Content uploaded by David Krieger
Author content
All content in this area was uploaded by David Krieger on Feb 21, 2024
Content may be subject to copyright.
From Systems to
Actor-Networks
A Paradigm Shift in the Social Sciences
By
Andréa Belliger and David J. Krieger
From Systems to Actor-Networks: A Paradigm Shift in the Social Sciences
By Andréa Belliger and David J. Krieger
This book first published 2023
Ethics International Press Ltd, UK
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Copyright © 2023 by Andréa Belliger and David J. Krieger
All rights for this book reserved. No part of this book may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical photocopying, recording or otherwise, without the
prior permission of the copyright owner.
Print Book ISBN: 978-1-80441-336-4
eBook ISBN: 978-1-80441-337-1
Contents
Introduction ............................................................................................. ix
I. General Systems Theory
Chapter 1: Basic Concepts of Systems Theory ..................................... 1
1.1 In the Beginning, There Is the Difference ............................. 1
1.2 What is a System? .................................................................... 2
1.3 System and Environment ....................................................... 4
1.4 Why Are There Systems? Reduction of Complexity
(Negentropy) ............................................................................ 5
1.5 How do we Recognize Systems? Function and
Functional Analysis ............................................................... 12
1.6 What are Systems Made Of? Order, Organization, and
Structure .................................................................................. 15
1.7 What Do Systems Do? Operation and Process .................. 16
1.8 The Mechanistic Model – The System as Machine ........... 16
1.9 Code and Information ........................................................... 17
1.10 Cybernetics and Feedback Loops ........................................ 24
1.11 Purpose, Teleonomy, and Homeostasis .............................. 25
1.12 Functional Differentiation, Subsystems, and Internal
Complexity ............................................................................. 26
1.13 Contingency and Degrees of Freedom ............................... 27
1.14 Indeterminate and Determinate Complexity/
Contingency ........................................................................... 29
1.15 Autonomy, Learning, Emergence, Levels of Emergent
Order, Evolution, Self-Organization ................................... 31
1.16 Excursus: The Self-Organizing Universe............................ 37
1.17 The Biological Model – The System as a Living Being.
Autopoiesis, Allopoiesis, and Autonomy ........................... 42
1.18 Operational and Informational Closure, Self-Reference .. 45
1.19 Structural Coupling, Adaptation, Viability, and Again
Evolution ................................................................................ 48
1.20 Excursus: Ecology and Sustainability ................................. 52
1.21 What is learning? ................................................................... 56
1.22 Meaning Systems, the Semiotic Code, and Cognition ..... 57
1.23 Organism or Society? On the Problem of Biological
Reductionism in Systems Theory ........................................ 60
Chapter 2: Meaning as a System: Limits and Possibilities of the
Systems Approach in the Social Sciences ...................................75
2.1 Observation............................................................................. 75
2.2 Self-Reference ......................................................................... 79
2.3 Meaning .................................................................................. 83
2.4 A Critique of Luhmann’s Theory of Meaning ................... 86
2.5 Subsystemic Differentiation or Internal Complexity of
the Meaning System ............................................................ 102
2.6 Psychic Systems and Social Systems ................................. 107
2.7 The Social System ................................................................ 125
2.8 Communication ................................................................... 139
2.9 Communication, Subject, and Action ............................... 142
2.10 The Differentiation of Communication: Functional
Subsystems ........................................................................... 156
2.11 Ecological Communication or Can Society be Modeled
as a System? .......................................................................... 163
II. The Network Paradigm
Chapter 3: Network Science ............................................................... 185
3.1 Network Science .................................................................. 185
3.2 Actor-Network Theory ....................................................... 192
3.3 Translation and Enrollment .................................................. 201
3.4 Information ............................................................................. 221
3.5 Society: System or Network? ............................................. 230
3.6 Network Differentiation ........................................................ 247
III. From Systems to Networks – A Conclusion and a New Beginning
Chapter 4: From Systems to Networks – A Conclusion and a
New Beginning ............................................................................ 265
4.1 Digital Transformation ....................................................... 268
4.2 Network Norms ................................................................... 273
4.2.1 Connectivity .................................................................. 273
4.2.2 Flow ................................................................................ 276
4.2.3 Participation .................................................................. 277
4.2.4 Transparency ................................................................. 277
4.2.5 Authenticity ................................................................... 278
4.2.6 Flexibility ....................................................................... 279
4.3 From Systems to Networks ................................................... 279
4.3.1 Networks Do Not Define Fixed Roles and
Functions ........................................................................ 280
4.3.2 Networks Have No Constitutive Boundaries ........... 289
4.3.3 Networks Cannot be Controlled Top-Down ............ 305
Literature ............................................................................................... 322
Introduction
This book attempts to document a tectonic shift in world views, or what
can be called a “paradigm shift,” not only in the sciences but also in
society. There are many indications that in our time, a particular way
of understanding the world, which in its depth and breadth could be
compared to a tectonic plate, is by a process of historical and cultural
subduction being pushed beneath another. A new way of seeing the
world is consequently rising to the surface. These two tectonic plates are
continents of knowledge, values, basic assumptions, and methods and,
for this reason, can be called scientific paradigms. The two paradigms
we will be concerned with in this book are the “systems paradigm” and
the “network paradigm.” Systems are everywhere, or at least they used
to be. Now we are seeing that where systems once were, networks are
appearing. Systems and networks are two different but related forms of
order. Order in all its forms is what science investigates and attempts to
understand. Striving for order is at the basis of societies, cultures, ideol-
ogies, and world views. Science is part of this program. We will argue
that there are three levels upon which order appears: 1) the physical
level of matter, energy, particles, and fields, which is the object domain
of physics, chemistry, and related disciplines; 2) the order of life, which
is studied by biology; and 3) the order of meaning which is described
by philosophy and the human and social sciences. A scientific theory
becomes a paradigm when it proposes to model order on all three levels.
A paradigm is, therefore, a general or universal theory because it claims
not only to explain one domain of reality but all of reality, including
thereby – and this is very important – itself. Physics may claim to be
a Theory of Everything, but it can neither account for life nor for the
observer, that is, for itself, at least not yet. The same can be said of
biology. Where in the brain, we may ask, is the observer to be found?
We find neurons and patterns of their interaction, but we do not see the
observer. We see, for example, correlations between the firing patterns of
neurons and the vision of a rose. However, correlation is not causation
and indeed not identity. Where is the similarity, sameness, identity of
patterns of firing neurons, and the idea of a rose? This leaves the human
and social sciences, which do explicitly study themselves. Their problem
From Systems to Actor-Networks
x
is to describe the kind of order they represent. Traditionally, philosophy
and the human sciences have located their object domain on the level
of meaning, albeit under many different names and mostly in opposi-
tion to nature. When these sciences include basic concepts and models
applicable on the physical and biological levels, they become universal
theories that claim to explain everything, including themselves.
In recent times, two general theories claim universality in that they bring
together concepts and methods that are claimed to be applicable on all
three levels of emergent order, the physical, biological, and the level of
meaning. These are general systems theory and network theory. For
general systems theory, all forms of order are systemic, from fundamental
particles to living beings, human culture and society, and even the world
(Earth, Gaia, Cosmos). Ecology, as we will see, has become a title for
universal theories, and both systems theory and network theory offer
their own “solutions” to the problem of a holistic understanding with
which the Anthropocene confronts us. For systems theory, at least since
the founding of cybernetics in the 1940s, it has become clear that both
nature and human society are forms of systemic order. Network theory,
at least actor-network theory, which we will examine in this book, explic-
itly rejects the typically modern distinctions between nature and society,
subject and object, mind and reality. This implies that when speaking of
meaning, one is speaking of being on all levels. According to these two
theories, whenever we find order, whether in matter, life, or meaning,
what we see are systems of one kind or another or networks in various
forms. General systems theory, as opposed to the various special forms of
systems analysis, is “general” because it describes systemic order as such,
the general principles of any kind of systemic order, no matter whether
physical, biological, or social. This is what allows it to claim the status of a
paradigm. Actor-network theory also strives for a level of generality that
encompasses everything. It does this by attempting to describe the mech-
anisms or processes that construct networks as such, no matter how small
or large and no matter what actors, whether humans or nonhumans or
both, are linked together. The principles of network order are claimed to
apply on the physical, biological, and social levels, mainly because these
distinctions are simply rejected. Whereas systems analysis emphasizes the
Introduction xi
reduction of complexity, holism, teleology, control, and stability, network
thinking emphasizes connectivity, flow, participation, flexibility, and
other values which we will discuss below that systems do not exhibit.
This book examines the claims of these theories critically and argues
that we are witnessing a paradigm shift from systems to networks in our
approach to science, society, and ourselves. This may seem a daunting
task, and it indeed is, but the focus on generality unburdens us of the
necessity of delving into details of systems or network analysis in the many
different disciplines in which they are to be found. In a way, describing
a paradigm is much easier than applying it. This means that the heroes
of our story are theories, their basic concepts, their coherence, heuristic
power, and their explanatory scope. The two thinkers that we will take as
representatives of the theories discussed in this book are Niklas Luhmann
for systems theory and Bruno Latour for actor-network theory. However,
even if we let Luhmann and Latour do most of the talking, this is not a
book about Luhmann or Latour or both together. It is a book about theory
construction. The goal is to construct at least the outlines of a theory that
answers to the needs of our time, and we are simply using what, in our
opinion, are the best available resources. Although we talk about the social
sciences, as do both the theories we will be examining; we are not doing
social science in any traditional sense of the term.1 We will accordingly not
be primarily concerned with the many different forms of social systems or
with the various networks making up the “collective,” which is Latour’s
word for society. Instead, our focus is on the principles of systemic order
as such and correspondingly on the general mechanisms of network
construction. We will be concerned with the question of whether meaning
can be successfully modeled as a system and, if not, whether the network
model could be more successful. Posing this same question somewhat
differently, we ask whether systems theory can explain itself and, if not,
whether the network model offers a more adequate answer.
We begin in Part 1 by examining a theory that has several names, cyber-
netics, the theory of self-organization, complexity theory, constructivism,
1 Both Luhmann and Latour reject the traditional self-understanding of the social
sciences.
From Systems to Actor-Networks
xii
or, as it is commonly known, systems theory.2 Whatever name it goes by, we
are dealing with an interdisciplinary and self-proclaimed universal theory
that seeks to “explain” order not only on the physical and biological levels
but also psychological and social phenomena. Just as a spider’s web is fixed
on all sides at different and often widely separated points, so today what
could be called “general systems theory” is based on different and quite
distant disciplines: cybernetics and computer science, cognitive science,
physics, engineering, biology, logic and mathematics, psychology, neuro-
physiology, anthropology, sociology, semiotics, organization theory, and
philosophy are among the sources of systems models and concepts. When
referring to a general systems theory, we must keep in mind that because
of these many sources and applications, we are dealing much more with
a “paradigm” (László, 1974) than with a unified theory with established
methods and particular domains of application. The systems paradigm
consists of a more or less closely associated set of basic concepts, terms,
approaches to phenomena, methods, and beyond this, a worldview that
understands all forms of order as in some way “systemic.” Whenever one
speaks of systems, whether in physics, chemistry, biology, psychology,
sociology, economics, computer science, engineering, neuroscience,
management studies, etc., one usually finds the associated concepts of
self-organization, complexity, adaptation, self-regulation, learning, infor-
mation, feedback, holism, teleonomy, control, homeostasis, etc.
Despite the tendency to use similar concepts in many different disciplines,
which is typical of a scientific paradigm, general systems theory “...cannot
currently be presented as a consolidated set of basic concepts, axioms,
2 The idea of an interdisciplinary, universal research program under the name of
systems theory originated with Ludwig von Bertalany. Bertalany’s systems
theory, together with Shannon and Weaver’s (1949) information theory and Wie-
ner’s (1948) cybernetics, Maturana and Varela’s theory of autopoiesis, as well as
Bateson’s work in psychology and anthropology, and Parsons’s sociology, Beer
in organization theory, and more recently theories in biophysics, cognitive sci-
ence, articial intelligence and associated disciplines form a model, basic con-
cepts, and methods that can be found in a wide variety of scientic disciplines.
Even though cybernetics and systems theory are not current buzzwords, their
inuence is everywhere, so it is justied to speak of a systems paradigm in the
sciences and beyond science in our current cultural vernacular. Since there are
many kinds of systems on the physical, biological, and human levels, we will
speak of systems (plural) instead of system (singular).
Introduction xiii
and derived propositions” (Luhmann, 1984, p. 34). This is because the
systems theory research program is not a specific discipline but rather a
discourse, i.e., a particular way of talking, wherein the most diverse topics
are approached using similar terms and similar assumptions. Therefore,
the question arises: “...on which common or at least related concep-
tual foundations this transdisciplinary cooperation is based” (Krohn/
Küppers/Paslack 1987:459).3 For example, what do we mean when we
talk about feedback loops, reduction of complexity, autopoiesis, infor-
mation construction, codes, self-organization, homeostasis, adaptation,
evolution, sustainability, etc.? To what objects and processes can we legit-
imately apply such terms? Do these terms have a common meaning when
applied to very different phenomena, or are they mere equivocations? To
what extent is it useful or even possible to transfer physical and biological
models to the humanities and talk about cognitive, psychological, and
social systems, or the other way around, to speak of physical systems
as information processing and life as cognition? And if this is common
practice, on what explicit and implicit presuppositions is this based?
The self-evidence with which systems theoretical concepts are used in
psychology, sociology, anthropology, philosophy, business, politics,
education, and indeed everywhere challenges us to ask these and similar
questions.4 The attempt to present an overview of the models and basic
concepts of systems theory thus seems to be the order of the day, and this
is especially so when systems theory actually functions like a “paradigm,”
informing, implicitly if not explicitly, all aspects of life and society.
Our discussion of the systems paradigm does not intend to trace the
historical lines of development of the various systems theoretical
approaches in the sciences, nor to describe the various versions of systems
3 All translations from German language publications listed as such in the Litera-
ture are by the authors.
4 On the ubiquity of systems thinking and assumptions, see, for example, the
self-evident value of “sustainability” as exemplied by the WHO Sustainability
Development Goals, which claim to “transform the world” (hps://www.who.
int/europe/about-us/our-work/sustainable-development-goals). Also notewor-
thy in this regard is the new critique of “environmentalism” and the idea of a
“General Ecology” (see Hörl/Burton 2017), and nally, there looms the prospect
of a data-driven society governed by algorithms (see Kalpokas 2019; Schuilen-
berg/Peeters 2020 for an overview).
From Systems to Actor-Networks
xiv
theory currently in use,5 but rather to attempt to set forth the conceptual
framework or the conceptual “common property” of systems theoretical
discourse. It should be emphasized that this is done from a particular
perspective. We are guided by the question of the extent to which systems
theory can successfully be applied on the specifically human level of order,
that is the level of meaning. We will not be primarily concerned with phys-
ical or mechanical or biological systems but with meaning systems. Our
focus will be on the level of emergent order that can be called meaning,
that is, the form of order that philosophy and the social sciences study
and to which they also belong. This is, therefore, not a book on physics,
biology, cognitive science, or artificial intelligence, even though we will
have occasion to refer to these disciplines. Our concern is with the extent
to which order on the level of meaning can be understood as systemic.
And if this turns out not to be the case, what theoretical model could do a
better job? This is where the comparison of systems to networks and the
proposals of the new network paradigm become relevant.
The justification of prioritizing meaning over matter and life lies in the
fact that whatever we know about matter and whatever we know about
self-organizing, adaptive systems, otherwise known as life, we know
this on the level of meaning and as forms of meaning. As Heidegger
never tired of pointing out, what is closest to us and what comes first
is not knowledge of objective facts that science presents but the world
of meaning that includes all aspects of our existence. Science, after all,
is a social phenomenon, part of our social practices, and not the other
way around, and this remains so even if society becomes the object of
scientific investigation. If everything we can know and be, including all
the different sciences, is primarily ordered on the level of meaning, then
5 This has been done by some of the principle actors themselves, various organiza-
tions, as well as later commentators. For an overview, see the Wikipedia article on
System Theory hps://en.wikipedia.org/wiki/Systems_theory#cite_note-:1-5. For
historical descriptions, see the website of the American Association for Cybernetics
hps://asc-cybernetics.org/foundations/history.htm; also, the website of Principia
Cybernetica Web hp://pespmc1.vub.ac.be/CYBSHIST.html. And the website of the
International Society for Systems Science hps://www.isss.org/home/. Recent mono-
graphs include Malapi-Nelson (2017) and the work of Bruce Clark (2019). Short his-
tories have been oered by Umpleby (see papers at: hps://blogs.gwu.edu/umpleby/
papers-on-cybernetics-and-systems-science/; 2008) and Adams et al. (2013).
Introduction xv
it is on this level that any claims for the universality of a theory must
be justified. It must be shown that a theory that claims to be universal
can explain meaning on its own terms and not reduce it to matter or
life. A theory that cannot adequately explain itself cannot claim to be
a universal theory. No matter how much of reality it covers, it leaves
something out, namely, itself. Taking this into account requires that
we pay more attention in this book to philosophical and sociological
systems theory than physical or biological versions of the same. Since
the most prolific and influential sociological systems theorist is Niklas
Luhmann, we will be following the exposition of systems theory in his
work as our primary source.6 Following Luhmann, we will focus on
the idea of communication as a system. In emphasizing communication,
Luhmann is not alone. For many authors and in many respects – see the
linguistic turn in philosophy, semiotics, structuralism, and poststruc-
turalism, and Habermas’ Theory of Communicative Action – the consti-
tution of meaning in linguistic (also non-verbal!) communication has
acted as an “attractor” for the study of human existence in society.
The fruitfulness, heuristic power, and above all, the universality of
systems theoretical models and concepts have been demonstrated on the
physical and biological levels, but can systems theory also successfully
model meaning? How does meaning arise? What forms of order does
meaning take on? Is meaning a systemic phenomenon or something else?
As mentioned above, from today’s scientific point of view, a theory can
only fulfill its claim to universality if it can “explain” not only everything
in the world but also itself. Physical, mechanistic, and biological theories
notoriously omit the observer even while paying lip service to the need
to explain who it is that is doing the science and building the machines
in the first place. Where in the standard model is the observer? Which
6 None of the theories of 2nd order cybernetics that have arisen in the wake of
Bateson and von Foerster in the English language come near to the theoreti-
cal sophistication and “systematicity” of Luhmann’s work which incorporates
their insights and goes beyond them. Furthermore, another reason for focus-
ing on Luhmann’s extensive work is that this work is not well-known in the
English-speaking world. Very lile of Luhmann’s many writings have been
translated, and even less discussion has been generated by his work. We hope
with this book to contribute to making Luhmann beer known outside the Ger-
man-speaking world.
From Systems to Actor-Networks
xvi
particle or field is it? Embarrassingly, the theory calls for the observer
and even depends upon it but cannot explain it.7 Perhaps the observer
is a fundamental force of nature beyond the four known forces. If so, it
has not been found, measured, or seen in the decay patterns of the LHC.
Could it be that physics would need to look beyond matter to find itself?
