematic specialization, which is at the
origin of the homophily principle in online communities, can lead to a
particularly high level of cohesion and solidarity within communities and
guarantee, in this case, the striking performance of supported projects. Our
results also highlight that in certain themes, such as games, comics, video,
publishing or charities, these specialized and supportive communities are
more likely to develop. This result is in line with existing work on the
particularly important role of communities in the production and/or
consumption of goods in these cultural sectors (Throsby 2001; Cohendet
et al. 2008; Pélissier and Chaudy 2009; Auray and Georges 2012).
To deepen the discussion on the problem of choosing a community

detection method, we invite the reader to discover the long version of this
work describing our methodology (see Bothorel et al. 2021). We discuss in
detail the choice of the community detection method, which is not always so
simple. In particular, we deepen the analysis of the partitions from the point
of view of the organizational forms they allow to highlight. More precisely,
beyond the profiles of the members composing the communities, it can be
interesting to focus on notions such as cohesion (clustering coefficient,
density of links), centralization (through the presence of dominant nodes
which alone ensure this cohesion), internal versus external connectivity, etc.,
and even the combination of these indicators through bivariate maps (Dao
et al. 2021).
1.4. Exploring the social network and communities on the

YouTube platform
1.4.1. YouTube: data and construction of interaction graph
The second example of platform analyzed in this chapter concerns the

social interactions on media channels within YouTube and the formation of
new public spaces of interaction. We will limit our analysis to the YouTube
channels of French-speaking professional media. In the field of media, the
distribution of journalistic information on social media, and on YouTube in
particular, offers various means of action, in the practices of editorialization
(in the videos, their themes, the diversity of positions represented), of
personalization, allowing for offers to be declined in “increasingly
specialized proposals” (Charon 2015, p. 88, author’s translation) and in the
framing of their reception (the debates and remarks among commentators).
Moreover, it makes possible the unprecedented encounter and crossover
between the readership and communities of different titles and media. This
phenomenon gives social media, such as YouTube, the ability to take on the
attributes of the public sphere (Habermas 2006).
Interactions on YouTube mainly take the form of users’ reactions to

different videos posted on channels. In these collective spaces, users come to
occupy a hybrid position between consumers, information producers and
opinion prescribers. As explained below, unlike the case of Ulule, the
problem of choosing a community detection method is not central to the
analysis of interaction logics around online information on this platform.
However, the question of detecting and analyzing the evolution of
interactions within these user groups, which are dynamic in nature and
whose perimeter is not known a priori at the scale of the platform, is of great
importance.
The analysis we will present in this section covers comments published

on 59 French-language media channels from June 2006 to June 2019. It
includes, after data cleaning, 2,209,222 comments (divided into 1,024,363
threads, see hereafter) published by 517,689 users on nearly 46,000 videos.
We thus have, for each video of each of the 59 channels, the list of
comments that have been published.
Exchanges between consumer-commenters can take two forms: messages

posted after a video, which are left unanswered by other commenters, and
comments which respond to other participants’ comments. The first
comment, followed by its replies, forms a thread of discussion. To make sure
that our analysis captures the real spaces of interaction, we focus only on the
discussion threads, leaving out of the analysis the succession of comments
without answers. This choice constitutes another original element of our
work as compared to research that is often based on the analysis of a “flat
social world” (Boullier 1988).
Since the studies we want to conduct are longitudinal, we have chosen to

structure our data in an ordered sequence of static graphs, one per
observation period. In terms of analysis methodology, this allows us to use
SNA methods based on static graphs, which are numerous and well
documented. In terms of interpretation, this allows us to segment and thus
simplify the analysis for a chosen observation period. In line with previous
results (Islam et al. 2013), we chose a one-week period.
We therefore create a set of consecutive observation periods that gather

all the events of the period. An event is here the publication of a message by
a user about a video. It is therefore impossible for an event to belong to
several observation periods.
Each time period t is defined by a start date ts and an end date te. The
events that we will analyze are thus divided into T consecutive periods (i.e. T
weeks in our case). We then construct, for each period, a graph that gathers
the interactions between individuals that took place during this period.
Let t be a period of time, where Gt = (Vt, Et) is the graph of interactions