The same can be said of biology. Where is the observer in an organism? Is
it the central nervous system? Can it be found in the brain? The fact that
this has not been done and does not seem anywhere near being done,
not even in the artificial brains devised by computer scientists, is called
the “hard problem” of consciousness.8 The problem is “hard” because it
seems that the observer is not anything physical or biological at all. If one
attempts to make the subject into an object, consciousness disappears,
and if you try to know consciousness, it appears to be something that we
do and not something we can know as an object. Neither physical nor
biological science seems able to find anything like an observer beyond
merely adaptive behavior in the objects they study. This means that they
cannot explain their own activities, how they themselves are possible, or
even what they are as scientists in society.9
Science itself is not a physical or biological phenomenon even if it admittedly
takes place in some kind of still, despite claims to the contrary, inexplicable
relation to physical and biological substrates. A substrate, whether it be
matter with regard to life, or life with regard to meaning, is not a cause; it
does not determine anything, and at the most, it constrains. What emerges
out of it, on top of it, in relation to it, cannot be derived from it. This is why
7 For a detailed discussion of the “nonseparation” of observer and observed in
quantum physics and the ongoing unwillingness of scientists to correspondingly
revise their basic notions of reality, see d’Espagnat (2006).
8 See Wikipedia: hps://en.wikipedia.org/wiki/Hard_problem_of_consciousness.
Generally, the term “system” is used in physics to describe any object of study
that consists of interrelated parts. The delimitation or designation of the object is
done by the observer.
9 Bateson (1979) notoriously proposed that mind and nature are one and empha-
sized the continuities between life and meaning as many after him have also
done, for example, Michael Levin and Karl Friston who, following Maturana
and Varela, equate life, that is, adaptive behavior, with cognition. Whether “cog-
nition,” as the term is used in cognitive science, can be equated with meaning
is a question we will discuss below. As we shall see, it was left for Latour and
actor-network theory to examine how science actually operates.
Introduction xvii
one speaks of emergence, that is, the appearance of something that cannot
be deduced or reduced to what precedes it. On the one hand, the question
of what science is must itself become a scientific question, while on the
other hand, the science of science has yet to find a theoretical foundation
since it can neither be derived from matter nor from life and can only be
explained by an adequate theory of meaning. Does systems theory give
us an adequate theory of meaning? Could it be that with general systems
theory meaning finally succeeds in understanding itself, or is it not that
with systems theory meaning misunderstands itself as a system? Consid-
ering that systems theory claims to be a universal theory, that is, a theory
of everything, we must ask if systems theory can do what physics and
biology cannot. We must ask: Can systems theory itself be understood and
explained as a system? If systems theory cannot adequately explain itself,
does this mean that systems theoretical concepts require further justifica-
tion on an even more general level? If the discourse of systems theory wants
to demonstrate its universality, then the epistemological and metaphysical
presuppositions, that is, the implicit or explicit theory of the meaning of the
systems theoretical paradigm, must be clarified. The observer must find
itself in the theory and appear as a result of the theory. This was precisely
the program of what has been called 2nd order cybernetics or constructivist
epistemology.10 Has this program been successful?
If it should turn out that systems theory cannot adequately model itself,
and this will be the upshot of our argument in this book, then we must
ask what can. This is where we introduce in Part 2 the network para-
digm.11 After examining Luhmann’s theory of social systems in detail
10 If one equates meaning with logic or conceptual coherence, as Luhmann in the
end does, but Latour clearly does not, then the legacy of Hegel in the systems
paradigm becomes apparent. For an overview of 2nd order cybernetics see Wiki-
pedia hps://en.wikipedia.org/wiki/Second-order_cybernetics#:~:text=More%20
generally%2C%20second%2Dorder%20cybernetics,it%20is%20termed%20
as%20such, and for constructivism: hps://en.wikipedia.org/wiki/Radical_con-
structivism#:~:text=Radical%20constructivism%20is%20an%20approach,the%20
world%20beyond%20that%20experience. Current inuential proponents of this
program are Levin (2022) and Friston (2010).
11 For network science in physics and biology, see the work of Albert-Laszlo
Barabasi (hps://barabasi.com/), and for the social sciences, see the work of Bru-
no Latour (hp://www.bruno-latour.fr/).
From Systems to Actor-Networks
xviii
and attempting to show that there are many problems with the attempt
to model meaning as a system, we turn to a competing scientific para-
digm that claims to be able to do this. The network paradigm, just like the
systems paradigm, is not a unified theory but a set of related concepts and
approaches to reality which share a common worldview. This is a view
of the world as being networked, that is, constructed by means of associ-
ations, links, relations, and, as we shall argue, information. Many prom-
ising developments use network models in physics, computer science,
and artificial intelligence. Wolfram’s multigraph12 model for a physical
theory of everything and Goetzel’s OpenCog13 model for AGI (artificial
general intelligence) could well be cited as examples of the network para-
digm. As interesting as these examples are, we will take our sociological
focal point as a guide and rely on the work of Bruno Latour and what is
known as actor-network theory to explicate the network paradigm. The
focus will be on the social sciences and the theory of meaning.
Although Latour does not explicitly develop a theory of meaning, he clearly
assumes that actor-networks are at once that of which the world consists
and that this world is not to be understood according to the traditional,
modern dichotomies of nature vs. society, subject vs. object, matter vs.
mind, language vs. reality. Indeed, all the pillars of modern epistemology
and ontology are pushed aside to make room for a close description of
how humans and nonhumans enter into associations in order to construct
the world, which Latour calls the “collective.”14 Since it consists of associa-
tions, links, or relations, the collective can be said to be “networked.” Here
is where actor-network theory is similar to what has come to be known
as “network science.” For the newly emerging science of networks, all
forms of physical, biological, or social order exhibit network characteris-
tics. They can be understood as relations or links between nodes. Where
systems theory finds order in functional relations between the elements
of a closed system, network theory finds order in the relations between
12 See hps://writings.stephenwolfram.com/2020/04/nally-we-may-have-a-path-
to-the-fundamental-theory-of-physics-and-its-beautiful/.
13 See hps://en.wikipedia.org/wiki/OpenCog; hps://opencog.org/.
14 The collective is everything that associations and relations gather together, and,
as we shall see, this includes everything that exists.
Introduction xix
actors in an open and flexible network. Where systems theory proclaims
there are systems, network theory finds everywhere not systems but
networks. We interpret Latour’s “principle of irreduction”15 namely, that
nothing can be either reduced to another, or reduce another to itself, by
which he describes the relational nature of all beings as a definition of
information. We argue that actor networks are the results of information
construction. This argument is based on a close analysis of the mechanisms
of networking. Latour speaks of these mechanisms or processes in terms of
“translating” and “enrolling” actors that are both human and nonhuman
into networks. He also calls this “technical mediation” since meaning is
always embodied in associations between humans and nonhumans, that
is, things and artifacts. For actor-network theory, there is no disembodied,
purely virtual, or ideal meaning. Meaning is Being. In short: If there is
nothing in the hand, there is nothing in the head either. After describing
how actor networks are constructed, we review Latour’s description of the
various forms of networks to be found in modern society. Both systems
theory and actor-network theory offer empirical descriptions of society.
For Luhmann, society consists of functional subsystems, whereas for
Latour, society is a “collective” of “modes of existence.”
In Part 3, we move beyond a purely theoretical comparison of systems theory
and actor-network theory and attempt to empirically describe the shift from
systems to networks in various areas of society. What does the advent of the
network paradigm mean for society? We attempt to answer this question
with reference to networking in business, education, and healthcare, which
are representative for other areas of society as well. We describe how the
guiding values arising from the affordances of digital technologies which
dominate our historical moment lead to the transformation of systems into
networks with regard to three characteristics: 1) As opposed to systems,
networks do not define fixed roles and functions, 2) as opposed to systems,
networks do not have constitutive boundaries, and 3) as opposed to systems,
networks cannot be controlled top-down. A description of how these basic
differences of systems to networks in the areas of business, education, and
healthcare illustrate the emergence of a global network society.
15 See Latour (1993).
Part I
General Systems Theory
Chapter 1
Basic Concepts of Systems Theory
1.1 IntheBeginning,ThereIstheDierence
Let us start at the beginning, i.e., with nothingness. Let us imagine that
there was only some kind of primal substrate at the beginning of the
world. The original substrate is not differentiated and not formed. We
can think of it according to the atomistic model as consisting of infinitely
many uniform elements, which have complete freedom to form
connections. Or we can use a distinction that Luhmann preferred, the
distinction between medium and form, wherein the medium consists
of “loosely coupled” (Luhmann, 2001) elements, and form arises when
these elements become “closely coupled” (verdichtet) or in some way
bound to each other such that order or a pattern arises and no longer
everything is possible. In the primordial substrate, or medium, every-
thing is the same. There are no differences. One cannot even distinguish
among the elements since they are all the same, and there is no pattern
or order, no relations binding them together in any particular way.16
If we now want to put ourselves in the role of God and create a world,
then we must begin with a distinction. In the beginning, was the
distinction, the difference. The idea of difference is a basic principle in
systems theory. We can even speak of “difference theory” (Luhmann,
2006) since everything is based on a difference or a distinction. If there
were no distinctions, there would be nothing. Identity, substance, and
bonded unity is consequently not the main concept of systems theory.
16 The fact that the primal substrate or medium consists of elements explains why
dierentiation is possible. If the original substrate is a single, massive “thing,”
then nothing could develop from it; internal dierentiation and, thus, negentro-
py would be unthinkable. The paradox of undierentiated dierentiation could
be overcome by asserting that the dierentiation of the primal substrate con-
tains no information. These are dierences that make no dierence. Order, on
the contrary, is information. Other words for this substrate in systems theory are
“complexity” and “entropy.”
From Systems to Actor-Networks
2
It is the boundary and not the thing that is more important. This implies
that the primal elements, whatever they might be, cannot be thought
of as some kind of things whose essential properties determine what
the world can become. Order is relational and not substantial. In this
respect, systems theory fundamentally differs from the Western philo-
sophical tradition with its emphasis on substance or things. Identity is
not what makes something to be what it is. Identity is not why some-
thing exists, its ground of being. In systems theory, something “is” only
by being different from something else, i.e., by being a relation and not
a thing. Systems ontology is not an ontology of things; it is a relational
ontology.
Where differences in the form of relations and not mere multi-
plicity come from is a question we can leave open for now. Whether
differences are arbitrarily introduced by a mythical creator god or a
perhaps equally mythical “observer,” whether differences arise in the
evolutionary process of the self-organizing universe (Bateson, 1979;
Maturana/Varela, 1987; Jantsch, 1980), whether they are historically
culturally predetermined (Schmidt, 1994) or, finally, whether they are
elements of a communication system specialized in the production of
differences (Luhmann,1990b), these are questions that will be discussed
below. For the moment, it is only important to note that systems are
based on a distinction.
1.2 What is a System?
We now know what needs to be done if something is to be created from
the primordial substrate: A distinction must be made. But how is a
distinction to be made? A distinction is a delimitation. Something must
be delimited. However, this is only possible if, to remain with the idea
of an undifferentiated original substrate/medium in which all elements
are equal, when some elements are selected and put together, that is,
related in a certain way. Elements must be selected and related to each
other or composed in a particular way. A composition of this kind is
called in Greek to systeme.
Basic Concepts of Systems Theory 3
A “system” is an ordered whole, something that has been put together
in a particular way. By putting things together in a certain way, a
distinction arises. It is not enough to just put some elements aside –
where? – and it cannot be enough to build a wall around some elements
– how do we imagine this if nothing is yet present in the world, not even
walls? The parts of the original substrate, which are to be selected and
related and thus distinguished from all other parts, must, to be distin-
guished, be ordered in a certain way. The parts must enter into certain
relatively stable relations with each other. Otherwise, they are the same
as all others, and nothing distinguishes them. Differentiation, therefore,
takes place via relationing. Selection without relationing makes no sense
and achieves nothing. However, selection and relationing, necessary
as they are, are also not enough. The elements selected are related in
a certain way such that something can be done, and some operation
becomes possible. The parts are steered or controlled in order to operate to
achieve some goal or purpose. A system, therefore, consists of processes
of selecting, relationing, and steering/controlling. These are the funda-
mental principles of systemic order. Selection without relationing is no
selection because the elements are not in any way distinguished from
all others in the substrate. Relationing without controlling or steering
operations has no purpose, and there would be no reason for anything
to be related in any particular way.
To make a distinction that builds a system, we must 1) select some
elements from the totality of everything and 2) relate these elements
among themselves in a certain way, that is, 3) in a way that allows
certain controllable operations. If one satisfies these three conditions,
one has a system. A system, then, consists of elements that stand in
certain relations to each other, which relations then enable certain
operations or processes that pursue a goal due to control or steering of
the elements and their relations. This steering function accounts for the
close relationship between systems theory and cybernetics. The Kuber-
netes (Greek) was the helmsman of a ship, the one who steers the ship or
also society. Therefore, the Kubernetes was also the governor. It is from
this Greek word that the title cybernetics comes. In a cybernetic system,
one speaks of a governor or controller that steers the system toward
From Systems to Actor-Networks
4
specific goals or setpoints. This third principle of steering explains why
systems theory and cybernetics are the same. Cybernetics emphasizes
the steering function, which, as we shall see, is associated with feedback
loops, circular causality, information, and stability. Systems theory
treats all three functions as equally important. One could therefore
argue that systems theory is a broader and more encompassing term,
whereas cybernetics emphasizes one albeit central aspect of systemic
order. These three basic principles of selection, relationing, and steering
will accompany us all the way and in every domain of systems theo-
retical discourse. They build the foundation of the epistemological
and ontological presuppositions of systems theory. They are the basic
principles of systemic order. However, as important as they are, there is
more to general systems theory than these principles.
1.3 System and Environment
If one has created a system by a distinction that arises based on selection,
relationing, and steering, the system must be different from something
else. That which is different from the system, that which is “outside”
the system, that which contains everything that does not belong to the
system, and that with which the system in some way interacts is the
environment. The system is fundamentally constituted by a difference
from the environment. This is because the system necessarily includes
its own elements and excludes everything else: the environment. The
system/environment difference is at once inclusive and exclusive. The
environment is everything that is not an element in the system and is
under the control of the system and thus excluded from the system.
The idea of selection implies that not everything can be included in
the system. This has important consequences. It means there can be
no environment in itself, and as a corollary, there can be no system of
everything, no world system.17 The system/environment difference is
also a relation. The environment is always also the environment for or
in relation to a system. Systems and environment necessarily imply
17 We will return to this later when discussing the extent to which meaning and the
idea of ecology, both of which include everything, can be modeled as systems.
Basic Concepts of Systems Theory 5
each other, condition each other, and belong to each other. There is no
system without an environment, and no environment without a system.
In our systems theoretical world creation myth, the original substrate/
medium cannot represent an environment before the world’s creation,
i.e., as long as there is no system that, by distinguishing itself, “makes”
what is excluded into an environment. In this sense, one can probably
speak of creation “out of nothing.” Luhmann (1995:16) states:
There is agreement within the discipline today that the point
of departure for all systems-theoretical analysis must be the
difference between system and environment. Systems are oriented
by their environment, not just occasionally and adaptively, but
structurally, and they cannot exist without an environment.
They constitute and maintain themselves by generating and
maintaining a difference from their environment, using their
boundaries to regulate that difference. ... In this sense, boundary
maintenance is system maintenance.
Moreover, concerning the environment, Luhmann (17) writes:
The environment receives its unity through the system and only
in relation to the system. It is delimited by open horizons, not
by boundaries that can be crossed. Thus, it is not itself a system.
It is different for every system because every system excludes
only itself from its environment. Accordingly, the environment
has no self-reflection or capacity to act.
1.4 Why Are There Systems? Reduction of Complexity
(Negentropy)
Let us return to the primordial substrate or medium. The original
substrate excludes nothing, so it cannot be a system. It is the “open
horizon” containing “no boundaries that can be crossed.” It is not a
product of processes of selection, relationing, and steering. Accordingly,
the number or quantity of possible elements in the original substrate is
infinitely large, and also infinite is the number of possible connections
From Systems to Actor-Networks
6
among the elements. In the primeval substrate, everything is possible
because every element can be connected with every other in all possible
ways. This follows from the fact that in the primordial state, there are
no given selections, relations, determinations, limits, or regulations that
could influence whether this or that event would happen rather than
any other event. It is impossible to say whether the elements would
enter into this or that connection rather than any other. Such a primor-
dial state, in which everything is possible, and everything is therefore
equally probable, can be called chaos or, to take up a term from thermo-
dynamics and information theory, entropy.
The original substrate is chaos or entropy, the equal probability of all
possible relations. However, the primordial substrate can also be called
complex because everything is possible. We can even say that it contains
absolute complexity. If a distinction creates a system, then this system
will necessarily be less complex than the initial state. This is because the
system necessarily contains fewer elements than the environment. In
addition, the system’s elements are ordered in a certain way. Because of
this, the system contains fewer possible relations among the elements
than in the original substrate. In short, when a system comes into being,
entropy is negated. Therefore, we can speak of negentropy, meaning
complexity has been reduced. Systems always and necessarily reduce
complexity.
No system can realize the logical possibility of connecting
every element to every other one. Systems necessarily reduce
complexity. (Luhmann 1995:44).
In general, a system emerges as a reduction of complexity. Accordingly,
complexity is the reason for the emergence of systems. Complexity
presents itself as a problem to be solved. Because there is complexity,
therefore there are systems. Suppose one wants to understand a system,
i.e., explain the system by tracing it back to its ground. In that case,
whether it is a machine, a living being, or a society, one finds at the
beginning not a causa sui or a prime mover or a God, or even some
primordial substrate, as we have been calling it, but rather a primor-
Basic Concepts of Systems Theory 7
dial problem, namely the problem of complexity. If systems emerge
to reduce complexity, complexity is the problem that all systems, in
one way or another, solve. Therefore, systems are dynamic problem
solvers, and the problem they solve by means of selection, relationing,
and steering is the problem of complexity.18 Theoretically, it would be
helpful to speak not of some primal substrate as we have done above.
Instead, one could say that the environment represents a problem space
in which systemic order arises as various proposed solutions. Solving
a problem means that certain constraints have successfully been met. If
the environment were chaotic, there would be no way of solving any
problem because there would be no constraints in a purely random
and chaotic space. For this reason, as we will see, systems always arise
within an environment that is in some way already ordered. Therefore,
the environment can be conceptualized as a set of constraints, however
unknown and large this set may be. As problem solvers, systems can
be more or less successful. They can solve the problem of complex
environmental constraints well or less well or not at all and disappear.
These considerations lead to ideas of adaptation, viability, and evolu-
tion and connect systems theory intimately to biology. As we shall see,
most models of systemic order are taken from biology, and systems
theoretical discourse is replete with biological metaphors.19
Admittedly, the notion of complexity is itself complex, and before we go
further requires some explanation and clarification. First, it is important
18 Systems are always functional entities, i.e., they “operate”, they “do” some-
thing, and thus exist as “agents” that solve problems. The problem they solve on
the most general level is always: How can complexity be reduced? As Luhmann
(1995:25) puts it: “Establishing and maintaining the dierence between a system
and the environment then becomes the problem because, for each system, the
environment is more complex than the system itself. Systems lack the ‘requisite
variety’ (Ashby’s term) that would enable them to react to every state of the
environment, that is to say, to establish an environment exactly suited to the
system. There is, in other words, no point-by-point correspondence between
system and environment (such a condition would abolish the dierence be-
tween system and environment). For this reason, establishing and maintaining
this dierence despite a dierence in the degree of their relative complexities
becomes the problem.”
19 One should recall that cybernetics and systems theory arose from the engineer-
ing problems of self-regulating machines (anti-aircraft guns steered by radar)
and the purposeful behavior of organisms.