during this period of time.
Vt is the set of nodes in the graph Gt. Each node represents an individual
who has published at least one message during the time period t.
Et is the set of edges of the graph Gt. An edge e = (v1, v2, w) links two
nodes v1 and v2, if both individuals associated with it have published at least
once in the same thread during the same time period t. w is the number of
threads in which v1 and v2 have a co-publication over the same period of
time.
Another random document with
no related content on Scribd:
fate of ardent, generous souls like hers, if sometimes she was
betrayed into the many nets which greed, jealousy and base cunning
are always at hand to spread, for rendering nobler natures wretched.
Mademoiselle de Montpensier was, in one word, a true descendant
of her grandfather, Henri IV.
Lauzun, exiled as he had been, from Versailles, soon after passed
over to England, where he contrived to make himself useful by
conducting the queen and infant prince of James II. safely to France,
during the revolution of ’88. Louis, who received the dethroned
English king with great demonstration of sympathy and
magnificence, and gave the exiles his palace of St Germains for their
home, was thus again brought into direct communication with
Lauzun, who, being readmitted to royal favour, was created a duke;
but he never really regained the confidence of Louis.
On the occasion of the death of Mademoiselle, he presented
himself at the palace, attired in a magnificent mourning cloak. This
so angered Louis, that Lauzun ran a parlous risk of once more taking
the road to Pignerol.
All that remained of la Grande Mademoiselle’s possessions was
now proposed to be given to the illegitimate and legitimatized
children of the king; but precisely how to deal with Lauzun and his
wealth, acquired from Mademoiselle de Montpensier, was not so
apparent, since the question still remained open, whether
Mademoiselle had been his lawful wife. No one knew for certain, and
Madame de Maintenon conceived the ingenious idea of trying to
worm the true state of the case from Ninon, whom she knew had
been summoned to Mademoiselle’s dying bed, feeling persuaded
that Mademoiselle de L’Enclos was acquainted with it. She
accordingly begged her, in a little note very affectionately worded, to
come to Versailles.
Ninon was greatly tempted to reply that if Françoise desired to
speak to her, she might be at the trouble of coming to the rue des
Tournelles. All circumstances taken into account, and the generosity
with which she had treated Françoise’s little ways, it did not appear
to her that she was bound to wait upon the woman, merely because
she had lighted upon the lucky number in life’s lottery. Ninon,
however, was but a daughter of Eve. Curiosity was strong to see
how Madame Louis Quatorze lived in the lordly pleasure-house, and
forthwith she obeyed the summons.
Queen Maria Théresa’s surroundings and retinue had been
modest enough even to parsimony. Madame Louis Quatorze was
attended by a numerous guard, a train of pages, Swiss door-
keepers, and the rest; while her Court and receptions were as
magnificent as those of the king. Madame took herself very
seriously, and her deportment had become most majestic. To Ninon,
however, she unbent, and was simply the Françoise of old times.
She led her into her own richly furnished private boudoir, adorned
with a curious conglomerate of pictures and statuary, Christian and
pagan, where an enormous, life-sized figure of Christ, in carved
ivory, was neighboured by painted Jupiters and other Olympian
deities, in curiously heterogeneous fashion. There Françoise
embraced Ninon with quite a prodigality of affection. Suddenly,
however, her manner changed; she congealed into gravity and tones
of great solemnity, and Ninon saw the tapestry folds along the wall
quiver slightly. It occurred to her that one only, His Majesty Louis
XIV., could have any possible right to be present in that most private
apartment, and even then she felt the need of putting a strong
restraint upon herself and her foot, to prevent it from bestowing a
kick upon the tapestry. Then the truth began to come out, the
lamentable truth that Madame and the king were greatly perplexed
as to the best mode of dealing with the Duc de Lauzan, whose
possessions, made over to him by the Grande Mademoiselle, those,
that is to say, which he still held, were much wanted for the king’s
children. He had so many, as Madame de Maintenon pointed out.
That, admitted Ninon, was true enough, “but I will engage, you will
not be increasing the number,” she added. “What is the point of the
question?” It was whether Mademoiselle had really married Monsieur
de Lauzun.
The full significance of it all now dawned upon Ninon. Had
Mademoiselle not been his wife, it would be a comparatively simple
matter to compel a revocation of the gifts which the princess had
made him in the course of her life, in order that these should enrich
the children of de Montespan. No consideration was yielded to the
fact that, be Lauzun what he might, the gifts had been tokens of
Mademoiselle’s affection for him. Ninon preferred complete inability
to afford any trustworthy sort of information on this head, and
suggested applying for it to Madame de Fiesque, who might be
better instructed: “but,” continued Ninon, “supposing Mademoiselle
was not his wife, surely to publish the fact, would create a scandal
which His Majesty would consider paying too dear a price for the
estates of Auvergne and St Fargeau. Either she was Lauzun’s
wedded wife or—”
Here the chronicle goes on to relate: Mademoiselle de L’Enclos’
words were interrupted by a tremendous disturbance at the door,
occasioned by an altercation with the guards, of some person
endeavouring to force his way in. The voice was d’Aubigné’s, and
the next instant he reeled in, far gone in a state of intoxication, and
staggering to his sister, he gripped her by the arm and thrust her
back into the chair from which she had risen.
This chronicle goes on to relate a terrible scene, over which, for
the honour of human nature, some kind of veil may be allowed to
hang, lest veracious history has been embroidered by the ample
material fact has afforded. The family differences of private domestic
relations are frequently unedifying; but when it comes to the base
humiliating of a great monarch, one in whose very vices and
mistakes grace and virtue had been apparent, until the widow
Scarron crossed his path, pen may well refrain from detail, and
explain only that the intruder, d’Aubigné, had burst in upon his sister,
to reproach her for her treachery in the matter of inducing him to
enter St Sulpice. Taking advantage of the absence of his mentor and
alter ego, Santeuil, she had contrived to trap him by false promises
and misrepresentation into the hated place. His liberty for one thing,
and of all things prized by d’Aubigné, would not, she had said, be
curtailed; it had, however, been so entirely denied him, that when he
had attempted to leave, he had been unceremoniously “clapped,” as
he phrased it, “into a cellar,” and he had only escaped by wriggling
through an air-grating. To any one possessed of the faintest sense of
humour, the notion of making a monk of any sort of this wild harum-
scarum would have seemed too preposterous; but the sense, always
so lacking in Françoise d’Aubigné, allowed her to indulge in only too
many absurdities whose ending was disastrous; and in any case, the
notion of removing the incommoding one from the taverns and cafés
and other public resorts where he freely gave utterance to his
estimate of Madame Louis Quatorze, and notably of her newly
acquired saintliness, was dominant in her, and to be achieved at any
cost. She earnestly desired his conversion, possibly if only to silence
the hideous music of the ditty, whose refrain he was for ever
chanting in the streets, echoed by so many ribald tongues—
“Tu n’as que les restes,
Toi!
Tu n’as que nos restes!”
Since the chronicle goes on to tell that Louis the king was
concealed behind the tapestry during the interview of Madame and
her old friend Ninon, the appearance of d’Aubigné, with his string of
furious reproach, was of course singularly inopportune; and at last
the king, unable any longer to restrain his wrath, dashed aside the
concealing Gobelins, and white with anger, and his eyes blazing with
indignation, ordered the culprit’s arrest by the guards, and carrying
off to the Bastille. Confounded by the unexpected apparition,
d’Aubigné’s sober sense returned, and he promised everything
required of him with the humblest contrition, adding that if he might
suggest the homely proverb in that august presence, there was
nothing like washing one’s soiled linen at home.
The king’s silence yielded consent, and d’Aubigné was permitted
to depart from his brother-in-law’s presence a free man, on condition
of making St Sulpice his headquarters. It was at least preferable to a
lodging in one of the Bastille towers, he said, but any restraint or
treachery on the part of Françoise, or of Louis, in the way of his
coming and going into what he called that black-beetle trap of St
Sulpice, would be at once signalised. And thus the difficulty was
adjusted, a compromise being effected by appointing a certain Abbé
Madot to shadow the ways of d’Aubigné when he took his walks
abroad.
But for Ninon the malice of her old friend took on virulence, and it
was found later that Françoise charged her with having planned the
scandalous scene, in so far as bringing d’Aubigné into it; that she
had connived at his coming just at that moment. Yet exactly, except
for the king’s concealed presence, what overwhelming harm would
have ensued, is not apparent, and certainly for that situation, Ninon
could not have been responsible. Henceforth all shadow of
friendship between the two women died out, and enmity and
bitterness were to supervene when opportunity should be ripe.
CHAPTER XXIV
The Falling of the Leaves—Gallican Rights—“The Eagle of Meaux”—Condé’s