From Systems to Actor-Networks
8
to note that complexity can only be reduced by complexity.20 Selection,
relationing, and steering not only exclude complexity but also create a
complexity that is intrinsic or internal to the system. When one looks at
the organization of elements in a system, the system appears to have its
own complexity:
We will call an interconnected collection of elements “complex”
when, because of immanent constraints in the elements’ connec-
tive capacity, it is no longer possible at any moment to connect
every element with every other element. (Luhmann 1995:24)
What Luhmann here refers to as the “immanent constraints” of the
connective capacity of the elements arises from the fact that a system
always constructs its own elements to be elements of the system. Selec-
tion and relationing transform the original condition of the elements
into system-specific elements. If we think of a table as a system, the
elements of the table are the top and the legs. However, tops and legs
do not lie around in the environment. The system constructs them, and
once something, whether wood, metal, glass, etc., has become a tabletop
or legs of a table, it is transformed, and its connective capacity has been
constrained. We will return to this below.21 What is important to note
here is that selection and relationing at once limit and create complexity.
They allow the system to have only certain states or operations. The more
states a system can have, the more complex it can be said to be. And the
more complex a system is, this was Ashby’s law, the more it can reduce
environmental complexity. Complexity, therefore, can paradoxically
only be reduced by complexity. Because of “immanent constraints” on
the connective capacity of the elements, absolute entropy, as the primal
20 See Ashby’s (1974:299) law of required diversity, “...only diversity can destroy
diversity,” which Luhmann cites in the previous note. For a recent theoretical
formulation of this principle, see Assembly Theory proposed by Lee Cronin and
Sarah Walker (2023).
21 As the reader has probably noticed, the term “construct” and “construction” do
heavy labor in this book. A precise denition is oered by Latour (2013:151.),
where he argues that “construction” should not be understood as “articial” in
the sense that opposed to that which is “real.” We will return to this later when
discussing actor-network theory.
Basic Concepts of Systems Theory 9
substrate characterizes it, is negated.22 Not all connections between
elements are possible. As already noted, one cannot conceive of the
emergence of systems as a reduction of complexity without presup-
posing an environmental complexity, that is, as a set of constraints.
Only under this condition can complexity be understood as a kind of
motivation, the “cause,” or the ground of system formation. Luhmann
asserts that the environment is always more complex than the system
and that, consequently, there is such a thing as “a totality of possible
events,” a “unit of reference that no longer has boundaries” (1970:115).
Luhmann calls this all-encompassing, boundless field of complexity
“world.”23 If we now remember that the environment itself is always
system-relative, i.e., that there is no environment per se, but only the
system’s environment, then we have a third kind of complexity, namely
the complexity of the system-relative environment. The environment is
always the environment for a system relative to a system. In the end, we
are talking about three kinds of complexity: 1) absolute complexity (the
primal substrate, chaos/entropy), 2) the complexity of the environment
relative to a particular system, the environment of the system, and 3)
the internal complexity of the system itself. This complex situation can
be represented as follows:
22 This is the point of departure for Karl Friston’s Free Energy Principle and the
application of Bayesian mechanics to systems theory.
23 On the concept of the world, see Luhmann (1970:115): “The world cannot be
conceived as a system, because it has no ‘outside’ against which it demarcates
itself. If one wanted to think of the world as a system, one would immediately
have to think of an environment of the world, and the concept of the world guid-
ing thinking would be displaced to this environment.” The concept of “world”
implies that complexity, although boundless, is nonetheless not chaos. See for
Luhmann’s concept of the world Günter Thomas (1992). We will return to this
when discussing meaning as world. If meaning is the world, how can meaning
be modeled as a system?
From Systems to Actor-Networks
10
System
Environment 2
Environment 1
Figure 1.1
With respect to the first two forms of complexity, one could distinguish
between “Environment-1” and “Environment-2.” Environment-1 is the
absolute complexity, which denotes the environment of all possible
systems. As mentioned above, one could also consider it the problem
of why anything should exist rather than nothing. Environment-2, on
the other hand, would be the respective system-relative environment.
This is what van Jakob von Uexküll referred to as Umwelt and which
later became important in ecology and cognitive science. It is the world
whose constraints a system must adapt to in order to exist. It is a
problem that the system must solve. In biology, it can be called a “niche.”
Regarding meaning systems, as we shall see, the environment could
be called a “cultural niche,” which differs from biological or ecological
environments not only in that it is constructed (see Niche Construction
Theory24) but above all because it is included within the meaning system
and not outside. We will return to this unique notion of environment
that applies only to systems of meaning later. If one wants to use the
idea of the “world” in this context, it is recommended that “world”
refer to the system/environment difference specific to meaning systems.
Metaphorically, of course, one could say that every organism has its
own “world.” The pond is the world of the frog that lives in it. The frog
24 For an overview of the literature on NCT, see hps://www.oxfordbibliographies.
com/display/document/obo-9780199941728/obo-9780199941728-0089.xml.
Basic Concepts of Systems Theory 11
has its own world, just as does any system capable of cognition. We will
return to the distinction between environment-1 and environment-2
below in the discussion of the concept of meaning. We will see that the
idea of two forms of the environment plays a significant role not only in
living systems but also in human-level systems theory.
For the moment, it is important to note that the environment for the system
cannot be absolutely entropic but must be thought of as always already
ordered or reduced in some way.25 The absolute complexity of a chaotic
primordial substrate, i.e., a boundless entropy containing all possibilities,
is not found in the real world. Moreover, it is also something in which no
system could survive. It is a limit concept like the materia prima of ancient
philosophy. It is only conceivable as the other of order, form, and struc-
ture, but in itself, it has no determinations or constraints. The necessity to
talk about the indeterminate and to think of it as the difference between
the determinate and determinable creates a certain fuzziness and ambi-
guity in the use of the concept of the environment. Environment, then, in
the proper sense, is always already co-ordered by its own nature as well
as systemic processes of selection, relationing, and steering:
Every self-referential system has only the environmental contact
it makes possible and no environment ‘in itself.’ But this ‘itself
makes possible’ is not possible in a structureless, arbitrary, and
chaotic environment because within such an environment, it is
impossible to carry out “internally” satisfactory proof of worth
and, from the perspective of evolution, to acquire permanence.
(Luhmann 1995:101)
Systems, then, emerge within an environment that is always already
ordered in a certain way and, in this sense, “complex.” The (pre-)
ordered environment sets constraints or conditions of possibility for
a system insofar as the elements available to the system are already
25 Mathematically, the relations between the system and environment can be mod-
eled as “Markov blankets,” which statistically relate a system’s external and in-
ternal states. There must be some regularity in the environment for a system to
be able to compare the model with reality and act in order to reduce dierences.
See Friston’s (2010) free energy principle.
From Systems to Actor-Networks
12
limited in what they are, what they can become, and their capacity to
connect. Luhmann (46) states:
The term “immanent limitation” refers to the internal complexity
of the elements that are not available to the system, which at
the same time makes possible their “capacity to unify.” In this
respect, complexity is a self-conditioning state of affairs: in that
the elements must already be complexly constituted in order to
be able to function as a unit for higher levels of system formation,
their linking capacity is also limited, and thereby complexity
reproduces itself as an inescapable given at every higher level of
system formation.
As said above, complexity can only be reduced by complexity. According
to Luhmann, this creates a “complexity gradient” (250) that “stabilizes”
the system/environment difference.
1.5 How do we Recognize Systems? Function and
Functional Analysis
If systems emerge as reductions of complexity, then complexity must
be regarded as a “problem” rather than a cause. Complexity represents
a problem to be solved by system formation. The reason why systems
exist in the first place is to solve a problem. Systems are thus not
self-sufficient substances resting in themselves simply lying around in
the world, but functional entities whose ontological status is dynamic
and processual. They do not simply exist; they operate; they do some-
thing. Systems exist because complexity has been reduced in a certain
way and with respect to a certain form of problem-solving.
Systems theory is a functionalist theory, not a causal explanatory
theory. Systems are always viewed as solutions to problems, not as
states mechanistically caused by other preceding states or what can be
termed external causes. For this reason, cybernetics and systems theory
have always seen themselves as distinct from regular science with its
strict exclusion of circular causality and teleology. When studying
Basic Concepts of Systems Theory 13
linear self-regulating machines and biological systems, it became
apparent that strict causal determinism could not be the last word in
scientific explanation. Complexity is neither a cause nor an effect of
system formation but a problem that different systems try to solve by
transforming greater complexity into lesser, manageable complexity.
Functional analysis does not look for causes but “explains” a problem
by discovering functionally equivalent solutions. For example, the func-
tionalist view of a chair as a possible solution to the problem of sitting
comfortably discovers not primarily how the particular composition of
the parts mechanically enables sitting but looks toward the goal that is
to be achieved by the chair and asks what functionally equivalent solu-
tions to the same problem could be, e.g., stools, cushions, benches, etc.
This aspect of systems theoretical thinking is particularly emphasized
by Luhmann (1995:54):
In this sense, the functional method is finally a comparative one,
and introducing it into reality opens up what lies at hand for a
sidelong glance at other possibilities. In the end, it ascertains
relations among relations: it relates something to a viewpoint
on a problem in order to be able to relate this to other problem
solutions. Accordingly, “functional explanation” can be nothing
other than the ascertainment (in general) and exclusion (in
particular) of functional equivalents.
In systems theory, we are no longer within the framework of the caus-
al-deterministic explanatory model of the natural sciences. And we are
no longer within the world of the classical Western ontology of substance.
The world does not consist of things but of functions. The main questions
are no longer: From which causal processes and preceding conditions do
certain phenomena arise? We are neither no longer primarily concerned
with discovering which events lead to this or that change, nor do we
consider a mathematical model that allows for prediction and control to
be an adequate “explanation.” We do not ask what external conditions
cause system properties to arise. Instead, it is a question of what external
or internal conditions allow or favor certain system states. We are not
interested in how the whole can be analyzed into parts and explained
From Systems to Actor-Networks
14
from the parts. And finally, related to this on certain levels, we need
no longer worry about how cognition corresponds to external reality.
The “fit” or “adaptation” of the inside (system) to the outside (environ-
ment) is a matter of viability and not of truth. The entire epistemological
quandary of how the subject can know the object falls aside. Whatever
information the system might construct, as long as this enables it to
continue its operations, it is “truth” for the system. These questions arise
solely within the classical paradigm of natural science and the Western
ontology of substance and subject. It is well-known and often said that
systems theory has done away with efficient and material causes and
replaced them with final and formal causes.26
According to the paradigm of systems theory, functional analysis
does not start from mechanistic thinking but from analyzing dynamic
systems as ongoing solutions to problems. This has the consequence,
however, that on the most general or “metaphysical” level, it must be
26 See Stadler/Kruse (1992:135-38) for a comparison of the systems theory paradigm
with that of classical natural science. Stadler and Kruse list ve assumptions of the
classical model: 1) the equivalence assumption (causa aequat eectum), namely, “that
in causal processes, to which all processes of nature are aributed, the eect is al-
ways equal to the cause, both quantitatively and qualitatively”; 2) the continuity
assumption (natura non facit saltus), which “disregard[s] the possibility of erratic
changes in natural processes.” 3) the mechanistic presupposition, according to
which “organismic processes [are] interpreted in terms of the interlocking of the
parts of a machine”; 4) the elementaristic presupposition, according to which “a
complex system [can be] analyzed by breaking it down into its elements and ex-
amining the functioning of these elements. The function of the whole system is
then explained by the composition of the detailed processes in the elements”; and
nally, 5) the realist notion, namely, “that the physical world surrounding us is
actually as it looks to us.” Against these assumptions of the classical paradigm,
Stadler and Kruse set the following results of systems research: 1) “the analysis of
complex dynamical systems shows that in nature very often bifurcations occur at
unstable equilibrium points, where minimal causes can produce greatest eects”;
2) in many dynamical systems “phase transitions occur which generate entirely
new order structures at a higher level of analysis”; 3) against the mechanistic as-
sumption, systems theory knows that “the components of dynamical systems [in-
teract] with themselves and, given initial and boundary conditions, produce auton-
omous states of order without being externally imposed or necessarily achieved by
a mechanism”; against the elementaristic assumption, “elementary microprocesses
do not run independently of each other, but cooperate with each other and generate
new qualities on a macroscopic level, which cannot be explained by the individual
elementary processes”; and nally 5) “the realist basic assumption...is opposed by
a constructivist one. Cognitive systems do not absorb (semantic) information from
their environment but generate it internally within the system itself.”
Basic Concepts of Systems Theory 15
assumed that complexity is somehow “perceived” as a problem; other-
wise, “God” would not have created the world from the original chaos.
In most creation myths, God or the gods are always trying to solve a
problem. Why is there something rather than nothing? If mythical or
theological talk is unwanted, we can talk about “self-organization” and
“emergence” to explain the appearance of order from disorder.
1.6 What are Systems Made Of? Order, Organization, and
Structure
Suppose complexity is reduced, and a system emerges by putting a
smaller set of elements than originally given into an order, i.e., into certain
relations to each other. In that case, this order can be considered the orga-
nization or structure of the system.27 The organization of the system makes
the system as a composite whole more – and at the same time less – than
the sum of the parts. It is more because the system exhibits behavior
that is not contained in any of the parts or the sum thereof, and it is less
because the functions of the system do not correlate with the quantity
of the parts. A system can have many parts but few functions. It is the
system’s organization and not any physical property, such as a shell, an
27 The concept of “organized complexity” goes back to Bertalany (1956). In con-
trast to the classical mechanistic view of causality as a linear cause-eect relation-
ship, Bertalany was one of the rst to develop the concept of system as a term
for dynamic wholes, such as organisms, which cannot be grasped within the
framework of classical physics. See the discussion in Kneer/Nassehi (1993:21):
“Organized complexity exists when individual phenomena are not simply logi-
cally coupled with each other in a linear fashion, but when there are interactions
between them. If this is the case, only the description of these reciprocal net-
working conditions is able to convey a picture of the unity of the sum of those
individual phenomena. Disorganized complexity can thus be described as a lin-
ear concatenation of individual phenomena: From A follows B, from B follows C,
etc. Organized complexity, on the other hand, cannot be represented according
to this simple model because it is conceivable that A and B are mutually the con-
ditions of their possibility: So then neither A follows from B, nor vice versa, rath-
er A and B are given to each other by their reciprocity, by their relation which
cannot be represented linearly. It should have become clear that the essential
object of systems theory is the organizational form of the complex interrelation
between individual elements.” It is out of the complex interrelations of elements
that behaviors appear that can be called “emergent.” Furthermore, it can be ar-
gued, as Levin (2022) has done, that living systems are organized not only by
genes but by a morphogenetic form or paern typical of a species.
From Systems to Actor-Networks
16
enclosure, a membrane, skin, walls, boundary posts, etc., that demar-
cates the system from the environment and thus maintains the system/
environment difference. Therefore, it is misleading to say that the system
consists of thing-like conceived elements. The “elements” of a system are
not things or substances lying around in the environment waiting to be
taken up into a system. The system does not consist of things but of its
organization, the specific processes of selection, relationing, and steering
that are responsible for “constructing” the elements as system elements
in the first place. What there is in the elements that are not the product
of selection, relationing, and steering – as we will see in a moment – and
what is not controlled by the organization of the system does not belong
to the system but to the environment. We will explain the notion of orga-
nization further when discussing the idea of code.
1.7 What Do Systems Do? Operation and Process
The organization of the system determines how the system operates,
i.e., what processes the system performs. The organization should not
be understood as a static structure. A system is not only composed of
elements in certain relations to one another but also of operations or
processes that depend on how the elements interact with each other.
Operations or processes are what the system does, and all systems do
something.28 Since processes are always system specific, we turn to a
concrete example to illustrate the preceding definitions.
1.8 The Mechanistic Model – The System as Machine
To illustrate the basic concepts of systems theory that we have presented
so far and also the concepts that follow, let us look at some examples.
28 It is based on this general characteristic of systems that the many methods and
models of the vast eld called “system dynamics” has emerged, which aempts
to identify the elements of a system and model, often mathematically or compu-
tationally, how the elements interact and thus to be able to predict and control
system processes. See for an overview: Wikipedia hps://en.wikipedia.org/wiki/
System_dynamics#Overview.
Basic Concepts of Systems Theory 17
The system, let us say, is an air conditioner. Although this is a mecha-
nistic model, a machine designed and built by humans, it nevertheless
provides the opportunity to present properties that apply to all systems,
whether natural, biological, or social.29
For simplicity, let us assume that the air conditioner consists of only
three components. There is a thermostat that registers the room tempera-
ture, a cooling system that produces cool air and blows it into the room,
and finally a heater that produces heat. These three parts make up a
system, not merely because they are distinct parts but because they are
put together in a certain way and operate in certain ways. In fact, they
could be far apart from one another. For example, the thermostat could
on the wall in a room of the house, the cooling unit on the roof, and the
heater in the basement. The parts make up a system because they are
related to each other in specific ways that allow certain operations.
1.9 Code and Information
With this simple mechanistic model in mind, we now introduce other
basic concepts of systems theory. The assignment of the parts of a system
to each other, i.e., the particular kind of relations they enter into among
themselves, constitutes the dynamic structure or organization of the
system. These relations have a selecting effect. In other words, the orga-
29 The following example is not presented by chance as a cybernetic system. The
machine was the basis of cybernetics. Wiener spoke of cybernetics as the study
of control in the machine and the organism. Ashby’s inuential book, Design for a
Brain (1952), emphasized the continuity between organisms and machines, while
today, terms such as “cognitive architecture” or “articial life” clearly associate
machines and intelligence as maers for engineering and design. According to
Flechtner (1970:10), the model of the machine is equally applicable to other sys-
tems: “If we now try to give a label to cybernetics itself in anticipation, we can
start from the concept of the system. Certainly, machines and organisms, hu-
man beings and their communities can also be regarded as systems, as organized
wholes whose parts are interrelated in manifold ways. And which themselves,
as a whole, are interwoven with other systems and generally with their envi-
ronment in numerous relations. Abstractly spoken, they are structured systems,
which behave in some way to other systems, thus are not rigid but dynamic sys-
tems. Thus, we arrive at the preliminary determination: cybernetics is the gener-
al, formal science of the structure, relations, and behavior of dynamic systems.”
From Systems to Actor-Networks
18
nization selects certain elements from all possible system components,
relates them to one another in certain ways, and finally controls or steers
the operations of the system. An organizing principle that simultane-
ously selects, relates, and steers is what we would like to call a code. Since
every system consists of elements that stand to one another in certain
relations by means of which the processes of the system are enabled and
controlled, it may be said that the principle of organization or structure
of every system is a code. Since the term code is usually associated with
signs and symbols, we must explain how we are suggesting that the
term be used within the scope of general systems theory.
Although the concept of code generally refers to rules, laws, and fixed
procedures that can be followed to translate signs or symbols or program
a computer, not all codes are related to signs. There are, in addition to
semiotic codes, which, as we shall see below, play an important role in
the organization of systems of meaning, genetic codes, which organize
biological systems, and physical codes, which, as in the case of a machine,
organize physical systems, for example, our air conditioner. The concept
of code, then, does not imply the concept of sign, as some semioticians
(Sebeok, 1991; Deely, 1990) assume. However, it does imply the concept
of “information.” Generally, a code, in the sense that we use it as a
concept of general systems theory, is a rule for translating information
into action. A code can be many things, but in some way, common to all
various meanings is the idea that a code contains a set of instructions or
rules that allow for the “translation” of certain things or processes into
other things or processes. In the most general sense, the code, as the
organizing principle of a system, “translates” information into functions
and operations. The tendency to conceive of all system operations as
“semiosis” and thus to speak of sign use and communication in phys-
ical and organic nature is based on the continuities between systems. It
overlooks the discontinuities generated by the emergence of different
levels of order. On the level of meaning, where semiotic coding is the
principle of order, a sign is not a signal. A sign represents something
“as” this or that. A sign must be understood. It is not merely a cause or
a trigger of some event, as a signal can be. If a code performs differential
relations of signs through syntactic, semantic, and pragmatic selections,
Basic Concepts of Systems Theory 19
relations, and controls, we are dealing with a specifically semiotic code.