Funeral Oration—The Abbé Gedouin’s Theory—A Bag of Bones—Marriage
and Sugar-plums—The Valour of Monsieur du Maine—The King’s
Repentance—The next Campaign—La Fontaine and Madame de Sablière—
MM. de Port Royal—The Fate of Madame Guyon—“Mademoiselle Balbien.”
And time passed on—passed on. The brilliant century was in its sere
and yellow leaf, and one of the best and most amiable of the glorious
band, le Nôtre, the gardener par excellence, faded and died, to the
great grief of Louis, who dearly loved his company, and would walk
by his chair in the garden of Versailles, when the invalid’s limbs had
failed him. Ninon keenly felt the loss of the kindly friend, who had
been one of the party to Rome with Santeuil—who had nearly
missed the papal benediction on his hymns, as he always believed,
by his witticisms about the carp. And now the good canon was to die,
victim of a practical joke on the part of the young Duc de Condé, who
amused himself with emptying the contents of his snuff-box into his
guest’s glass of champagne. Unawares, Santeuil drained the glass;
and the hideous concoction produced a fit of such convulsive
sickness, that he died of it. Bitterly enough Condé repented, but that
did not bring back his friend.
About the time that the zenith of Louis’s power was attained, when
his very name was uttered on the bated breath of admiration, hatred
and terror—and the yoke of the widow Scarron had not yet
entangled him—and while the Doge of Genoa was compelled by
Duquesne to sue for mercy at the feet of the French monarch—
accused of complicity with the pirates of the Mediterranean—the
Court of Rome was compelled to yield to the demands of the Church
in France, in the matter of the régale. This right, which had ever
been the strength and mainstay of religious Catholic independence
in France, had fallen in later days somewhat into abeyance; and
when, some nine years earlier, it had been put into active force
again, the pope opposed it. To establish it on a firm footing was the
work of Bossuet, who set forth and substantiated with the bishops of
the dioceses of France the existing constitution of the Gallican
Church under the ruling of the four famous articles: 1. That
ecclesiastical power had no hold upon the temporal government of
princes. 2. That a General Council was superior to the pope. 3. That
the canons could regulate apostolical power and general
ecclesiastical usage. 4. That the judgment of the Sovereign Pontiff is
only infallible after the universal and general consent of the Church.
The pope and the Court of Rome had no choice but finally to
accept these propositions; but unpalatable as they were, they came
between the worse evil threatening Catholic Unity, of a schism such
as it had suffered in England under Elizabeth and Henry.
The splendid gifts of Bossuet place his memory on a lasting and
lofty eminence, as it placed him, living, in distinguished positions,
Bishop of Meaux, preacher at the Louvre, preceptor to the Dauphin.
From his profound theological learning welled forth the splendid
eloquence which thrilled the vast assemblages flocking to drink in his
orations. One of the most magnificent among these was that at the
obsequies of the great Condé, beginning—
“Cast your gaze around; see all that magnificence and piety has
endeavoured to do, to render honour to the hero: titles, inscriptions, vain
records of what no longer exists, the weeping figures around the tomb
and fragile images of a grief which Time, with all the rest, will bear away
with it, columns which appear to lift to high heaven their magnificent
testimony to him who is gone; and nothing is lacking in all this homage
but him to whom it is given.... For me, if it is permitted to join with the
rest in rendering the last duties beside your tomb, O Prince! noble and
worthy subject of our praise and of our regrets, you will live eternally in
my memory. I shall see you always, not in the pride of victory ... but as
you were in those last hours under God’s hand, when His glory was
breaking on you. It is thus I shall see you yet more greatly triumphing
than at Fribourg and at Rocroi.... And in the words of the best-beloved
disciple, I shall give thanks and say—‘The true victory is that which
overcometh the world—even our faith.’”
A noble purity of spirit and deep conviction inspired Bossuet’s

eloquence. His knowledge was limited by his Jesuit training, though
he studied anatomy at a later period, by the king’s desire, in order to
instruct the Dauphin in the science; but with science generally and
physics he was unacquainted. As a Jesuit he was opposed to
Jansenism and the Port-Royalists; but for long the gentle piety of
Fénelon retained the respect and admiration of Bossuet’s more fiery
spirit. Both these great men gave instruction at St Cyr, by the desire
of Madame de Maintenon and the king.
Time must indeed have passed lightly by Ninon; for once again, at
the age of eighty years, she inspired a young abbé, named Gedouin
—a distant relative on the maternal side—with deep fervent
admiration. Ninon at first believed that he was jesting with her, and
rebuked him severely; but it was a very serious matter on his part,
and though she told him of her fourscore years, he declared that it in
no way altered his sentiments. “What of that?” he said; “wit and
beauty know nothing of age,” and the Abbé Gedouin’s pleading,
which was not in vain, terminated Ninon’s last liaison with an
affectionate and endearing friendship. When he was rallied on his
conquest, the abbé’s rejoinder was that—
“Ah, mes amis, lorsqu’une tonne
A contenu d’excellent vin,
Elle garde un parfum divin
Et la lie en est toujours bonne.”
Monsieur de Lauzun, on the other hand, being now over sixty