We are dealing with a semiotic code only if the system’s operations are
constitutive of meaning. When semiotic codes organize systems, they are
meaning systems rather than physical or biological systems. This does not
mean that “cognition,” “problem-solving,” or “modeling” cannot occur
in biological systems. In fact, as we shall see, some of these concepts have
their natural home on the biological level of emergent order and lead
to problems when applied to the level of meaning. It should be noted
that general systems theory commits meaning systems to “understand”
everything as a system, including itself. Whatever kinds of systems and
codes there may be other than semiotic coding is a question that is asked
and answered only within semiotic coding.
In our example of the air conditioner, the code, i.e., the blueprint of
the machine based on the laws of nature, relates certain parts made
of certain materials and associates them with certain other parts in a
certain way. This is selection and relationing. What about steering, the
third necessary component of systemic order? What does the system do?
How does the system operate? First, not only are certain parts, heating
and cooling units, and thermostats selected and related, but also certain
events in the environment are selected and assigned to specific values
of the thermostat. Let us assume that there are only three values on
the thermostat: “too hot,” “too cold,” “comfortable.” The thermostat is
built in such a way that out of all possible events in the environment,
only these three make up the system-relevant environment and can
interact with the system. Nothing else that happens in the environ-
ment, how many people are in the room, what they are talking about,
whether music is playing, etc., is relevant or can become information
for the system. The thermostat is one element in a system related to
other system components, i.e., to the heating and cooling units, but it
operates based on information.
The three values of the thermostat are translated or assigned to
specific operations in the system, namely, turning on/off the cooling
and heating systems or doing nothing. This is steering or control. As
Wiener said, control via “communication,” or information transfer, is
From Systems to Actor-Networks
20
the basis of cybernetics. The thermostat or whatever plays the same
role in any machine, such as the flywheel of a steam engine, is called
the “governor.” The system solves the problem of an environment that
has three degrees of complexity, too hot, too cold, and comfortable. It
reduces complexity in terms of environmental events, elements inte-
grated into the system, and relations among elements in such a way
that certain operations are enabled, and by means of these operations,
certain goals, and set points, can be attained. The system is self-regu-
lating but not necessarily “adaptive.”30
As a reduction of complexity, the code acts based on the system/
environment difference and constructs the system and its specific
environmental references. By considering system organization as code,
the environment becomes structured, partitioned, and differentiated.
This is to say, it becomes transformed into potential information. As
mentioned above, there is still an absolute environment (environ-
ment-1), which is the entire world in which the house containing the
air conditioner stands. However, it is the system-relevant environment
(environment-2), which is the three values that the thermostat can
register.31 The code selects certain environmental events as potential
information and constructs a system-specific environment, i.e., the
environment for the system. Everything else that happens in the envi-
ronment of the air conditioner, except for the three temperature events,
is not “perceived” by the system as information but as noise. The noise
in the room, whether people are present or not, whether it is winter or
summer outside, whether the room is a schoolroom, a living room, or a
government building; all this and infinitely(!) more is irrelevant to the
air conditioning system and cannot become information for it or, what
amounts to the same, cannot be “observed” by it. It can, of course, affect
the air condition and disturb it.
30 “Complex adaptive systems” are usually understood from the biological model
of operational and informational closure, autopoiesis, reproduction, learning,
and evolution. These are not usually characteristics of machines.
31 Speaking of the environment as “relevant” and not merely “relative” to the sys-
tem emphasizes that the environment does not consist of things or events but of
information or noise.
Basic Concepts of Systems Theory 21
What does it mean to say that only certain events in the environment
become information for the system? The system/environment difference
is to be thought of as a code in that only three environmental events
have informational value, i.e., they are differences that make differences in
the system. Information, according to Bateson (1979:99): “is a difference
that makes a difference.” This is illustrated by the fact that the infor-
mation “too hot” does not “mean” or represent something intrinsically
present in the world, but it refers to the difference between values on
the scale of the thermostat. Information is not a thing but a relation, or
as Bateson puts it, a difference, and differences are not things:
The difference, being of the nature of the relationship, is not
located in time or in space. We say that the white spot is “there,”
“in the middle of the blackboard,” but the difference between
the spot and the blackboard is not “there.” It is not in the spot; it
is not in the blackboard; it is not in the space between the board
and the chalk. I could perhaps lift the chalk off the board and
send it to Australia, but the difference would not be destroyed
or even shifted because the difference does not have location.
(Bateson 1979:98)
There are, of course, innumerable differences on a temperature scale,
but only those are information that are selected as relevant to the
system by its constitutive code. Information, then, does not exist in and
of itself. Information does not lie around in the world until some system
stumbles upon it. Information is constructed by the system. The code
constructs information by constructing differences which then change
the state of the system and select operations. Again: for the system, the
world does not consist of things but of differences, which the system
constructs. This fundamental principle of all systems is the basis of
what has come to be known as constructivism and 2nd order cybernetics,
about which more will be said later.
Another important property of the code, apart from constructing the
relevant environment, is constructing the system-specific elements. The
code draws the system/environment boundary in such a way that the
From Systems to Actor-Networks
22
elements are not, as it were, first found lying around in the environ-
ment outside the system and then somehow gathered up and brought
into the system. Coolers and heaters do not exist in the environment;
they are constructed by the system with regard to solving a specific
problem. The system itself constructs the elements of which the system
consists. This means that the concept of elements, components, parts,
etc., must also be understood in a system-relative way. There is not a set
of existing elements, say atoms, pieces of wood, stones, etc. which are
first discovered in the environment and then somehow taken up and
made into building blocks of different systems. The elements become
only elements of the system by the system. This is why “translation” can
be used in this context and why the principle of system organization
can be called a code. It also implies that there are different elements
at different system levels; atoms, cells, organs, and signs are all to be
considered as elements of certain systems on system levels, each coded
– and that means constructed – in their specific ways. Organisms are
made up of different elements than machines, and semiotic systems
are, in turn, made up of different elements than organisms. A system
of meaning does not consist of cells, tissue, skin, organs, neurons,
metabolic processes, etc.32 On each level of emergent order, the code
“translates” elements into functions in different ways. This means
elements, like relations, are not optically or substantively given but are
constructed by the system through its specific code. We thus are left
with a completely relativistic and functionalist notion of element:
Then an element would be what functions for a system as a
unity that cannot be further dissolved (even if, viewed micro-
scopically, it is a highly complex compound). When one says
“cannot be further dissolved,” this also means it can constitute
and change itself only by interrelating its elements, not by
dissolving and reorganizing them. ... Elements are elements
only for the system that employs them as a unit, and they are
such only through this system. (Luhmann 1995:22)33
32 It is a commonplace error to think that neurons in the brain construct meaning or
that signs are elements of the brain along with neurons.
33 This means that systems analysis has to pay aention to the level of emergent
Basic Concepts of Systems Theory 23
Let us take another example to illustrate this general principle of
systems theory. Consider, for example, a table as a system. The table, let
us say, consists of a top and legs. The top and the legs are the necessary
elements of the table system. However, the function alone determines
what can serve as a tabletop and what as legs. Table tops can be made
of any material that is solid enough to support weight, and the legs
can be made of wood, steel, glass, plastic, etc., so long as they function
to support the top. Even a person can serve as a table if they go on all
fours and hold that position long enough for someone to use their back
as a table. The elements of the table system are not just lying around
in the environment until the code of the table stumbles over them and
integrates them into itself. Instead, the code selects certain elements
that can serve the functions that need to be fulfilled to create a working
space at mid-body height.
The same is true of living organisms. Let us say the elements of an
animal are the organs, the skin, tissue, the bones, etc. The organism does
not find these things just lying about somewhere in the environment
and then integrates them into itself, but it constructs them according
to the genetic and morphogenetic code that organizes it. And finally,
a language does not consist of signs that can be found in the environ-
ment, but out of all the possible sounds that the human voice can utter,
only certain are selected to bear meaning. The same, of course, is true
of non-verbal signs. Therefore, along with the principle that all systems
are constructed by processes of selection, relationing, and steering that
create a system/environment difference, we can say that it is a second
principle of general systems theory that all systems construct their own
order on which it operates. If one disregards level dierences, e.g., by analyzing
communicative actions neurophysiologically, then one abolishes the system/en-
vironment dierence, i.e., the code that is constitutive for the system, and sub-
ordinates the system to an alien code, where one then can no longer compare it
with the system’s own elements, e.g., signs, but with elements foreign to the sys-
tem, e.g., nerve impulses. For this reason, it has become commonplace to speak
of a “hard problem” of consciousness. One cannot nd meaning and experience
in the brain because meaning is a higher level of emergent order than life. Mean-
ing constructs meaning, and life constructs cells, organs, metabolic processes,
etc. The entire materialism/idealism controversy that has plagued Western phi-
losophy for ages and dominates cognitive science until today could be avoided
by carefully distinguishing levels of emergent order.
From Systems to Actor-Networks
24
elements by means of a code that “translates” elements into functions
and operations.34
1.10 Cybernetics and Feedback Loops
The operations of a system are outputs that have effects on the envi-
ronment. For example, our air conditioner turns on the cooling system
or the heating system, which changes the temperature in the room so
that “comfortable” again registers on the thermostat. From the system’s
point of view, it does not “know” what is happening out there in the
environment. It only “knows” that the values on the thermostat are
changing. According to the code, which is its ordering principle, the
system operates until the value “comfortable” is registered again, i.e.,
until new information triggers new processes.
This form of operation is called a feedback loop, a control loop, or
circular causality. The output of the system into the environment is
received back into the system as input. Outputs become inputs. This
is the loop. It is feedback because it “informs” the system of the effects
of its outputs, allowing the system to control its operations such that
the goals or setpoints it has can be attained. Thus, we can speak of a
cybernetic or self-regulating system. Although all systems are self-reg-
ulating, the kind of self-regulation we are talking about depends on
the code constituting the system because output and input are linked
according to the possibilities of the code. The system could cause many
things in the environment to happen through its operations. The air
conditioner could produce many things, such as noise, stench, high
costs, etc., but because of the code constituting it, these are all neither
relevant nor “perceptible” to the system, i.e., they have no informa-
34 This does not imply that there are no constraints on what elements can be trans-
lated into what functions. The laws of nature and characteristics of maer and en-
ergy are constraints for living systems, just as the laws governing living systems
constrain meaning. Despite constraints on the part of the substrate, higher-level
codes contextualize and constrain systems on lower levels. Life, for example, can
do things with maer that maer could not do on its own, and meaning can do
things with maer and life that neither could do on their own. A case in point is
articial intelligence and, of course, all technologies.
Basic Concepts of Systems Theory 25
tional value; they do not become inputs or feedback, unless, of course,
they are so selected.
1.11 Purpose, Teleonomy, and Homeostasis
Because the system has a goal, setpoint, and some information that
when it is registered does not trigger an operation, i.e., it does nothing;
it looks to an observer as if all the system’s operations are aimed at
achieving this one state over and over again. The system aims at
stability or equilibrium. As long as nothing triggers some operation,
the system does nothing. This value is then called the setpoint or target
value because the system appears to act as if this value is to be main-
tained. Whenever relevant values are registered, e.g., “too hot” or “too
cold,” the system operates until it registers its target value “comfort-
able” again. For this reason, it can be said that cybernetic systems are
goal-directed or teleonomic systems. The target value guides (controls!)
the operations of the system and thus appears as a “purpose,” a “goal.”
Moreover, because system operations are undertaken only to bring the
system back to the same point, i.e., to (maintain) a stable equilibrium,
such systems are called homeostatic systems or systems that operate to
maintain a specific constant state.
This is not only true for cybernetic machines like the air conditioner.
Organisms are the prime example of homeostatic systems. They act
to maintain their equilibrium or, as we will say later, their “autopoi-
esis,” for example, a certain oxygen level in the blood or a certain body
temperature. In the case of living systems, the set points are not set by
humans according to a mechanical code, but they are set by the genetic
code and by interactions with the environment. However, it must be
noted that organic systems never do nothing. They are constantly oper-
ating to ensure their further operations. This is called “autopoiesis,”
which we will discuss later. Furthermore, living systems strive for
stability and maintenance of certain parameters, but they also generate
instability and contingency through mutations and interactions with
the environment. Our air conditioner does not do this. There is some-
From Systems to Actor-Networks
26
thing about life that does not strive for stability, order, regularity, and
predictability. This is, as we shall see, also true of meaning systems. To
the extent that human beings and societies may also appear to maintain
stability, they can be modeled as homeostatic systems, that is, systems
that operate to maintain given parameters.35 According to the theory
of social systems that Luhmann proposes, social subsystems such as
economics, politics, jurisprudence, education, business, etc. are guided
by autopoietic, operational, and informational closure. The question of
whether the respective codes of social systems operate homeostatically
is a matter of interpretation. Do human societies strive for stability or
change? Can history be understood as social and cultural evolution? We
will return to these questions later.
1.12 FunctionalDierentiation,Subsystems,andInternal
Complexity
Up to now, we have mainly dealt with the relationship between the
system and the environment. If we now look at the system alone without
reference to the environment, it becomes apparent that it is functionally
differentiated. The system has different elements that perform various
operations with different functions. In the case of our air conditioner,
one function is to cool, the other is to heat, and the third is to register
the temperature. For these three functions, the system has different
subsystems, namely, the cooler, the heater, and the thermostat. The
system, therefore, is not simple. It has an internal complexity that has
come about through a process of differentiation. According to Luhmann
(1995:53), differentiation is “the establishment of new system/environ-
ment differences within the original system.”
The cooling system, the heating system, and the thermostat can be
said to be subsystems within the overall system of the air conditioner.
Similarly, the organs and functions within a living system that are
essential to its viability can be seen as arising from the differentiation of
35 The mathematical model of such systems is what the “free energy principle”
proposes. See Friston (2010).
Basic Concepts of Systems Theory 27
subsystems.36 The idea of a subsystem implies not only that it is part of a
larger system but that it is a system itself and not merely an element. We
would not say that a table’s top and legs are subsystems, but a cooler
and a heater are. Nonetheless, the legs and top are distinct functions
and, from this perspective, constitute the internal complexity of the
table. To the extent that internal complexity depends on the possible
states of the system, and these, in turn, depend on the functional subsys-
tems it has, internal complexity is not merely a matter of the number of
elements in a system.37 Furthermore, as in the case of the table, it is
often unclear to what extent a subsystem constitutes a function within a
system or whether functional elements that do not seem to fulfill all the
requirements of a system still contribute to the internal complexity of a
system. Finally, following Ashby’s law that only complexity can reduce
complexity and given the tendency of systems to attempt to reduce
complexity, there looms on the horizon of systems theory the problem
that, at a certain point, systems will become too complex to maintain
their coherence and disintegrate into the environment or another possi-
bility is that a higher level of order will emerge.
1.13 Contingency and Degrees of Freedom
Because the system has internal complexity, it has different possibilities
for operation or action. In the case of our air conditioner, it can either
cool or heat. This means that in a certain sense, the operation of the
system is not determined, i.e., it does not have only one operation. If it
can have different operations, there is contingency. Contingency means
“it could be different.” For example, the system could act this way or
that way. If we don’t know what the thermostat is currently registering,
and if we don’t know how the system is constructed – that is, if it were
a “black box” – then we don’t know what the system will do. It could
36 Miller/Miller (1990) identies twenty dierent subsystems necessary for life.
37 Klaus and Liebscher Wörterbuch der Kybernetik dene complexity as follows:
“The complexity K of a system is directly proportional to the number n of its
elements, the number z of the possible states of these elements, and the number
k of couplings between the elements”(1979:314).
From Systems to Actor-Networks
28
be heating or cooling. The term contingency denotes this possibility
or indeterminacy. Contingency means that the system is not mecha-
nistically determined such that only one course of action, a “reaction,”
is necessary. If we have a system whose operations are not causally
determined, as they are for our air conditioner, these actions can be
said to be “free.” If there are two options, one can say the system has
two degrees of freedom.38 The more internal complexity a system has, the
more degrees of freedom it has. And since there is a direct correlation
between information and operation, the more information a system can
construct, the more degrees of freedom it will have. Only systems that
have internal complexity can exhibit contingency and can be said to
have degrees of freedom. In the case of our air conditioner, we would
have to assume that it could react to a drop in temperature in different
ways and that the “decision” of which way to solve the problem is left
to it and not the environment.39
Contingency is, in many respects, just another word for complexity.
Contingency, for example, also represents a problem for the system;
the problem of “deciding” what it should do. This amounts to reducing
contingency, just like the emergence of the system in the first place
can be thought of as reducing complexity. When the system “chooses”
which course of action it will take, it has reduced contingency. A deci-
sion can be defined as uncertainty reduction. Out of the different possi-
bilities, only one is selected. Before the decision one didn’t know what
the system would do. Afterwards, the situation is no longer uncertain.
We can distinguish between system contingency and environmental
contingency. The environment consists of a multitude of possible events
that could happen. Taking our air conditioner again as an example, it
could get too hot or too cold, or it could stay comfortable, and of course,
many other events that are irrelevant to the system could happen.
38 Mechanical or physical systems of course also have degrees of freedom, but the
term refers to the number of variables or independent parameters which deter-
mine states of the system. See Wikipedia for an overview hps://en.wikipedia.
org/wiki/Degrees_of_freedom_(mechanics).
39 Levin (2022) denes “intelligence” as the ability to solve a problem in more than
one way and nds this ability throughout the living world from the single cell to
human brains. For Levin life as such is intelligent.
Basic Concepts of Systems Theory 29
Depending on what happens in the environment, the system reacts.
But the system does not “know” in advance what will happen. From
the system’s point of view, the behavior of the environment is unde-
termined and, in this sense, uncertain and contingent. Some systems
have no internal complexity and, consequently, no contingency. They
are completely determined and “simple.”40 The environment, however,
is always complex and contingent, and the primal substrate we postu-
lated at the beginning of the world is at once absolutely complex and
absolutely contingent.
1.14 Indeterminate and Determinate Complexity/
Contingency
As noted above, there are three kinds of complexity and contin-
gency. Two of them concern the environment: 1) absolute or (world)
complexity or contingency, i.e., anything at all can happen, everything
is equally probable, and there is no reason why there is something
instead of nothing, and 2) complexity/contingency of the system’s envi-
ronment, i.e., not everything can happen, and only certain events that
do regularly happen are considered relevant for the system. Let us call
these two forms of complexity indeterminate complexity/contingency
and determinate complexity/contingency. For the air conditioner, only
three values exist as possible or perceptible events in the environment:
too hot, too cool, and comfortable. These three states represent the
determinate complexity/contingency of the environment. Determinate
complexity forms the horizon of possible events, the “perceptual field”
of the system, or everything that can become information for the system.