years old, contracted a marriage with an English girl of sixteen. She
was so fearfully thin, that the Duc de St Simon, who was one of
Mademoiselle de L’Enclos’ cercle, said de Lauzun might as well
have wedded all the bones of the Holy Innocents Cemetery, where
the skulls and bones were piled in pyramids.
St Simon was a delightful conversationalist. He was the son of the
old favourite of Louis XIII. He could be very caustic with his
anecdotes. One night he greatly amused the company with an
account of the marriage of the son of the Grand Dauphin, the little
Duke of Burgundy. He was of the tender age when ordinary and
everyday little boys are occasionally still liable to chastisement by
their elders. The duchess to be, who was still very fond of her doll,
was presented on the occasion by the Queen of England with a very
elegantly trimmed shift, handed to her by the maids of honour on a
magnificently enamelled tray. In this garment she was attired, while
her youthful husband, seated on a footstool, was undressed in the
presence of the king and of all the Court. The bride, being put to bed,
the Duc of Burgundy was conducted in and also put into bed, beside
which the Grand Dauphin then took his seat, while Madame de Lude
took her place beside the young duchess. Then sugar-plums were
offered to the bride and bridegroom, who cracked them up with the
greatest enjoyment. After about a quarter of an hour, the Duc was
taken out of bed again, a proceeding which appeared greatly to
displease him, and he was led, sulking enough, back to the
antechamber, where the Duc de Berry, some two years his junior,
clapping him on the shoulder, told him he was not a bit of a man. “If it
had been me,” he added, “I should have refused to get out of bed.”
The king imposed silence on the little rascal’s rebellious counsel,
and placed the bridegroom back into the hands of his tutors,
declaring that he would not permit him to so much as kiss the tips of
his wife’s fingers, for the next five years to come. “Then, grandpapa,”
demanded the little brother, “why have you let them be married? It is
ridiculous.” It was all certainly something like it.
After that the child was placed for his instruction in the care of the
Abbé de Fénelon, whose rapid advancement at Court had been
attained by his lofty character and talents.
But Louis had far more affection for his illegitimate children than
for these, and aided by Madame de Maintenon’s intrigues, he finally
succeeded in securing a large portion of the heritage of la Grande
Mademoiselle for the Duc du Maine and the Duc du Vendôme; but
the brave spirit of heroes and conquerors he could not endow them
with, for all his desire. It was to no effect that he confided command
to them of his troops in Holland. The Duc du Maine specially
undistinguished himself. Just as the enemy was escaping scot-free,
he found he was hungry, and asked for a cup of bouillon to
strengthen him. “Charge! Charge, Monseigneur!” urged Villeroy’s
messenger, coming to him in a fever of excitement.
“Oh, well, patience,” replied the warrior; “my wing is not in order
yet.”
Finding no sort of response to his repeated messages, Villeroy
went in search himself of the prince, and found him in his tent, at his
confessor’s knees. The first duty of a good Christian, he said, was to
make his peace at such times with Heaven. So the religious
discipline of his governess and stepmother, the widow Scarron and
Madame de Maintenon, had borne fruit. It was of a different flavour
from the prayer of the brave servant of King Charles I.—Sir Edmund
Verney—before Edgehill: “Lord, Thou knowest how busy I must be
this day. If I forget Thee, do not Thou forget me.” And there was no
battle won or lost that day on the Dutch frontier, and Louis, when
they brought to Versailles news of the enemy’s safe retreat, was at a
loss to understand the situation; for no one cared, or dared, to tell
him the truth, until Lavienne, his valet-de-chambre in chief, in the
days of Louis’s amours, hazarded the observation that, after all,
proverbs could speak falsely, and that “Good blood could lie;” and
then he went on to add the other truths concerning Monsieur du
Maine. In the face of the fulsome praise following in the journals—
which lied as only journals know how—the king was overwhelmed
with grief and chagrin; and, beside himself, he broke his cane in a fit
of anger on the back of one of an unlucky servant, whom he
happened to detect surreptitiously eating a bit of marchpane. This
ebullition, creating the consternation of all the Court, just sitting down
to dinner, brought Madame Louis Quatorze and Père la Chaise upon
the scene. “Parbleu, mon père,” said the king, gradually regaining his
senses, “I have just chastised a wretched creature who greatly
merited it.”
“Ah!” gasped the confessor.
“And I have broken my cane on his back. Have I offended God?”
“No, my son, no,” replied the holy man. “It is merely that the
excitement may be harmful to your precious health.”
Fortunately the cane, being of slenderest rosewood, had easily
snapped.
Before the end of the next campaign, the redoubtable Duc du
Maine was recalled: d’Elbœuf hastened to say to him, making a
profound bow, “Have the goodness, Monseigneur, to inform me
where you propose entering on the next campaign.”
The duke turned, smiling, and extended his hand to d’Elbœuf,
whose ironical tones he had failed to perceive.
“Wherever it is,” added d’Elbœuf, “I should wish to be there.”
“Why?” demanded the duke.
“Because,” replied d’Elbœuf, after a silence, “at least one’s life
would be safe.”
Monsieur du Maine gave a jump, as if he had trodden on a
serpent, and went away without replying, not being better furnished
with wit than he was with valour.
And the autumn leaves of Ninon’s life were ever fast falling around
her. In her Château de Boulogne Madame de la Sablière passed
away, and la Fontaine, finding life a sad thing without her, quickly
followed her.
The Jesuit conception of religious faith, great as were its merits as
originated in the mind of Loyola, theoretically, and in its code drawn
up by his gifted successor, Lainez, had displayed its imperfections in
its practical working, as time passed. This was more apparent in
France even than elsewhere on the Continent; since there papal
authority was tempered by regulations which afforded wider scope to
thoughtful and devout minds ever occupied by the problem of final