But there are other events in the environment than just these three. For
example, an earthquake could destroy the building, or it could burn
down, or it could be deliberately demolished, or it could be noisy in
the room. Many other things could happen, and all of these would not
“mean” anything to the air conditioner because its constitutive code
has not determined such possibilities to be possibilities for it. The
40 Heinz von Foerster (1992) called such simple systems “trivial machines.”
From Systems to Actor-Networks
30
code makes certain events in the environment relevant to the system
and leaves everything else undetermined. It constructs a system, i.e.,
creates a system/environment difference, by reducing the indetermi-
nate complexity of the world to determinate complexity/contingency
of the world for the system. Every system has its environment, and no
system faces the task of selection, relationing, and steering over against
absolutely indeterminate complexity. Determinate complexity consti-
tutes the constraints that condition system organization. The universe
is always already ordered in some way. And so, it makes sense to speak
of “natural selection.”
It should be kept in mind, that determinate complexity/contingency
depends directly on the internal complexity of the system. The code,
as said, selects only certain events from all possible events as relevant
to the system. The system thus constructs its environment, which
consists of the possibilities that the code of the system has determined
to be possibilities for it. This is not to say that the system can wish for
any environment. The environment contains many, perhaps infinite,
constraints that are not at any time all relevant for the autopoiesis of
the system. These are ignored, or if relevant constraints are ignored, the
system disappears. For all such events, the system is “blind” because it
has no differentiated internal subsystems or operations corresponding
to them. If a system has no internal complexity, its environment can
also have no determinate complexity. In such a case, there is only one
relevant information and only one operation corresponding to it. One
cannot even speak of “information” in this case but “cause” or “trigger.”
Thus, the determinate complexity/contingency of the system’s environ-
ment is related to its internal complexity/contingency, i.e., to its func-
tional differentiation. The more internal system complexity, the more
determinate environment complexity. The more complex a system is,
the bigger its world. We will return to these ideas when discussing the
conditions under which life and meaning can emerge.
Basic Concepts of Systems Theory 31
1.15 Autonomy, Learning, Emergence, Levels of Emergent
Order, Evolution, Self-Organization
Imagine a much more complex system than our simple air conditioner.
Suppose the system has not only the two subsystems with operations
of cooling and heating available in its repertoire but also other possi-
bilities, e.g., instead of turning on the cooler when it is too hot and the
heater when it gets too cold, the system could be set up to open and
close the windows depending on the outside temperature. Or one could
imagine that an alarm would go off when it gets too hot, alerting people
of the danger or automatically contacting the fire department. All of
these and other conceivable possibilities are functional equivalents that
become possible because of the functional differentiation in the system,
i.e., greater system complexity.
Let us further assume that the system can “decide” which functionally
equivalent operations it wants to perform. Depending on the situation
in the environment and on the internal state of the system, it could
“choose” this or that operation. One time to open the windows, another
time to send people out, a third time to turn on the cooling, and so
on. The system has sufficient autonomy to solve a problem in different
ways.41 Let us further assume that the system is capable of learning,
based on experience, which of these functionally equivalent operations
works best in each case. We could also assume that the system is capable
of a certain planning. It can predict probable environmental states and
adjust its operations to meet future needs. Furthermore, we may assume
that the system is self-productive and can maintain itself and produce its
parts, e.g., by being connected to an automated factory that produces
spare parts and other similar systems according to instructions from the
system. And finally, let us assume that the system can even reorganize
itself such that it can still function when the environment changes in
such a way that its former organization no longer functions; that is, it
can adapt to environmental changes by changing its organization and
41 As mentioned above, Michael Levin (2022), following Maturana and Varela and
William James, calls the ability to solve a problem in more than one way “intelli-
gence” and equates life with cognition or intelligence.
From Systems to Actor-Networks
32
becoming something different than what it was before.
With these assumptions, we notice that the model of a machine and the
basic concepts that apply to it and can be derived from it are no longer
adequate to describe the system. Despite all the continuities and similar-
ities that cybernetics has discovered between self-regulating machines
and organisms, there are still fundamental discontinuities and differ-
ences between mechanisms and life. When a system has many different
behavioral possibilities, the ability to learn from its actions, and the
ability to reproduce itself, anticipate the future, and plan, it is no longer
what is usually understood to be a machine but a living being.42 At a
certain point, the system’s increasing complexity forces us to speak of
an entirely different form of systemic order. The system now becomes a
living organism. There is a need for a specifically biological model instead
of a mechanistic model. This doesn’t change, even if we can also speak
of “artificial” life and “artificial” intelligence in the case of engineered
systems exhibiting the characteristics we have just described but which
have been constructed by humans and are products of “design.” Such
artificially living and intelligent systems are modeled on living systems
and designed to operate similarly to living systems. At what point do
artificial systems become “life,” or are we forced to redefine what “life”
means, or must we create new hybrid categories to account for beings
that heretofore did not exist or that always existed but were ignored?43
42 Von Foerster’s (1992) well-known distinction between trivial and non-trivial
machines could be cited here. The trivial machine is a determined input-output
loop, whereas the non-trivial machine is not determined by inputs but has au-
tonomy with regard to outputs and therefore is unpredictable. The non-trivial
machine is non-linear and historical. Von Foerster and cybernetics in general has
chosen to emphasize the continuities between machine and organism. Michael
Levin (2022) also emphasizes continuity and proposes that there is agency and
intelligence down to the atomic and even subatomic levels as well as unbroken
continuity between machine and organism. This can be admied as an eect of
scale of observation while also arming that emergence, major transformations,
or discontinuities are apparent at larger scales.
43 See for a discussion Deplazes/Huppenbaur (2009); Wikipedia Articial Life
hps://en.wikipedia.org/wiki/Articial_life. Here to the work of Michael Levin
(2022) should be included as challenging the clear distinction between mecha-
nism, life, and mind, and of course, Latour has challenged the traditional dis-
tinction between the articial and the real, things that are constructed by humans
and things that construct themselves.
Basic Concepts of Systems Theory 33
To introduce and define further basic concepts of systems theory, we
now have to jump to a higher level of emergent order and speak of
living systems. Our models necessarily become biological. This jump
is not arbitrary but results from the development of the complexity of
our subject. As we saw, the system we were describing became increas-
ingly complex, and suddenly it reached a point where it “jumped”
to a higher level of order. This jump can be called emergence.44 Since
living systems are not an effect of physical causes, they can be said to
be self-organizing.45 There “emerges” something that cannot be derived
from the elements and operations of systems on the merely physical or
mechanical level of order. Even today, biologists do not know what life
is or how it originated. What is known, however, is that the emergence
of life has something to do with complexity.46 It would seem that when
a certain degree of complexity arises such that existing codes and forms
of organization can no longer reduce it, a jump occurs to a higher level
of emergent order and a more complex form of coding.
The result of such a fundamental transformation can be called the
transition to a higher level of emergent order. This occurs on the basis of
a new kind of coding. DNA is the code of living systems.47 Once life
has emerged as a level of order beyond physics and chemistry, many
new aspects of systemic order appear. Living systems are autopoietic,
self-referential, informationally and operationally closed systems.48
They operate to maintain their operations, their “autopoiesis,” that
is, self-production. They do this within an environment that “selects”
44 Heylighen (1989:382) denes emergence as “a qualitative change, where a new
organization or system appears, with properties (potential appearances) that did
not exist in the old system.”
45 Self-organization is a widely documented phenomenon found in many kinds of
systems on all levels of emergent order. For an overview, see Wikipedia: hps://
en.wikipedia.org/wiki/Self-organization.
46 See, for example, Assembly Theory by Sara Walker and Lee Cronin. hps://
en.wikipedia.org/wiki/Assembly_theory
47 The work of Michael Levin (hps://www.drmichaellevin.org/) has raised the
possibility that electronic morphogenetic coding or bioelectric networks also
play an important role in organizing living systems. Levin uses the metaphor of
hardware and software to explain the relation of DNA to morphogenetic electri-
cal paerns.
48 Maturana/Varela (1980; 1987).
From Systems to Actor-Networks
34
for viability, that is, successful adaptation. Since the environment is
constantly changing and is changeable, contingency comes from the
outside. But the genetic code is subject to more or less random variation
that changes the organization of the system such that unforeseen possi-
bilities of action internal contingency can lead to successful adaptations.
This kind of interaction and relation between system and environment
can be termed evolution. Evolution means that the autopoietic, system-
building reduction of complexity allows for variability that is then
selected by the environment. Evolution implies that the system reduces
internal and external contingency so that autopoiesis can continue. The
environment allows some mutations of the organism to survive and
reproduce. This is called natural selection. Variation and natural selec-
tion are principles that all genetic coding is subject to. They determine
the relationship of the living system to its environment. It should be
emphasized that evolution is not emergence. Evolution is the ongoing
play of variation and selection, allowing the genetic and morphogenetic
code to explore the possibility space of any environment. However, the
emergence of life as such is not a product of evolution. Instead, life is
what makes evolution possible. Only living systems evolve. Life did not
evolve from physical or chemical systems. Life emerged from physical
and chemical complexity as a higher level of order and a fundamentally
different form of coding.
Generally, organic systems have emerged from inorganic systems,
just like meaning systems have emerged from life. The genetic code
organizes systems that are subject to the pressures of variation and
selection. The semiotic code organizes meaning systems that are not
subject to evolutionary pressures and should not be modeled as living
systems. Meaning is not subject to variation and selection and cannot
be said to “adapt” to any environment. For this reason, as we shall
see, it is misleading to speak of cultural or social “evolution” or social
Darwinism. It is, however, commonplace to do precisely this. Despite
all tendencies to the opposite, concepts such as evolution and adapta-
tion have no place in social theory. Considering that everyone, at least
since Spencer, assumes there is such a thing as social evolution and
that evolutionary theory applies to society and culture, to claim the
Basic Concepts of Systems Theory 35
opposite as we do, is so unusual that we will have to argue for these
claims at length later in this book. We will discuss semiotic systems that
represent a higher level of emergent order, and which cannot be under-
stood on the basis of a biological model. For the moment, it is sufficient
to note that the dynamic relations of the system to the environment
are different on each level of emergent order, even though there are
continuities and analogies that permit using certain common concepts.
The claim of general systems theory, which we are critically testing in
this book, is that the basic systems theoretical concepts on one level
can be carried over to the next higher level. Just as one can speak of
self-organization on the level of physical systems, this concept can be
applied to living systems and even meaning systems. Systems theory
offers general concepts like self-organization and emergence that can be
fruitfully applied to understand various phenomena. We will see that
this may be the case for certain systems theoretical concepts, whereas
for others, the systems model breaks down.
With regard to emergence and self-organization, let us look at some
current definitions. Ebeling (1989:17) writes:
By self-organization, we mean an irreversible process that
leads to more complex structures of the overall system through
the cooperative action of subsystems. Self-organization is the
elementary process of evolution, understood as an unlimited
sequence of self-organization processes. In this sense, the
processes on Earth and in the cosmos are generally evolu-
tionary processes that can only be understood in the context of
their history, i.e., the entire chain of causative self-organization
processes.
Closely linked to self-organization is the concept of emergence. Emer-
gence is defined by Krohn/Küppers (1992:389) as follows:
In the ‘classical’ sense, emergence means the emergence of new
layers of being (life vis-à-vis inanimate nature or mind vis-à-vis
life) that are in no way derivable, explicable, or predictable from
the properties of an underlying level. Hence, they are perceived
From Systems to Actor-Networks
36
as ‘unexpected’, ‘surprising’, etc. In a modern version, we speak
of emergence when a new quality arises through microscopic
interaction at a macroscopic level, which is not derivable (caus-
ally explainable, formally inferable) from the properties of the
components, but which nevertheless exists solely in the interac-
tion of the components.
To a certain extent, emergence can be seen as a process of self-organi-
zation. This is because 1) systems show a tendency to increase internal
complexity and 2) from the growing internal complexity of the system,
which conditions increased environmental complexity, the need arises
for a new organizing principle capable of solving the complexity
problem. Only the system itself can solve this problem. The very idea
of emergence implies self-organization since emergent phenomena are
not caused by external factors. At a certain point, chemical or molecular
systems became so complex that this complexity could only be reduced
by a new kind of organizing code, the genetic code. The genetic code is
more complex and variable than the physical code, so it can do things
with matter that matter alone could not do. Life organizes matter in
ways that matter would not be able to do by means of physical codes
alone. As we shall see, the semiotic coding of meaning systems is more
complex and variable than genetic coding and can therefore do things
with life that life alone could not do. This explains how technology and
genetic engineering are possible. It also explains why meaning systems
can produce artificial intelligence on a non-biological substrate. But
more of this later. In a certain sense, it could be said that life is more
“powerful” than matter and meaning more “powerful” than life.49 But
as we shall see, these different forms of order are “powerful” in different
ways. They do not all do the same things in the same ways.
49 For meaning systems, life is cheap. Following Socrates, it can be said that the life
without meaning is not worth living, and history demonstrates how many oer
their lives for their beliefs.
Basic Concepts of Systems Theory 37
1.16 Excursus: The Self-Organizing Universe
As a scientific paradigm and a supposedly universal theory, systems
theory is also a worldview. Its claim to universality has a visionary
character. This can be seen in the following story of the self-organizing
universe.50 The concept of self-organization is comprehensive and
has been used to describe everything from nano-structures up to the
cosmos. With regard to systems theory, self-organization always refers
to a system of some kind, that is, a composite of parts related in specific
ways such that the composite can be distinguished from an environ-
ment and specific tasks can be accomplished. The relations among the
parts are the organization of the system. The concept of self-organiza-
tion implies that this organizing principle and process comes from the
system and not from any external cause. God cannot create a system.
The system is the agent pulling itself up, as it were, by its bootstraps.
In self-organization, there is always some spontaneity, autonomy, auto-
catalytic event, etc. The concept does not claim, however, that nothing
outside the system “triggers,” “occasions,” “influences, or “constrains”
self-organization. It does not claim that something comes from nothing.
The trigger or occasion for self-organization is usually some change
in the environment that often acts as a selector for new structures of
the system, but it can also be some “mutation” in the system itself. In
our story of the self-organizing universe, we are primarily concerned
with the internal complexity of the system – which, of course, is always
related to external complexity – as the driving force behind the self-or-
ganizing transformation of the system.
As we have said, systems arise to solve the problem of complexity. What
is most complex is what we have referred to as the primal substrate or
medium, which is absolutely complex. In the primal substrate, all events
are equally probable. This is the chaos before the creation of the world,
that is before any forms of systemic order come into being. The first
systems at the beginning of the world were purely physical. The prin-
ciple of their organization was the laws of nature that structure physical
50 See Jantsch (1980) for a detailed discussion of self-organization and evolution.
From Systems to Actor-Networks
38
and chemical systems. Classical natural science did not acknowledge
the possibility of self-agency of any kind. Things were passive until
acted upon by external forces. The world did not consist of systems
but of objects, things, and inanimate matter in motion. This changed
with the systems paradigm. Systems theory subscribes to circular and
teleological causality and holism and applies these explanatory models
also in the physical sciences.
Generally, even for physical systems, it could be supposed that to
reduce environmental complexity, the system must build up internal
complexity. In cybernetics, the need to build internal complexity to
respond to external complexity is called Ashby’s Law.51 To cope with
environmental complexity, the system must create internal complexity
through structural differentiation, that is, via self-organization. The
more internal complexity a system has, the more environmental
complexity it can successfully reduce. As in the air conditioner model,
the system could successfully respond to a few different environmental
events thanks to its internal differentiation into a cooler and heater. A
living system can respond to many more external perturbations since
it has greater internal complexity. Systems theory postulates an innate
agency or tendency of systems to increase internal complexity, which
lays the foundation for self-organization and the emergence of ever-
higher levels of order.
Returning to our story, at a certain point in time, chemical systems
became so complex that they could no longer be efficiently organized
by the physical code. As a solution to the problem of over-complexity,
life emerged based on genetic coding. Emergence is a radical form of
self-organization that no longer merely changes structures and adapts
to an environment; this is normal evolution. Emergence changes the
principles that organize the system. Life is a fundamentally different
kind of system than physical systems. No amount of physics is going
to “explain” life. The methods, instruments, and theories successfully
51 Usually stated as “only variety can destroy variety,” Ashby also stated it as fol-
lows: “The larger the variety of actions available to a control system, the larger
the variety of perturbations it is able to compensate.” Ashby (1956).
Basic Concepts of Systems Theory 39
used to observe and explain physical systems cannot describe living
systems. Quantum mechanics will not explain life or consciousness.
Not respecting this discontinuity in levels of emergent order leads to
physicalist reductionism and, on the other side, to panpsychism.52
At a certain point, living organisms became so complex, for example,
with the development of a central nervous system consisting of billions
of neurons and trillions of connections, that only a new form of coding
could reduce this complexity. Meaning or semiotic coding emerged.
Meaning systems are an entirely different form of system than living
systems. Nonetheless, it is a thesis of this book that systems theory, with
few exceptions such as Luhmann, never clearly distinguished between
life and meaning. Indeed, the idea that life is cognition, that cognition
is a process of the brain, and that mind and nature are one has become
a common motif in systems theoretical discourse and cognitive science.
We will return to this problem later. For the moment, let us continue
our story of the self-organizing universe.
The general tendency of systemic development dictates that meaning
systems, that is, societies and cultures, become more and more complex
by processes of internal differentiation. History tells us how origi-
nally simple societies, languages, forms of government, technologies,
beliefs, etc., have become more complex and differentiated over time.
Civilization is a story of progressive expansion of knowledge, art, tech-
nology, etc., which leads to ever more complexity that the system must
somehow control. Presently, some voices claim that society has become
so complex that it can no longer be organized by means of semiotic
coding, and some new form of order must arise to prevent society from
collapsing into chaos.53 Typical symptoms of this precarious situation
are the many global problems of climate change, migration, uncon-
52 The seemingly endless debate between materialism and idealism can be traced
back to this disregard for levels of analysis. For a critique of physicalism see the
work of Bernardo Kastrup (hps://www.bernardokastrup.com/). Levin’s (see
Fields et al 2022) panpsychism is anchored in application of the free energy prin-
ciple to quantum systems.
53 For an overview of the discussion see Wikipedia Complex Society (hps://
en.wikipedia.org/wiki/Complex_society)
From Systems to Actor-Networks
40
trollable financial systems, pandemics, international conflict, etc. The
rise of Earth Systems Science, ecology, and global governance can be
seen as the attempt to model the world as a systemic whole that can,
in principle, be predicted and controlled. Whether world complexity
can be successfully analyzed by any system dynamics models and thus
become accessible to prediction and control is an open question.54 Not
only is it an open question, but it is also a controversial issue because
increased control correlates with decreased freedom. A world ecolog-
ical system that is in every possible way predictable and controllable
seems to resemble the most radical form of totalitarianism.
If it is not possible to mathematically and computationally model the
world, one may suppose that the problem of societal over-complexity
will be solved by a fourth leap to an even higher level of emergent order,
a level of order higher than meaning. Of course, we cannot know what
this would be since whatever we can know is known on the level of
meaning. Complex molecules could not have foreseen life, and complex
living structures could not have foreseen meaning. We are “trapped”
in our semiotic order; we cannot think of any other ordering principle
beyond meaning. Such a possibility is not within the horizon of our
perception. As Wittgenstein said, the limits of one’s language are the
limits of one’s world. We cannot imagine what the next developmental
leap will look like. Every attempt to imagine a transhuman form of being
is done within the realm of meaning. Otherwise, we couldn’t talk about
it. Whatever might appear to solve the problem of the over-complexity
of meaning, whether it be some form of transhumanism, cyborgs,
robots, or AIs, it appears for us as an open horizon of possibility, as the
transcendent par excellence, as the unmastered contingency of human
existence.