salvation and its attainment.
“Two such opposed foes encamp them still
In man as well as herbs, grace and rude will,”
says Friar Lawrence, musing over his “osier cage,” of weeds and
flowers. There had been no time on Christian record that the
question had not exercised theologians, and when it had burnt into
fuller flame, fanned by the ardent soul of Luther, it spread through
Europe and was called the Reformation; but the spirit of it had been
ever present in the Church, and to endeavour to stamp out the
Catholic faith had, in Luther’s earlier days at all events, formed no
part of his desire. Yet scarcely had his doctrines formulated, than the
fanaticism and extravagance of the ignorant and irresponsible seized
upon them, and wrung them out of all size and proportion to fit their
own wild lusts and inclinations, “stumbling on abuse,” striving to
impose their levelling and socialistic views, and establish a
community of goods, and all else in common—even their wives,
though dispensing with clothing as a superfluity and a vanity
displeasing in Heaven’s sight. So Anabaptism ran riot in Germany
under John of Leyden and his disciples; while upon its heels Calvin’s
gloomy and hopeless tenets kept men’s minds seething in doubt and
speculation over grace and free-will, his narrow creed and private
enmity bringing Servetus to hideous and prolonged torture and death
at the stake, for heresy.
Stirred by the revolt of Protestantism on one side, and the claims
of Rome on the other, supported by the Jesuits, speculation gained
increased activity within the pale of the Catholic Church, animated
further by the writings of Jansenius, Bishop of Ypres, whose theories
on grace and the efficacy of good works were grounded mainly on
the viâ media, and it was the following of his opinions by the
illustrious students gathered at Port Royal which created the school
of Jansenists that included such names as Fénelon, Pascal, and so
many others, headed by the Abbé Arnauld, whose sister Angélique
was the Superior of the convent of Port Royal, and whose father, the
learned advocate, had been so stern an opponent to the Jesuits as
to have caused their expulsion from France in the reign of Henri IV.
Readmitted later, they found as firm an opponent in his son, who,
when still quite young, wrote a brilliant treatise against the danger of
Jesuit casuistry.
The convent of Port Royal des Champs was situated on the road
from Versailles to Chevreuse, and hard by, in a farmhouse called La
Grange, “Messieurs de Port Royal,” as the Jansenist priests and
students were called, made their home. They had for their friends the
most distinguished men, scholars and poets of the time; Boileau,
Pascal, Racine were of the band. The place itself is now scarcely
more than a memory. It was then, wrote Madame de Sévigné, “Tout
propre à inspirer le désir de faire son salut,” and hither came many a
high-born man and woman of the world to find rest and peace. Now
a broken tourelle or two, the dovecote and a solitary Gothic arch
reflecting in a stagnant pool, are all that remain in the sequestered
valley, of the famous Port Royal, which early in the next century was
destroyed by royal decree, when its glory had departed, following the
foreordained ruling of all mundane achievement; and the
extravagance of the convulsionnaires and later followers of
Jansenism was stamped out by the bull “Unigenitus” against heresy.
Arnauld’s heart was deposited at Port Royal at his death, with the
remains of his mother and sisters. Louis XIV., as ever his wont had
been to genius and intellect, had invited him “to employ his golden
pen in defence of religion;” but that was before the great king came
under the direction of Madame de Maintenon and Père la Chaise.
But that Madame and her Jesuit confessor would long continue to
regard the Port-Royalists with favour was not possible. Intolerance
succeeded to patronage, and Fénelon was deported to Cambrai,
sent afar from his friend, Madame Guyon, whose order of arrest and
incarceration in the Château de Vincennes was issued very shortly
after Mademoiselle de L’Enclos’ interview with Madame Louis
Quatorze in her Versailles sanctum.
In her dismay, Madame Guyon contrived to fly to Ninon, seeking
protection; but it was of no avail. Without a moment’s delay, Ninon
drove to Versailles, and sought an interview with Madame de
Maintenon on behalf of Madame Guyon. The interview was not
accorded. Nanon—the Nanon of Scarron days, but now
“Mademoiselle Balbien”—was delegated to speak with her.
—“Mademoiselle Balbien,” who gave Ninon to understand that she
was to be addressed no longer as “tu” (“thou”), but as “vous” (“you”),
that the question of Madame Guyon could not even be entered upon,
and under threat of being herself again lodged in the Répenties she
was bidden to depart.
Ninon was at first amazed at this strange reception and insolent
behaviour of mistress and maid. But she was not left long in
perplexity, since “Mademoiselle Balbien” permitted the truth to
escape her prim lips, that Madame de Maintenon had credited Ninon
with the design of introducing d’Aubigné into the boudoir in the
middle of that memorable interview, with the intention of disgracing
Madame in the estimation of the king. That Ninon was not made of
the stuff for this, it is almost superfluous to say. Any sins she might
have to answer for, did not include the hypocrisy with which Madame
de Maintenon had clothed herself about, and almost equally
needless is it to repeat that by no possible means the concealed
presence of the king could have been known by any but the two
most immediately concerned. It could be but a matter of their dual
consciousness.
For six years Madame Guyon remained in prison. Monsieur
Fénelon’s Maximes des Saints was condemned by the Court of
Rome, and the bigotry and hypocrisy ruling Versailles swelled daily.
Molière, alas! was no more, to expose the perilous absurdities and
lash them to extinction; but the comedy of La Fausse Prude,
produced some weeks later at the Italiens, was a prodigious
success. The world greatly enjoyed and admired the fitting of the
cap, built upon the framework supplied by one who had befriended
and sheltered under her own roof the forlorn young orphan girl,
Françoise d’Aubigné.
CHAPTER XXV
The Melancholy King—The Portents of the Storm—The Ambition of Madame Louis

Quatorze—The Farrier of Provence—The Ghost in the Wood—Ninon’s
Objection—The King’s Conscience—A Dreary Court—Racine’s Slip of the
Tongue—The Passing of a Great Poet, and a Busy Pen Laid Down.
The disastrous thrall holding Louis XIV. to Madame de Maintenon,

was an endless theme of wonder and speculation among his
subjects. Very few of them ascribed it to pure unadulterated love and
affection for his old wife—for she was his elder by three years—while
Louis himself was now at an age when the enthusiasm of life slows
into some weariness and languor as it recognises the emptiness and
futility of all mundane things. There were times when he was lost in
brooding thought, and he would wander about his splendid galleries
and salons and magnificent gardens, absorbed, if his dull aspect
expressed the inward spirit, in melancholy reflection. The glory had
departed of his earlier ruling, leaving the nation loaded with debt.
The price had to be paid for those brilliant victories of long ago, and
accumulation of debt on the many later reverses cried for settlement.
The provinces had been deeply impoverished by the absenteeism of
their overlords, whose presence the Grand Monarque had for so
many years required to grace Versailles, attired in their silks and
velvets, sweeping their plumed, diamond-aigretted hats to the
polished floors, bowing and crowding to gaze at the sublime process
of His Majesty’s getting up, promenading with the great ladies
among the fountains and bosquets of Trianon, spending the heaven-
bestowed hours in the sweetness of doing nothing but manipulate
their rapier-hangers and snuff-boxes; while Jacques Bonhomme,
away down in Touraine and Perigord and Berri, and where you will in
the length and breadth of fair France, was sweating and starving to
keep those high-born gentlemen supplied with money in their purses
for the card-tables, and to maintain their lackeys and gilded coaches
in the sumptuous style which was no more than Louis required of the
vast throng. It was in its way an unavoidable exaction, since the few
of the nobility who remained on their own estates had done so at the
peril of incurring the severe displeasure of the king, the Sun-King—
Le roi le veut—whose centre was Versailles.
And still the full time was not yet when all this should be changed.
Even for Louis, the absolute reckoning day was but shadowing in.
“After us the deluge”: that prophetic utterance was spoken long after
Louis was borne to his rest in St Dénis, but when the records of his
life tell of those long-brooding, silent pacings amid the grandeur and
treasures of his splendid palace, comes the question if from afar off
there did not sound the murmur of the flood that was to break some
hundred years hence, if in some dim yet certain way the cloud no
bigger than a man’s hand was not apparent to his introspective gaze,
for as yet the domestic misfortunes of his latest years had not
befallen, death had not robbed him of his heir, and the rest dear to
him; but discontent, not unmingled with contempt, seethed round the
proud King of France. How were the mighty fallen, and how great the
political mistake which indissolubly linked the ambitious woman,
clothed about in her new-found meretricious garb of piety, with his
great responsible destiny—Louis, Dieudonné and elect ruler.
Nor did it stop at the secret, sufficiently open and acknowledged,
of his marriage with Scarron’s widow. The fear was well enough
founded that she was moving earth, and if possible all heaven, to be
Queen of France; but righteousness had small part in the endeavour,
and trickery and chicanery failed to prevail to this crowning end upon
the king’s consciousness and conviction. Pride, and the sense of his
irrevocable bondage, mingled with the poison of the hypocritical
devoutness instilled into him by his wife and her confessor, kept him
silently deferential to this woman, spoiled by prosperity; but she
herself says that all her endeavours to amuse him or bring a smile to
his lips, failed. He had—mildly construing the homely proverb—put
off from shore with a person—more or less mentionable—and he
was bound to sail to land with her.
The diablerie at work was untiring, and had many strings, and
there seem, small, if any, question that to the genius of the
Marseilles merchant’s wife, formerly Madame Arnoul, the curious
tale of the Farrier of Provence is due.
From extreme southward of France came this poor man, who said
he was shoemaker to all the horses of his grace, Monsieur
d’Épernon, at his country mansion near Marseilles—to speak to the
king’s Majesty upon a subject concerning him alone.
The major of the guards to whom he explained his wish, told him
such an interview was impossible. A letter of audience was first
required, and that was to be had only with utmost difficulty. Besides,
he added, the king did not receive all the world. The man objected
that he was not all the world. “Quite so,” said the guard. “By whom
are you sent?”
“By Heaven.”
“Ah!”—and all the bodyguard went into fits of laughter at this reply.
The man stoutly insisted, however, that he had most important
matters to disclose to “the Master of ‘Vesàilles,’” as he phrased it. At
this point of the conversation, the Marshal de Torcy, Colbert’s
nephew, happened to come by. Overhearing what had passed, he
directed that this emissary of Heaven should be conducted to the
ministers, just then sitting in council. They, impressed with the
honest and earnest air of the farrier, informed the king of the affair.
Listening with grave attention to their representation, Louis
commanded the man to be brought before him. Alone with the king,
the farrier unfolded his tale. It was fantastic enough. He was
returning, he said, from the duke’s stables, where he had been
shoeing some of the horses—to his own home, in a hamlet situated
not far off, and was passing through a wood. It was night, and quite
dark; but suddenly he found himself enfolded in a brilliant light, and
in the midst of it stood a tall woman, right in his path. She addressed
him by his name, and bade him repair immediately and without an
instant of delay to Versailles, where he was to tell the king that he
had seen the spirit of the dead queen, his wife, and that she, the
ghost of Maria Théresa, commanded him in the name of heaven, to
make public the marriage he had contracted, which hitherto he had
kept secret.
The king objected that the man had probably been the victim of
hallucination. “I thought so too at first,” replied the farrier, “and I sat
down under an elm-tree to collect myself, believing I had been
dreaming; but two days afterwards, as I was passing the same spot,
I again saw the phantom, who threatened all sorts of terrible
misfortunes to me and mine if I did not immediately do what it had
directed.”
Then the king had another doubt; and asked him whether he was
not trying to impose upon him, and had been paid to carry out the
affair.
The man replied that in order for His Majesty to be convinced that
he was no impostor, he should wish him to reply to one question he
had to ask. “Have you,” he went on, when the king willingly
consented to this, “have you ever mentioned to living soul a syllable
about the midnight visit the late queen-mother paid you in the
Château de Ribeauvillé years ago?”
“No,” said Louis, with paling lips, “I never confided it to anyone.”
“Very well; the ghost in the forest bade me remind you of that visit,
if you expressed any doubt of my good faith; and,” added the man,
as the king said it was very strange, “before disappearing, the tall
white woman uttered these words—‘He must obey me now, as he
then obeyed his mother.’”
The king, in an access of dismay and perplexity, sent for the Duc
de Duras, and related to him in confidence what had passed during
the interview with the peasant. The duke, who was an intimate friend
of Ninon, told her the wondrous tale.
It took no time for her to arrive at the conclusion that Madame
Louis Quatorze and her faithful card-divining friend and fortune-teller,
Madame Arnoul, were at the bottom of the business, and under
promise on the duke’s part of inviolable secrecy, she told him of the
adventure in the Vosges and the very conspicuous part she had
played in it, actuated by her enmity towards de Montespan. The
farrier, she did not doubt, was honest enough; but, simple and
credulous, he had been made the tool of the two women—an easy
prey to Madame Arnoul, who, living at Marseilles, had seen him, and
reckoned him up as suitable for her design.
The duke was of opinion that there was no doubt Ninon’s solution
of the mystery was correct, and he added that, this being the case, it
was her duty to inform the king of it—“For who knows,” said Duras,
“that he may not be weak enough to obey the ghost’s behests, and
disgrace himself and his throne in the eye of all Europe and the
universe, by seating the Maintenon upon it.” It was a most serious
matter—most serious.
Ninon, however, shrank from the suggestion. She was a woman of
courage; but recent experience had taught her the lengths of malice
to which her old friend Françoise could go, and she had no mind to
measure weapons with her again. To make clean confession of the
affair to the king, was simply to bring down upon herself all the
thunderbolts of the hatred of the woman whose ingenuity was never
at fault in plausibility, and the finding the way to retain the kings good
graces at no matter what cost to anyone.
Ninon saw a far better plan than sacrificing herself for the
destruction of the scheme. She begged the duke not to compromise
her to the king; but to represent to him the advisability of sending
competent and trusted persons to the Ribeauvillé château,
accompanied by the duke himself, and there to sound and search
the recesses and panelling of the haunted room and the adjacent
one she indicated, and little more would be necessary to prove to His
Majesty that he had been duped.