The moral of this story is that generally, the development of systemic
order runs in the direction of ever higher system complexity and ever
higher contingency, variability, and transformability. It seems that the
goal is to absorb the complexity of the environment into an all-encom-
passing system that in every way equals environmental complexity.
54 Wofram (A New Kind of Science, hps://www.wolframscience.com/nks/
Basic Concepts of Systems Theory 41
As Ashby predicted, what was outside is reproduced inside. And as
Jantsch (1987:181) says, “Not in the construction of hierarchy, but in the
unfolding of complexity lies the real work of evolution.” Complexity is
not only at the beginning of the world but is apparently also the final
and apocalyptic goal of its development. Systems become increasingly
complex to cope with the absolute complexity of the environment until
they become so complex that they no longer differ from the environ-
ment, and the constituting difference from the environment disappears.
Suppose we grant that internal complexity is formed against the back-
ground and as a response to environmental complexity. In that case,
we must ask if it will come so far that systems become equally complex
as their environment, i.e., that they finally and definitively solve the
problem of complexity. The system would be exactly as complex as the
environment. Indeed, this is the stated goal of datafication or applying
descriptive, predictive, preventive, and prescriptive analytics to big data
and modeling reality by so-called “digital twins.” Can this be done for
small subsystems such as a machine, a factory, a city, the human body,
etc., and finally for the Earth System? If so, there would be no difference
between the system and the environment. System complexity would
equal environment complexity. But if there were no difference between
system and environment, there could be no system because systems are
constituted necessarily by a difference to the environment. Is the goal of
evolution its dissolution, the destruction of all systems, and the return
to absolute entropy? Suppose evolution aims at returning to the state
from which it started, namely the equal probability of all events. In that
case, we have a story that is like physical theories of the “big bounce,”55
namely, the universe comes into being, expands, and then collapses
only to come again into being and so on infinitely. Accordingly, systems
theory is open to evolutionary as well as devolutionary interpretations.
We do not know whether we are going up, down, forward, or back-
ward. The entire story can signify the fulfillment of all possibilities and
a senseless cycle. There seems to be no reason internal to the theory
for optimism. Finally, it is a matter of faith that we must leave to the
55 See the Wikipedia article: Big Bounce hps://en.wikipedia.org/wiki/Big_Bounce.
From Systems to Actor-Networks
42
“theologians” of systems theory.56 Another option we will explore in
this book is that when systemic order becomes too complex, the world
does not collapse into chaos. Instead, systemic order is transformed
into network order. At a certain level of complexity, systems become
networks. We will argue that this is what we are currently witnessing
in all areas of society.57
1.17 The Biological Model – The System as a Living Being.
Autopoiesis, Allopoiesis, and Autonomy
Before taking off on our excursion into the implications of the idea of
the self-organizing universe, we introduced a new kind of system that
brings new basic concepts into general systems theory. This was the
living system. According to Maturana (1985), a system that produces
itself is called autopoietic. “Auto-poiesis” means “self-producing” from
the Greek “auto” = “self” and “poiein” = “to make,”, “to bring forth.”
Systems, as we saw, consist of elements in certain relations, which
make certain operations possible. If the operations of a system consist
in producing the elements and relations of which it consists – and do
nothing else – then this system is an autopoietic system. It is a system
whose goal is nothing other than producing and reproducing itself. In a
definition that has already become classic, Maturana says:
An autopoietic machine is a machine organized (defined as a
unity) as a network of processes of production (transformation
and destruction) of components that produces the components
which: (i) through their interactions and transformations contin-
56 For an optimistic interpretation see e.g., Moltmann (1985:213), who equates the
absolute world-complexity or the environment of the universe with God: “All
individual maer and life systems as well as their communicative interrelation-
ships as a whole ek-sist into a transcendence and subsist from this transcendence.
If we call this transcendence of the world ‘God’, then we can say tentatively: The
world is in its details and as a whole a God-open system. God is its extra-worldly
environment, of which and in which it lives. God is its extra-worldly anteroom,
into which it develops. God is the origin of the new possibilities from which it
gains its realities.”
57 The doubtful reader can scroll forward to Part III of this book to see many illus-
trations of this tendency.
Basic Concepts of Systems Theory 43
uously regenerate and realize the network of processes (rela-
tions) that produced them; and (ii) constitute it (the machine)
as a concrete unity in the space in which they (the components)
exist by specifying the topological domain of its realization as
such a network. (Maturana/Varela 1980:78)58
Consider, for example, the case of a cell: a cell is a network of
chemical reactions that molecules produced in such a way that
they 1. produce, or recursively participate in, through their
interactions, the very network of reactions that they themselves
produced, and that 2. realize the cell as a material entity. The
cell, therefore, remains as a natural entity, separable topograph-
ically and operationally from its environment, only so long as
this organization is consistently realized by perpetual turnover
of matter, regardless of the changes in its form or the specificity
of the chemical reactions constituting it. (Varela et al., 1974:188)
What characterizes a living system is that its operations are not directed
to any goal beyond itself. This is what “self-reference” for a living
system means. A system whose output or product is something other
than the system itself, such as the air conditioner, which outputs heat or
cold and thus aims at producing a particular state in the environment,
is not autopoietic but allopoietic. The goal of the non-living system is
not the production of its own elements and organization. Nevertheless,
there are similarities. Self-organizing systems operate like allopoietic
systems in that they “try” to maintain their set point or at least some
steady state or equilibrium. Thus, they can be said to be homeostatic
and cybernetic, but unlike allopoietic machines, living systems are
58 See the explanation by Roth (1987:264): “...the production of the components
must take place in a more or less precisely dened way, and in such a way that
the component A (partly together with other components) creates exactly those
conditions under which component B can emerge, which in turn (together with
others) creates exactly those conditions under which C can emerge, and so on,
until nally a certain component (again together with others) creates the con-
ditions under which A can emerge again and the whole (in itself many times
intertwined) cycle of mutual production and repair can be run through again.”
This entire process has since been described in detail in terms of RNA protein
synthesis, ribosomes, transcription, and translation.
From Systems to Actor-Networks
44
oriented to their own organization, i.e., their set point is not imposed
on them from the outside but is set by their organization. They do not
“operate.” They “act.” It needs to be emphasized that they do not act to
cause some effect in the environment, but instead, to be able to continue
to act, to maintain their organization, or their autopoiesis. Life is agency
for its own sake.
Whatever events in the environment autopoietic systems respond to,
they respond not to change or to restore some state in the environment,
but whatever they do, whatever the operations might be, the purpose
is to enable the organism to continue its autopoiesis. This is important
because one cannot model living systems as agents of stability and
regularity. They do not seek the most probable outcome but the
outcome most conducive to autopoiesis.59 The air conditioner begins to
operate when certain information is registered about the difference of
the temperature to the setpoint. Such systems operate to minimize or
eliminate differences between what their setpoint demands and what
the environment is doing. Their agency is inherently conservative and
stabilizing. On the contrary, the organism responds to inputs according
to a genetic code that governs the autopoiesis of the system. For example,
an organism can react to cold or heat by running away instead of trying
to change the environment. It can respond by generating mutations that
allow it to withstand greater or lesser temperatures. The living system
is not concerned with the environment but with itself. Lions don’t hunt
zebras to reduce the population of zebras in the environment but to be
able to go on hunting. Lions are only concerned with themselves. A
system that is concerned with itself is autopoietic and also autonomous.
Autonomy means that the system, by its own structure, determines the
series of states through which it passes. The system’s state changes are
not controlled “from the outside,” i.e., heteronomously or mechanisti-
cally, but by the internal genetic code.
59 This is why Friston’s free energy principle and its Bayesian mechanics do not
oer a good model of living systems. Living systems are more than adaptive
behavior.
Basic Concepts of Systems Theory 45
1.18 Operational and Informational Closure, Self-Reference
The fact that the living system is concerned only with itself means
that it is autonomous, implying that it is an operationally closed system.
A system is operationally closed in a minimal sense if its states are
determined only by its own structure.60 Whatever the system does, it
operates only by virtue of its own organization and, thus, in a minimal
sense of self-reference, only in relation to itself.61 Operational closure
and self-reference mean the system’s operations are directed toward
itself. In this sense, the machine system of the air conditioner is also
operationally closed and self-referential since it operates recursively
according to its own structure. It is not concerned, and cannot be, by
anything else that happens in the environment other than whether it is
too hot, too cold, or comfortable. And in a certain sense, its goal is not to
change the environment, even if this always happens, but to maintain
its structurally given setpoints. In this sense, operational closure and
self-reference trivially apply to all conceivable cybernetic systems since
a system always operates by virtue of the organization/structure that
defines it and which delimits the system from the environment.62
At the level of autopoietic systems, however, operational closure and
self-reference imply informational closure. The system constructs infor-
mation out of perturbations from the environment; there is no exchange
of information between the system and the environment. An autopoietic
60 See Roth (1987:259): “Although materially and energetically open, it [the system
AB/DK] itself determines its state sequences on the basis of its specic internal
structure, which makes autopoiesis possible. ... Structurally determined and thus
autonomous systems may be externally stimulated or ‘perturbed,’ but these in-
uences do not determine the state sequences of the system.” Roth (264-5) notes
that the term “autopoiesis” has two meanings: self-production and self-preser-
vation. The rst meaning, self-production refers not only to living things but
also to physical systems. Only self-preservation describes autopoiesis in living
things.
61 The term “self-reference” can be applied to all recursive, cybernetic systems, e.g.,
also to the air conditioner. In this case it is of course not a question of self-con-
sciousness. Self-consciousness is the form of self-reference that takes place at the
level of meaning systems. Keeping this in mind would contribute to clearing up
much confusion in debates between materialism and idealism (panpsychism).
62 Thus, Luhmann (1984:31) can understand his claim that “there are systems” as
equivalent to the claim “there are self-referential systems.”
From Systems to Actor-Networks
46
system acts only to maintain its autopoiesis and thus refers only to itself
in the ways in which its internal differentiation allows. Information is
internally constructed out of undifferentiated perturbations coming
from the environment. The eye of the frog constructs information out of
quickly moving black smudges within its field of vision. The information
says, “This is good to eat.” The selection of a system-relevant environ-
ment, for example, the quickly moving black smudges, is done by the
system and not the environment. According to Luhmann (1990b:321),
“information...is not an input, but only the differentiation of connection
possibilities” within the system amounts to information. Information is
always only information for the system by the system. Information is
generated internally in the system and not read from the environment.
As noted, this need not mean that the system acts “self-referentially” in
the sense of “self-aware,” but only that it has no informational “input/
output” interface. Environmental events have no informational value for
an informationally closed system. They are merely “perturbations,”
i.e., undifferentiated disturbances that can only become information
due to the system’s organization or code. Of course, if there were no
structure and regularity in the environment, the frog would never have
been able to select quickly moving black smudges as information in the
first place. Indeed, life would not be possible without regularities in the
environment.
With regard to semiotically coded meaning systems, what has meaning
and can have meaning must be coded by the system as meaningful.
The meaning system does not find signs lying about in the environ-
ment and only needs to read them. If anything has meaning, and one
can talk about it, this is because it has been coded as information. As
Wittgenstein put it, “Whereof one cannot speak, thereof one must be
silent.” Operational and informational closure and self-reference are
constitutive of all autopoietic systems. In the case of systems that do
not produce themselves, that is, allopoietic systems, they are informa-
tionally open and, therefore, can “know” from the environment what
its code has selected as relevant. The autopoietic informationally closed
system does not “know” anything about the environment, and when
it comes to organisms, instead of meaning systems, the system does
Basic Concepts of Systems Theory 47
not even “know” about itself. Maturana (1987:105) describes this for an
organism and its nervous system as follows:
With respect to their state dynamics, however, the organism
and nervous system operate exclusively as closed systems
that merely generate structure-specified states, much as a pilot
behaves in instrument flight. When the pilot must fly and land
without visibility, he must keep the indicators of his aircraft’s
instruments within well-specified limits or follow a series
of specific variations. When the pilot leaves the aircraft after
landing, his wife and friends might come up to him and say,
‘That was a wonderful flight with an excellent landing!’ We
were afraid because of the fog!’ The pilot might reply, ‘What
flight? What landing? I didn’t fly, I just kept my display units
constant in certain areas.’ In fact, there was a flight only for an
external observer, and exactly this happens with an organism
and its nervous system.
When speaking of operationally and informationally closed systems,
this does not mean they are closed to everything. As Bertalanffy pointed
out, all systems are open with respect to the exchange of matter and
energy. Allopoietic systems are open to informational steering from
the environment and are not operationally and informationally closed.
Autopoietic systems, above all living organisms, are operationally and
informationally closed self-referential systems.
We can now add a third principle of general systems theory in addition
to the first two. The first principle states that all systems are organized
by selection, relationing, and steering, which constructs a system/
environment difference. The second states that all systems construct
elements that are not things but functions. And now, third, all systems,
at least those relevant for our purposes focusing on social systems, are
self-organizing, autopoietic, self-referential, and thus operationally and
informationally closed.
From Systems to Actor-Networks
48
1.19 Structural Coupling, Adaptation, Viability, and Again
Evolution
Structural plasticity, or the ability of an autopoietic system to change its
structure even though it is structurally determined, is called adaptation.
The term “adaptation” is misleading in that it presents the image of an
intentional act aimed to optimize responses to the environment. However,
an operationally and informationally closed system cannot “know” the
environment per se. It can only react to impulses, disturbances, or pertur-
bations coming from the environment according to its organization. If
the system has the ability to change its structure or organization, what
could be called “adaptability,” and if indeed the structure of the system
changes through interactions with the environment in such a way that
the system can continue its autopoiesis, then this appears to an observer
as if the system has “adapted” to its environment. Instead of adaptation,
one could speak of a structural coupling between organism and the envi-
ronment. This is described by Maturana (1987:101-2) as follows:
Since a structurally specified system can undergo only state
changes specified by its structure, the range of structural plas-
ticity of a composite entity is determined by its structure and
not by the medium in which the entity operates and is realized
as an entity. The medium can only perturb a structurally plastic
system and trigger a change of state that it does not specify.
Under these circumstances, the perturbations by the medium
act as selectors of the structural transformations of the perturbed
entity; and the sequence of perturbations that the medium trig-
gers in the history of the interactions of a given entity acts as a
selector of the sequence or course of structural changes that the
entity follows in that history. This results in the establishment
of a structural correspondence between the given entity and the
medium in which it operates, which appears to an observer as
adaptation or structural coupling.
Structural coupling explains the “correspondence” between system
and environment, or even between different systems, whereby it must
Basic Concepts of Systems Theory 49
be kept in mind that other systems are only environment for a system.63
Maturana uses the adaptation of the muscles of an organism to the
Earth’s force of gravity as an example of structural coupling. Gravity is a
remarkably constant feature influencing all life on Earth. Gravity selects
for bodies with specific muscular capabilities. Those organisms that do
not adapt to the force of gravity do not survive. Viability or adaptation
on Earth results from the structural coupling between organisms and
gravity. For this reason, it would be misleading to use the notion of
structural coupling to explain internal differentiation and, thus, internal
complexity/contingency in autopoietic systems. The heater and the
cooler subsystems within the air conditioning system could be said to
be adapted to each other by structural coupling, but this is only possible
insofar as they can each be considered autonomous systems in their own
right and not merely functional elements of the air conditioner. Are the
heater and the cooler really autopoietic, operationally, and information-
ally closed systems? Do they “adapt” to each other and to the air condi-
tioner? Or are they merely elements of the air conditioning system? The
same question can be asked of cells and organs in the body.
Internal differentiation of subsystems cannot easily be regarded as the
result of evolution, that is, the self-organization of autonomous auto-
poietic systems that are then structurally coupled to other components
of a system so as to build a systemic unity. Subsystems may, at best, be
“semi-autonomous” systems.64 It would thus appear that subsystems
are not operationally and informationally closed but are functional
elements of some over-arching system. They do not self-organize and
63 The correspondence of organism and environment has been discussed since an-
tiquity under the title of “teleology.” See for an overview the Wikipedia article:
hps://en.wikipedia.org/wiki/Teleology. Recently structural coupling has been
theorized as a “Markov blanket” under the free energy principle. Friston (2020:
1-2) writes “the free energy principle was introduced to describe how living
self-organizing agents have evolved to exist in a conned ‘state space’ with a
bound on their long term entropy.”.. which means they evolve “to avoid improb-
able … sensory encounters, i.e., those associated with existential risk.”
64 Interestingly, Luhmann speaks of social subsystems as “semi-autonomous” and
not as autonomous systems. Can systems be “semi” autonomous, semi autopoi-
etic, semi self-referential, semi operationally and informationally closed? We will
see later why this is an important question for understanding whether society
can be successfully modeled as a system.
From Systems to Actor-Networks
50
therefore do not evolve independently and relate to other body parts
via structural coupling. Does the liver, stomach, lungs, or skin self-orga-
nize as autopoietic systems and then enter into structural coupling with
the heart, brain, etc.? Or are they not much rather constructed within
the overall system out of the autopoietic operations of the interactions
of elements of the system following the genetic and morphogenetic
coding of the organism? This is what “autopoiesis” means. This is what
it means to say that a system constructs its elements.
Present-day biologists, such as Michael Levin, consider every cell an
autonomous system with its own information processing and adaptive
operations. Thus, every cell of a multicellular organism would be in
structural coupling with every other cell. Of course, Levin (2022) admits
that cells form “collective intelligence,” which implies operational and
informational openness and collective self-reference. According to
Levin, bioelectrical networks do this. They could be called a morphoge-
netic code related to DNA as computer software is related to computer
hardware. Out of this collective intelligence arise functional organs such
as lungs, heart, skin, etc. To what extent the software of an organism or
the morphogenetic bioelectrical networks can themselves be conceived
of as autonomous systems structurally coupled to other such software
modules within an organism is an open question. How are morpho-
genetic bioelectrical networks related to each other? Furthermore, the
extent to which individual cells that take instructions about what kind
of collective entity they should form, heart, skin, bone, etc., can still
be modeled as autonomous, operationally, and informationally closed
systems is also an open question. After all, they do what they are told to
do and not what they want to do themselves. Where is the autonomy?
Where are the informational closure and internal information construc-
tion? Indeed, cells that beak out of group conformity and go their own
way are those cells that Levin calls “cancerous.” The model of structural
coupling suggests, as Levin does too, that one must think of individual
cells in an organism or of subsystems in an organization as autonomous
systems. These issues are important because they lay the basis for what
it means to speak of a social system which perhaps cannot be modeled
in ways similar to biological systems.
Basic Concepts of Systems Theory 51
There is a great temptation to generalize and apply the ideas of struc-
tural coupling and adaptation to every element, for example mito-
chondria, within an organism, that is, to see them as tiny organisms
within organisms, where each is constituted by a system/environment
difference, is autopoietic, and operationally and informationally closed.
However, we do not find this extension of the theory advisable. On the
contrary, cells, tissues, organs, and subsystems should only, in a deriv-
ative sense, be said to be “adapted” or “structurally coupled” to the
organism as a whole or to each other. Instead, it could be said that they
are integrated into the organism. There is a difference between adap-
tation and integration. Can one meaningfully say that the digestive
system is “adapted” to the cardiovascular or nervous system within
an organism? Could the heart and lungs somehow mutate and put
“adaptive pressure” on the stomach? Does an organism’s genetic and
morphogenetic code represent a kind of ecosystem in which various
autonomous systems, over time and under the pressures of variation
and selection, have adapted to each? Is it an accident that systems
biology drops the word “system” when describing an organism’s
internal complexity and speaks of “networks of networks”?65 Why does
Levin speak of bioelectrical networks? With regard to later consider-
ations on the relationship between living systems on the one hand and
semiotically organized meaning systems on the other, it is important to
keep clearly in mind the distinction between structural coupling and
integration, or as it is also called interdependence, or again, as Luhmann
will say interpenetration.