Diversity of Methodological Approaches in Social Sciences Example of The Analysis of Media and Online Information Inna Lyubareva Full Chapter

Uploaded by

Copyright:

Available Formats

You might also like

Diversity of Methodological Approaches in Social Sciences Example of The Analysis of Media and Online Information Inna Lyubareva Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Diversity of Methodological Approaches in Social Sciences Example of The Analysis of Media and Online Information Inna Lyubareva Full Chapter

Uploaded by

Copyright:

Available Formats

Diversity of Methodological

Approaches in Social Sciences:

Example of the Analysis of

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2023

Library of Congress Control Number: 2023936103

British Library Cataloguing-in-Publication Data

Overview: Media, Information Pluralism and Methodological

Chapter 1. Online Platforms and Analysis of Community

Chapter 2. Echo Chambers and Opinion Dynamics:

Chapter 3. Diversity of Sources and Pluralism of

Chapter 4. Interviews as a Research Method for

Chapter 5. Media and Information Pluralism Through the

List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

Media, Information Pluralism

O.1. Introduction to the theme of digital transformation of media

The debate about information pluralism – a guarantee of a well-

Overview written by Inna LYUBAREVA and Roger WALDECK.

The current discussions on pluralism take on a new importance due to the

In particular, the development of digital platforms and services has lowered

This proliferation leads us to rethink the meaning and relevance of

1 That is, journalism focused on the speed of information production.

institutions have played the role of gatekeepers by filtering news according

Therefore, it is worth questioning to what extent this increase in content

diversity of sources mechanically improve the diversity of content, which in

In summary, the digitization of the media and the current explosion of

O.2. Methodological choices in social science research on digital

In the social sciences, traditional paradigms are regularly questioned

material constraints has been particularly criticized because it neglects a

Today, the massive diffusion of digital technologies and their uses

One of the important consequences of the associated epistemological

At the heart of this approach is a fruitful collaboration between different

development, in addition to the hypothetico-deductive approach, which is

When we talk about the digital context – characterized by actors

For instance, when it comes to the construction of a research project, one

In addition, given the extremely rapid pace of innovation and

O.3. Introduction to the chapters in this book

The analysis of media and information pluralism as a complex and

perspective different methodological approaches, some of which are well

Each chapter is devoted to a particular methodology and aims, beyond a

Axis 1: Platforms master the function of infomediation by putting

Chapter 2 proposes the agent-based modeling approach to analyze the

Axis 2: A number of studies have evaluated information pluralism

radio stations or television channels would be more conducive to the

In order to study the pluralism of sources and journalistic framings in the

Camerer, C., Loewenstein, G., Rabin, M. (2004). Advances in Behavioral

Lyubareva, I. and Rochelandet, F. (2016). L’évolution des modèles d’affaires dans

Online Platforms and Analysis

Online platforms, such as digital social networks, structure new ways of

Chapter written by Cécile BOTHOREL, Laurent BRISSON and Inna LYUBAREVA.

and networks by structuring their organization and by making them more

Our work suggests that instead of imposing an a priori perimeter of

This perspective highlights the relevance of social network analysis

SNA designates a set of methods, notions and concepts based on graph

Finally, from an intermediate point of view, SNA allows for the

Therefore, in SNA, the relationships between individuals (the nodes),

More specifically, in the case of Ulule we highlight that, based on the

In the case of YouTube, where the choice of the community detection