A system that allows structural changes to be “controlled” by selection
within the framework of a postulated ecosystem is considered to be
viable. Viability means that the system, however it may change struc-
turally, meets the constraining conditions of the environment. This is
called “natural selection.” The environment “selects” which organisms
are viable and which are not. If the system does not adapt to the envi-
ronment, it cannot continue its autopoiesis and disappears.66 When the
65 See Kiani et al (2021).
66 Maturana (1987:107): “In fact, organism and medium are operationally indepen-
dent systems in their state dynamics, each following its independent structural
From Systems to Actor-Networks
52
structural changes arising from system mutations are so radical that the
organization of the system, i.e., its identity, is affected, then we speak
of evolution or speciation. Evolution is not mere structural change but
organizational transformation, i.e., the development or emergence of
new forms of systems out of the autopoiesis and operational closure of
existing systems. This presupposes that systems can reproduce them-
selves, which is the very meaning of autopoiesis, and that variation can
and does occur in this process. Thus, different species and branches of
the tree of life, or phylogenesis, occurs. Governed by the evolutionary
principles of variation and selection, life explores all possible niches
of morphogenetic space. Evolution is the process by which the many
different forms of life emerge, transform, and disappear. Evolution is a
natural process. If, for example, a genetic engineer creates a new living
being in the lab, which is called synthetic life, this is not evolution since
the organizational transformation of the genetic code in question did not
arise in the process of autopoiesis. Synthetic life is an allopoietic inter-
vention. But after the living being once exists and can reproduce, organi-
zational transformations in the sense of evolution become possible.
1.20 Excursus: Ecology and Sustainability
There is much uncertainty in terminology between the idea of subsys-
tems within systems, the mechanisms constructing internal complexity,
and the relation between systems and their environments, including
other systems. It would seem that systems theory offers only structural
coupling or adaptation to explain how systems relate to anything beyond
their constitutive boundaries. Internally, if elements of a system are not
defined as autonomous systems in their own right, relations among
elements can be modeled as dynamic interrelations. Most systems
analysis does this, whether in physics, biology, or society. Almost
everywhere the word system is spoken, units are being analyzed into
parts that are dynamically interrelated to each other to produce specific
specication. Therefore, appropriate behavior is necessarily possible only as a
result of structural coupling, and an organism is therefore either in a state of
structural coupling or in a state of dissolution.”
Basic Concepts of Systems Theory 53
operations or emergent behaviors. The modeling is almost always done
for purposes of simulation, prediction, and control. This is especially
the case where mathematical models are used in engineering to build
autonomous machines. Regardless of whether the relations among
elements are understood as adaptation or structural coupling, irrespec-
tive of whether the system being analyzed is a machine, an organism, an
organization, a society, or whatever, the assumption is that the processes
of some circumscribed and bounded composite can be described in such
a way that simulation and intervention in the operation of the system
become possible that allow for prediction and control.
In this context, what do we mean when speaking of an “ecosystem”? In
most talk about ecosystems and all of ecology as a science, the environ-
ment is falsely conceived of as a system. The various organisms within
the ecosystem are then automatically thought of as subsystems within
some encompassing “eco”-system. According to Luhmann (1990:21),
interestingly, the ecological question does not belong to systems theory
at all since it “[makes] the unity of the difference between system and
environment the subject, but not the unity of a comprehensive system.”
Let us recall that systems thinking became popular at about the same
time and is closely connected to ecology. The idea of systems entered
public awareness with the first Club of Rome Report, “The Limits of
Growth,” published in 1972 under the leadership of Dennis and Donella
Meadows. The Club of Rome proposed a World Model as the frame
used to analyze factors affecting the limits to growth. The model was
based on systems theoretical ideas of holism, circular causality, homeo-
stasis, and the interdependent relations among elements leading to
emergent properties, properties of the whole that could not be derived
from the elements individually. The basic concepts, the methodology,
as well as the focus on the interaction of society with the natural
environment falsely branded the new science of ecology as a systems
science. The systems model became unavoidably related to ecology.
The term “ecosystem” expresses this connection, which has continued
until today in the idea of “Earth Systems Science.”67
67 See Steen et al (2020).
From Systems to Actor-Networks
54
It is interesting to note that the object of ecological investigation, what
is described by ecology, that is, the interaction of organisms with an
environment, cannot itself be considered a system. It is a fundamental
tenet of the systems model that the system/environment difference
constitutes a system. The unity of the difference between system and
environment cannot itself be a system. For example, a frog may be
considered an organic system adapted to life in a pond which is its
environment. Suppose the interactions of the frog with the pond are
considered to build a system, as talk of an ecosystem suggests. What
distinguishes this ecological system from its environment, an envi-
ronment that it must have to be a system? What is the environment
of the pond? If we suppose it is the forest in which the pond exists,
the ecological principle of observing all interactions of organisms and
their environment together as one unity requires the entire forest to be
considered the ecosystem. The whole forest then becomes the system
that is being observed. This only pushes the question back. What is the
environment of the forest? If the forest is located in the mountains or
on a plain, if it is near a big city and thus used as a recreation area, all
this becomes the object of analysis. What is the ecological impact of all
these factors? Must they not also be taken into account and considered
as the larger ecosystem of the pond and the forest? The forest-city or
forest-mountain complex becomes the ecosystem. The ecosystem’s
boundaries must continually be pushed back and enlarged indefinitely
until, as is today the case with climate change, the entire planet Earth
and beyond the whole universe must be taken account of. Ecology leads
directly to Earth Systems Science and beyond to Gaia and the cosmos.
Does not the moon, cosmic rays, solar storms, etc., also influence the
pond? Is not the ecosystem, in fact, the whole world, as Earth Systems
Science in the very name proposes?
Whenever there is a system, the environment, according to systems
theory, must be conceptualized as different from the system. How could
the system select, relate, and steer elements, refer its operations to itself,
and establish and maintain autopoiesis, if there were no difference
between the system and environment? Ecology necessarily includes
the environment in the system. This means that the system/environ-
Basic Concepts of Systems Theory 55
ment difference is continually dissolved and pushed outward only to
be re-established on a higher level. The consequence is that ecology
cannot be a systems science because it demands that the interactions
between the system and the environment themselves are the object of
investigation. Ecology studies what systems science cannot envisage:
the unity of the difference between system and environment. For
ecology, the environment is unavoidably included within the system,
which automatically establishes a new system/environment difference,
which is then transgressed until we reach the entire universe, or at
least, the planet Earth. As often said, ecology leads directly to the Gaia
hypothesis. The world as a whole must, in principle, if not in practice,
be included within the ecosystem. This, however, is practically impos-
sible. How can the immense complexity of the world be modeled, simu-
lated, predicted, and controlled? How can models of “sustainability”
for the world as a whole be designed? What must be held stable, and
what can be allowed to change? How many setpoints must be taken
into account? How many different processes? How could the effects
of interventions by geoengineering with regard, for example, to the
average global temperature, be calculated for all foreseeable outcomes?
But this is not only a practical problem that perhaps some supercom-
puter and unimaginable quantities of data could solve. It is a funda-
mental problem in theory. According to systems theory, there can be
no world system, no system that includes everything. Systemic order is
based upon selection, exclusion, and the reduction of complexity. The
system is always less than the environment. The ecological model, on
the contrary, is based on the inclusion of all factors and the exclusion of
nothing. For these reasons, it can be argued that ecology forms a kind
of proof case for systems theory. Can the complex interdependencies
typical of ecology be modeled as a form of systemic order, and if not,
what alternatives are there? Perhaps we cannot limit ecology to the
question of life but must include meaning. Perhaps Latour’s “critical
zone”68 must include systems of meaning and not only geology and
biology. Perhaps Gaia is not a system at all but a network. Perhaps Gaia
is not only geology and biology but also culture, society, and meaning.
68 See Critical Zones. The Science and Politics of Landing on Earth, Latour/Weibel (2020).
From Systems to Actor-Networks
56
Maybe the ecological crisis is a crisis of meaning, not life. We will return
to this question when discussing meaning systems and networks below.
1.21 What is learning?
The appearance of new behaviors due to structural changes in an auto-
poietic, operationally, and informationally closed system can be called
learning. However, learning presupposes that behavior changes are not
caused by changes in the organization, i.e., in the system’s identity. You
may be able to teach an old dog new tricks, but you can’t teach an old
dog to become a new dog that is a new species.69 Organizational change
is evolution and not learning. Learning is ontogenetic structural trans-
formation and not phylogenetic and evolutionary change. In this sense,
one can also speak of “machine learning.” This presupposes, however,
that the machine can develop new behavior based on its interaction
with the environment or data. Simply programming a computer is
not machine learning or what could be called “artificial intelligence.”
Insofar as neural networks can set their weights and create their models
to achieve goals, this could be understood as learning. If a human
programs the computer, then the behavioral transformation no longer
comes from the operational closure of the system but is determined from
the outside. In organic systems, one can distinguish between instinctive
behavior, where there are no degrees of freedom, and behavior that
arises due to the system’s interaction with the environment. Michael
Levin (2022) has argued that every organism can learn and thus can
be attributed some kind of “cognition.” Cognition Levin defines as
problem-solving, and whenever a system has more than one degree of
freedom and can solve a problem in more than one way, it can be said
to be able to learn the best way using trial and error. For Levin, this is
true of the behavior of single-celled organisms up to human cognition.
Levin, and most cognitive science, do not consider the possibility that
meaning systems operate entirely differently than biological systems.
This confusion leads to many misconceptions about the nature of society
69 This does not imply that epigenetics and interaction with the environment can-
not lead to inherited changes.
Basic Concepts of Systems Theory 57
and of meaning. It is advisable, therefore, to distinguish different kinds
of learning on different levels of emergent order. Machine learning
differs from problem-solving on the biological level, which in turn is
different from problem-solving on the level of meaning. What they all
have in common, from a systems theory perspective, is structural or
behavioral change resulting from system operations.
1.22 Meaning Systems, the Semiotic Code, and Cognition
Let us imagine a very complex system. More complex than anything
we find in physical or biological nature. Let us assume that the system
regards its behavior, i.e., its structural transformations or everything it
has learned, as “signs,” that is, as something which means something.
Furthermore, let us assume that the system is self-referential in the
sense of being self-conscious.70 It can be said that the meaning system is
a sign to itself, or as Pierce once said, “Man is a sign.” Furthermore, we
will assume that the system can store and retrieve information at will
and that it has the ability, perhaps even the necessity, to communicate
these stored and retrieved “ideas” to similar systems using language.
If all this is assumed, we have again come to a threshold where the
previous model of the organism is no longer sufficient to represent the
phenomena we are describing.
Just as at a certain level of complexity, the model of the machine had to
be replaced by the model of the organism, so the model of the organism
70 Animals of course are also conscious and most probably self-conscious, but they
cannot tell us about it so there is fundamental gap between animal consciousness
or sentience and the human realm of meaning that can best be characterized
as the technology gap. Although life can be equated with cognition and intelli-
gence, these concepts mean something else on the level of meaning. On the level
of meaning they mean technology, the “articial,” the mediation of tools in all
activities. Again, one must weigh continuity against transformation. Even if all
levels of emergent order from the physical to the semiotic are based on one sub-
strate, the substrate is transformed at each level. Despite all continuity, there are
undeniable dierences in levels of emergent order. The question is whether the
dierences so outweigh the continuities that the concepts at issue become mere
equivocations and the theory loses its generality. We will argue that this is in fact
the case.
From Systems to Actor-Networks
58
must now be replaced by a new model that exists on a higher level of
emergent order.71 Just as the genetic code as an organizing principle
replaced the mechanical/physical code of the air conditioning system,
the genetic code itself must now be replaced by another code that is
more complex and variable than life and which can do things with life
that life alone cannot do. What kind of code is this? What type of model
allows us to replicate the complexity of the above assumptions?
If self-reference occurs as conscious and meaningful self-reference,
and learning can be stored, retrieved, and communicated, then we are
dealing with what could be called a meaning system. A system capable
of meaningful operations can be organized by what we shall call a semi-
otic code. It must be emphasized that although semiotics refers to signs,
signs are not coterminous with meaning. Long before Homo sapiens,
with their big brains and unique linguistic abilities, appeared on the
scene, hominins were using tools and creating something that could
be called “culture.” We are using the term “semiotic,” therefore, in a
very broad sense that includes, of course, language but goes far beyond
that which the discipline of linguistics studies. A semiotically organized
system can therefore be called a meaning system or, since meaning and
language of some kind are inseparable, a communication system. Thus,
communication can become the defining feature of social systems for
Luhmann (1986:269):
A social system comes about when an autopoietic communi-
cation context always emerges and demarcates itself from an
environment by restricting the appropriate communications.
Accordingly, social systems consist not of people, nor of actions,
but of communications.
These statements demand an explanation. For Luhmann, the model of
a meaning system as a communication system is society. What needs
71 It is estimated that life emerged on Earth about 4 billion years ago. The origin
of meaning can be dated to more than 3 millions years ago, that is, with the rst
stone tools. Homo sapiens, however, with its big brain and linguistic abilities
appeared about 300,000 years ago. Meaning is much older than Homo sapiens
and also much oder than language.
Basic Concepts of Systems Theory 59
to be understood is no longer the biologically or genetically organized
organism, and not even the human individual with its large brain and
the experience of consciousness, which is the subject matter of cognitive
science, but instead it is human society, which exists on its own level of
emergent order. This evolutionary leap is described by Wilke (1990:27)
thus:
The leap from the realm of necessary stimulus-response linkage
to the realm of heightened degrees of freedom and contingency
marks a threshold of evolution associated with the concepts of
meaning, identity, reflection, self-understanding, and self-the-
matization.
Accordingly, we must no longer orient systems theory towards biology,
and not even towards psychology or cognitive science, but towards
sociology and semiotics. This suggestion is highly controversial. That it
takes a radical reorientation to describe a system of meaning adequately
is not accepted by biologists and most cognitive scientists.72 These disci-
plines are forced into a reductionist standpoint by disregarding the leap
to a new level of emergent order. Physicalist or materialist reductionism
claims only particles, fields, matter, and energy exist, and all other
phenomena, including life and meaning, must be “explained” in these
terms. Your Grandmother is nothing but atoms in motion. Along these
lines, it is supposed that when quantum gravity is finally adequately
theorized, this would be a Theory of Everything (ToE).73 But of course,
such a purely physical theory leaves life and meaning out of account
and can hardly claim universality. And it cannot explain itself and thus
alone on this criterion cannot claim to be universal. But not only the mind
is left out of account; physicalist reductionism cannot explain life either.
The same game, however, can be played on the biological level. Because
of the informational closure of living systems, it is supposed that we,
as living systems, are locked within our brains and cannot know the
72 The exception that proves the rule is non-Cartesian cognitive science. See for an
overview Rowlands (2010).
73 See Wikipedia for an overview. hps://en.wikipedia.org/wiki/Theory_of_every-
thing.
From Systems to Actor-Networks
60
world “out there” in the environment. All information is constructed
within and by the system according to its own organization. Therefore,
the world we know is not the real world but an illusion that is more or
less viable.74 Much of current systems theory is, in one way or another, a
form of biological reductionism. Reality is reduced to being an informa-
tional construction of the brain or, as is often said, a “model,” a “simula-
tion,” or even, as Hoffman (2019) claims, an “illusion.” This thesis is not
new. It goes back at least to Kant and has recently been discussed under
the name of “radical constructivism” or “second-order cybernetics.”75
Because of the prevalence, if not omnipresence, of biological meta-
phors and models in systems theory, there is a widespread tendency
to explain meaning systems “biologistically” and thus not to recognize
that meaning is a level of emergent order beyond life and, therefore,
that meaning systems must be modeled differently from living systems.
Since systems theory has been developed almost exclusively as a theory
of cybernetic machines and living organisms, it is necessary to address
the limitations this imposes on the theory before discussing systems of
meaning in their own right.
1.23 Organism or Society? On the Problem of Biological
Reductionism in Systems Theory
In contrast to the thesis that meaning must be conceived of as a third
level of emergent order beyond matter and life, Maturana, Levin,
Friston, Hoffman, and many others claim that the biological model is
sufficient to “explain” language and meaning.76 As Maturana puts it,
74 This is fundamentally a Kantian position to which most cognitive science ad-
heres. See radical constructivism in psychology and epistemology and Homan
(2019) for a current exposition of this form of biological reductionism.
75 See the Wikipedia article on radical constructivism and second-order cybernet-
ics: hps://en.wikipedia.org/wiki/Radical_constructivism; hps://en.wikipedia.
org/wiki/Second-order_cybernetics.
76 The reader should keep in mind that there are many confusing crossovers and
interferences between philosophy of mind and its many debates about mind/
body dualism, psychological behaviorism, deeply rooted metaphysical commit-
ments to materialism or idealism, and various schools of cognitive science that
make the issue discussed here dicult to restrict to systems theory alone.
Basic Concepts of Systems Theory 61
life can “model” meaning: “Cognition is a biological phenomenon and
can only be understood as such” (1985:33).77 Surprisingly, Maturana
(1987:112) openly admits that this approach has nothing to do with
meaning or mind in the sense that philosophy of mind has usually
understood these terms. He openly admits that the biological model
reduces language to “contentless consensual behavior.”
Observers and language have been generated by the linguistic
description of living systems and their operations. But in
this process, language appeared as a contentless consensual
behavior whose validity owed solely to its consistency within
the consensus domain.
And even more directly (113):
The actual operation of language as an ontogenetically estab-
lished system of coordinated mutual elicitation of behavior
between organisms is contentless.
The word “content” refers to semantic meaning. This biologistic view
of language deviates so radically from the usual concept of language
that Maturana (112) must postulate another, higher domain of opera-
tion, which he calls a “consensual domain,” to do some justice to the
phenomenon of meaning, communication, and intersubjectivity:
An observer operates in two overlap-free phenomenal domains.
As a living system, he operates in the realm of autopoiesis. As
an observer proper, he operates in a consensual realm that
exists only as a collective realm determined by the interactions
of multiple organisms.
The decisive difference between these two non-overlapping domains is
that the observer, as a living system, is either individual or collective.
77 See also (39): “Living systems are cognitive systems, and life as a process is a pro-
cess of cognition. This statement applies to all organisms, whether or not they have
nervous systems.” See also Maturana/Varela (1987:33). Levin (2020) also follows
this line of biological reductionism in his theory of life as cognition and Homan
(2019) takes it to its logical conclusion with his claim that reality is an illusion.
From Systems to Actor-Networks
62
The individual is concerned only with its autopoiesis, but the collec-
tive is somehow concerned with the mutual interactions of organisms.
What is this collective? As Luhmann will claim when speaking of social
systems, it is a special kind of system? Or is it merely a matter of two
different perspectives on the same system? Does the collective interac-
tion of organisms create a new and different system on a higher level of
emergent order? If one were to consider this “consensual domain” as a
system in its own right with its own organization, this would require
that one leave the purely biological level and abandon the model of the
organism.78 We dealt with this question above when discussing the idea
of ecology. We saw that an ecosystem cannot be a system because it
cannot maintain a constitutive difference from an external environment.
The same argument applies in the present context. The “consensual
realm” of which Maturana speaks as the domain in which the observer
supposedly exists is a “collective realm determined by the interactions”
of many individual organisms. What informs, relates, and organizes
these interactions? What constitutes the unity and characteristics of this
collective realm? With regard to ecology, it is uncertain whether Gaia is
to be thought of as a super-organism or as something else not describ-
able as a system of the same kind as life. What type of code organizes
the consensual domain Maturana ascribes to the “observer proper?”
If we remain within the systems paradigm, we must postulate a funda-
mentally different kind of system other than autopoietic, self-refer-
ential, operationally, and informationally closed systems, or we must
abandon the idea that the observer can operate in any kind of special
domain. The observer as observer, again, falls out of the theory. This
implies that language and meaning cannot be thought of as a system
on the biological level of order. Biology cannot account for biolog-
ical science itself. Within the biological model, meaning becomes an
epiphenomenon of life somehow generated by the structural coupling
78 Levin seems to say both at once. The individual cell is at once an autonomous,
operationally and informationally closed system and also bound up into a collec-
tive that is following instructions from a morphogenetic, bio-electrical network
in order to form a multicellular organism, which in turn can be bound up into
colonies, groups, and societies. It is unclear whether the units being observed are
systems or networks.
Basic Concepts of Systems Theory 63
of individual organisms or, from the point of view of cognitive science,
even individual neurons. It is difficult to explain why individuals
should experience this as meaning, as the “world” in which they exist.
We find ourselves facing the “hard problem” of consciousness again.
One cannot derive meaning from non-meaning, but what we can do,
as we argue, is postulate the emergence of a higher level of order upon
which language, meaning, communication, and human society appear.
Whether they appear as systems is the question systems theory must
answer convincingly to maintain its claim to universality. Language is
the construction of information by means of communication and not
by means of perturbations triggering information construction internal
to a living system, even if many organisms somehow do this together.
If organisms are integrated into some higher level of systemic order and
not merely structurally coupled in some kind of consensual domain,
what is this higher domain? How can the “interactions of multiple
organisms” give rise to a system of a different kind such that this system
is not merely a super-organism? Does life “self-organize” into meaning?
Or does meaning “emerge” based on complex forms of life?79 To answer
these questions, Maturana helps himself with metaphors and analogies.
First, it is asserted that the observed coordinated behavior between
organisms is not something different from human linguistic behavior.
For an external(!) observer, what humans do when acting together is
no different from what ants or bees do, or at least for such an external
observer, it seems to be similar. Bees dance in front of one another just
as humans gesticulate and make noises in front of each other. Apart
from the fact that neither Maturana himself nor anyone, for that matter,
is “external” to language and communication, Maturana goes on to
suppose that the structural coupling of behavior among living systems
appears to be “like” language (109); thus, a continuity is postulated
between genetically and semiotically coded systems:
When two structurally plastic composite entities interact with
each other and thus act as selectors of their respective paths
of structural change, mutual structural coupling takes place.
79 Self-organization is continuous, whereas emergence is discontinuous.
From Systems to Actor-Networks
64
This results in the state changes of one system recursively
triggering the state changes of the other, thus constituting a
domain of coordinated behavior between the mutually adapted
systems. When this occurs between living systems during their
ontogeny, a domain of coordinated behavior is formed that is
indistinguishable from a consensus domain that has been estab-
lished between humans.
And further:
To an observer who sees two or more organisms operating in a
consensual domain, the organisms appear to be operating with
consensual representations; therefore, to the observer, a consen-
sual domain functions like a linguistic domain in which the
behaviors of interacting organisms create consensual indexes
or descriptions of their interactional domain.
Reflexivity is said to bridge the gap between meaningless mutually
conditioned behavior on the one hand and meaningful linguistic
communication, that is, “consensual representations” or “descriptions”
on the other. Meaning appears out of nowhere as “representations” or
“descriptions” that the observer somehow knows, or at least supposes,
are present for the observed organisms. There are two different things
here that cannot be equated. There is the fact that the behavior of a
collective of organisms “appears” to an observer as though they were
using shared representations to understand each other and coordinate
their behavior, and there is the entirely different fact of whether they are
actually doing this. It is one thing to appear in a certain way and another
to actually be as one might suppose. Where does this knowledge of a
linguistic domain come from? What is supposed to be explained by this
idea is simply assumed but immediately suppressed and relegated to
external structural coupling. According to Maturana, “the interactions
within the consensual domain operate as selectors for further struc-
tural couplings within the domain.” This leads to a “recursion” from
operations to operations. From this, meaning and language are said to
emerge (109):
Basic Concepts of Systems Theory 65
...then a recursion takes place in which the participating organ-
isms make consensual determinations of consensual determina-
tions. A linguistic domain with such features is a language, and
to operate in it is – from the observer’s point of view – to operate
in a descriptive domain that allows recursive descriptions of
descriptions.
That meaning refers to meaning, and communication leads to further
communication is not what is observed, but what is experienced. One
does not simply observe two people making noises and gesticulating in
each other’s presence, but one knows they are talking about the weather
or who won the football match. Making “consensual determinations
of consensual determinations” cannot be understood as structural
coupling. Meaning referring to meaning or communication referring
to communication is not an example of structural coupling but of the
self-referential operations of a meaning system. How the recursion of
structurally coupled operations upon themselves magically is supposed
to create semantic content, i.e., meaning remains unclear. We speak
of magic because one is getting something, meaning, from nothing,
non-meaning. How is this to be explained? The model lying behind this
sleight of hand is the central nervous system. The brain is supposed
to be an operationally closed system of nerve impulses, i.e., “relative
neuronal activities” always lead only to other neuronal activities. The
system is operationally closed. It operates upon itself “recursively” in
that its own neuronal operations lead again to neuronal activity; or, as
von Foerster (1993:34) puts it, we are dealing with “the computation of
computations.” Recursion alone, however, will not do the work that
needs to be done here. That nerve impulses operating upon themselves
at a certain point of complexity and scale somehow transcend the
purely biological level of order and trigger a “higher” level of neuronal
networking and processing which is then equivalent to conscious-
ness and meaning, is an interesting and perhaps within the biological
model, necessary assumption. It would imply that the lower impulses
are somehow meaning-referentially mapped to the higher ones or, as
we may also say, semiotically coded. The metaphors of “lower” and
“higher” are, of course, misleading. There is no hierarchical structure
From Systems to Actor-Networks
66
involved here. There is a leap into another level of emergent order. It
is not convincing to suppose that the operational closure of either the
neuro-system or the structural coupling of behavior upon itself, that is,
mere recursion, can explain the emergence of language and meaning.
Where does genetic coding turn into semiotic coding? There seems to
be no alternative to postulating the emergence of a higher level of order
that is semiotically coded. Maturana (1987:114), and following him,
many cognitive scientists are determined to explain the ability to think,
i.e., the ability to use meaningful signs linguistically and communica-
tively, solely based on the operational closure of the nervous system:
However, as observers, we can always specify a metacognitive
domain from whose point of view we are external to it and
consider it. We can do this, of course, because in our neural
system everything happens in the same phenomenal domain of
closed relations of relative neural activities.
As Küppers/Krohn (1992:22) note, semiotic reference or meaning cannot
be explained by recursion alone:
Often...the notion of reference is used synonymously with that
of recursion. But while recursion merely specifies that an output
of systemic operations becomes its input again, the notion of
reference in the narrower sense means that some states repre-
sent or refer to others. More sharply, representation or reference
is a state that is a representation (a model) of another state.80
In conclusion, we can assert that the emergence of language and meaning
cannot be explained by means of structural coupling and recurrence
of neuronal activity and the resulting appearance to some observer of
what is construed as “consensual” behavior. There is no “consensus”
here; there is only mutual adaptation. The organisms or nerve cells do
not “communicate” or “ agree” on anything. Coordination of behavior
is not in itself communication or meaningful consensus or agreement.
80 See also Schiewek (1992:154), who distinguishes between “operative” and
“symbolic” generalization, the crucial point being “that these symbolic general-
izations are no longer subject to biological laws....”
Basic Concepts of Systems Theory 67
Ants, bees, in fact, almost all species in some way coordinate their
behavior, indeed, the entire world of life is always a more or less stable
state of mutual adaptation. Any “balanced” ecosystem can appear as
“coordinated” behavior. But we do not think of this, except metaphor-
ically, as communication and meaning. According to this model, how
recursively, i.e., cybernetically interconnected nerve impulses or how
a semantically contentless mutually coordinated behavior becomes
semantically filled language remains unexplained. That there are
cybernetic recursions at mechanical, biological, and semiotic levels is
not denied by anyone. But the model of the organism (or of the nervous
system) is insufficient to represent the specifically linguistic forms of
self-reference and meaningful communication observed and experi-
enced on the level of meaning.
It may well be that, from a particular observer’s perspective, structurally
coupled behavior looks like linguistic communication. But this does not
mean that it actually is language, as is shown, for example, by talk of
the “language” of bees or ants, etc. As Jantsch (1987:180) puts it:
For species representing an early stage of evolution, e.g., insects,
the close coupling by physical and chemical processes (such as
by chemotaxis) extends further to the association of organisms
into societies. But in the later and more complex species, such
as mammals and especially humans, the formation of societies
is based not only on material processes, such as production and
division of labor, but also primarily on mental and spiritual
processes, which arguably find their counterpart in neural,
chemical, and electrical processes within the organism, but
represent a higher level of self-organization.81
81 It is equally dicult to understand how “observation” should be possible on
the basis of the biologistic model. If observation is seen as a synthesis of “dis-
tinction” and “indication” following Spencer Brown (1972), then “indication”
is a semiotic term that can only have meaning within a system of meaning. It
is impossible to see, as Kneer/Nassehi (1993:97) claim, that a thermostat, which
is not part of a meaning system, can “designate” temperature dierences. The
thermostat designates temperature dierences for people who read the device
as a sign. For itself, the thermostat does not signify anything and thus it does not
observe anything. We will return to the notion of observation below.
From Systems to Actor-Networks
68
We know from Wittgenstein that all behavior is interpretable as
language when viewed from the outside, but that is precisely why no
external observer can discern language or meaning. Speech, for Witt-
genstein, is rule-governed activity, and only those who “know” the
rules and are able “to go on” in any communicative situation can recog-
nize language. Merely knowing the rules cannot mean interpreting
any particular behavior as rule-governed on the basis of observation,
for, as Wittgenstein pointed out, everything can be interpreted as both
rule-conforming and rule-violating. Instead, understanding a language
means that one knows how to go on, how to participate in a language
game. Thus, Wittgenstein can say (1984b: n.102):
Our paradox was this: a rule could not determine a mode of
action since every mode of action was to be brought into confor-
mity with the rule. The answer was: If everyone is to be brought
into agreement with the rule, then also to contradiction. There-
fore, there would be neither agreement nor contradiction here.
The paradox can be resolved only if the specifically linguistic rules – we
can also speak of a specifically semiotic code – function normatively and
non-determinatively. This implies that semiotic coding is recognized
by being subject to the rules so that one can bump against them and
be corrected by other participants in communication. Intersubjective
corrigibility is what makes a rule into a rule. Meaning is not a matter of
observation but of mutual recognition and participation. Only based on
participation in communication can there be “right” and “wrong” in the
normative sense. Contrary to this, Maturana/Varela (1987:191) argue:
Let us be clear, therefore, that the assessment of whether there is
knowledge or not is always in a relational context in which the
structural changes triggered by perturbations in an organism
appear to an observer as an effect on the environment. And
the observer estimates the structural changes triggered in an
organism in terms of the vortices he expects to see. From this
point of view, every interaction of an organism – all its observed
behavior – can be evaluated by an observer as a cognitive action.
Basic Concepts of Systems Theory 69
Thus, the fact of life itself – the uninterrupted maintenance of
the structural coupling as a living being – is nothing else than
cognition in the realm of existence. Formulated as an aphorism:
Life is cognition.
According to Wittgenstein, only those can determine language and, thus,
cognition who can, in principle, participate in communication. But this
requires that observer and observed both together already act within a
semiotically organized meaning system, i.e., the system reference is a
semiotic meaning observable only from the inside and not an organism
and not the brain. We find ourselves automatically and necessarily on a
higher level of emergent order. In this case, when the observer describes
cognition, recognition, meaning, or language, they describe themselves,
and thus what they observe is no longer something external to them-
selves. Observation on the level of meaning systems is always self-ob-
servation and self-definition. This is indeed recursive, but it is nothing
other than the self-reference of the meaning system. Self-description – we
know this from the explanation vs. understanding debate in the meth-
odology of the humanities – is fundamentally different from the objecti-
fying description of external objects and is subject to other semantic and
pragmatic rules.82 The meaning of terms like “observation” and “cogni-
tion” changes substantially when it comes to self-observation, that is, to
the operations of a meaning system. Although a general systems theory
must claim that there are continuities and sufficient similarities so that
the same concepts can be used on the biological and the semiotic levels,
this, we argue, is not at all certain.
That systems exist only for an observer nevertheless remains valid since
observation, as we shall see in a moment, is defined as introducing a
distinction. A system can only “exist” when the system/environment
difference is applied. The introduction of this distinction is itself an
observation. If the operation of distinguishing necessarily implies
that the observer distinguishes themselves from the observed, then all
observation must be merely external observation. There would be no
self-observation. The notion of observation is problematic insofar as it
82 See Apel (1979).
From Systems to Actor-Networks
70
almost inevitably implies the idea of a mere external observation while
simultaneously implying self-observation. It is another edition of the
subject/object distinction, which entails the possibility that the object
can be the subject and vice versa. But if we hold that observation means
the articulation of differences, then the classical problems of episte-
mology concerning subject and object and the paradoxes of the subject
that is also object, etc., can be formulated differently.
Meaning systems, which imply cognition, recognition, and communica-
tion, cannot be adequately understood by objectifying methods. Indeed,
meaning systems do not “exist” for an external observer. Physics and
biology do not and cannot explain meaning. The observer, as Heidegger
pointed out, is always already “in-the-world” of meaning, indeed, exists
as meaning. Furthermore, a theory of meaning systems cannot be based
on self-observation as long as observation is understood as objectifying
description. With the transition from a physical or a biological theory
to a psychological or sociological theory, self-observation becomes
self-understanding in the communicative and hermeneutic sense. For
this reason, we speak of meaning as a higher level of emergent order.
The disregard of this transition in current formulations of systems
theory or constructivism explains why these theories are incapable of
solving the problem they must solve to justify themselves, namely the
problem of meaning. As long as the theory itself is founded on an insuf-
ficient analysis of the conditions of meaning – as is the case for all forms
of physicalist or biologistic reductionism – it will be self-contradictory.83
Furthermore, the fact that there are systems organized on the basis of
meaning, i.e., of a semiotic code, is shown by the fact that biological
survival, as all those who have died for their beliefs show, becomes
subordinate to the principle of meaning. Social phenomena cannot be
explained as attempts of organisms to pass on their genetic material or
to adapt to an environment to ensure viability. There is no “free energy
principle” guiding the emergence of meaning and society. Millions
of organisms are sacrificed on the world’s battlefields for the sake of
83 See Bernardo Kastrup for the critique of physicalist and biologistic reductionism,
hps://www.bernardokastrup.com/.
Basic Concepts of Systems Theory 71
beliefs, ideologies, struggles for influence and power, etc. These organ-
isms are not concerned with passing on their genes but with passing
on symbols such as “freedom,” “justice,” “the will of God,” or “father-
land.” Indeed, a life worth living, as Socrates emphasized, is not merely
biological but a meaningful life. Meaning makes organisms willing to
die for their ideals, religious and ideological beliefs, leaders, etc. The
symbolic world of meaning and symbolically mediated group identity
is demonstrably “worth” more than mere life. As Wilke (1990:29) argues
on behalf of a sociological systems theory, such phenomena cannot be
explained biologically:
At the level of individuals (mental systems), political or religious
martyrs of all kinds prove that for people, the preservation of
an idea (i.e., a certain form of meaning) can be more important
than the preservation of their own lives. The same happens
when believers do not touch religiously tabooed animals or
plants, although they starve to death. Further examples could
be found effortlessly. On the level of social systems, especially
‘crusades’ of all kinds, point to the fact that groups or whole
peoples put their existence at risk in order to preserve or enforce
certain contexts of meaning such as religion, political values, or
moral postulates. That individual and social levels are as a rule,
inextricably intermingled is shown by one of the most ludi-
crous human inventions, war. For mere ideas (like fatherland,
freedom, communism), millions of people let themselves be led
to the battlefield. This is no longer biologically comprehensible.
As previously noted, the setpoint of an autopoietic system, its homeo-
static parameters, is the preservation of its organization. For organisms,
this means the preservation of their life. In the case of meaning systems,
or systems organized based on a semiotic code, the system functions to
maintain meaning and not to maintain any biological processes. There-
fore, we must conclude a higher level of emergent order is involved.84
84 In this context, Roth’s (1987:269-70) criticism of Maturana can certainly be agreed
with: “It is thus factually incorrect to equate life with cognition. What neurons
do does not enter into the maintenance of themselves. Simply said: what the
muscle cells of my heart and the glandular cells of my liver do has a direct eect
From Systems to Actor-Networks
72
According to Luhmann (1984:141-2):
Meaning is the very ‘substance’ of this emergent level of evolu-
tion. ... It is completely mistaken to seek a ‘carrier’ for meaning.
Meaning carries itself by making its own reproduction self-ref-
erentially possible. ... There are indeed highly complex evolu-
tionary preconditions of meaning formation, but there is no
privileged carrier, no ontic substrate of meaning.85
The leap from genetically organized biological systems to a semiotically
coded system of meaning does not imply, at least this is the theory’s
claim, that we must leave the conceptual terrain of systems theory
and involve concepts outside of the systems model. On the contrary,
much more than in any reductionism, the strategy of systems theoret-
ical analysis seems to lie in indicating the level of emergent order and,
thus, the framework within which systems concepts can be applied and
specified. So, at the sociological level, “every social contact is conceived
as a system up to society as the totality of the inclusion of all possible
contacts” (Luhmann 1995:33). As is true for the transition from physical
to biological systems, so it is true for the transition from biological to
on the maintenance of the existence of my organism; but for the activity of many
billions of neurons, which are active while listening to Bach’s music and at the
same time thinking about its complex structure, this does not apply. It is, after
all, the characteristic of the cognitive activity of the brain that, if only in some
way the continued existence of the organism is secured, it is released from the
obligation to promote survival. The autonomy of the brain is quite essentially
a release from the maintenance of existence: the brain can occupy itself more
and more with things that have only very indirectly or nothing at all to do with
survival (or even work against it in the long run). This is precisely the basis of
the specic performance of human cognition, namely constitution of reality and
thus the possibility of action-planning, i.e., of doing something that is not yet of
use to the organism.” Roth concludes, “Cognition creates – contrary to Matur-
ana’s view – something that does not remain on the same ontological level as
autopoiesis.” Which of course does not mean that cognition has nothing to do
with the organism. The organism together with the nervous system forms the
necessary environment of the cognitive system (= sense system). The nervous
system, although its operations do not make sense, could be said to be integrated
into the meaning system as an environmental condition.
85 Interestingly cybernetics and systems theory have from the very beginning pro-
pose the development of articial intelligence, that is, intelligence not based
upon a biological substrate.
Basic Concepts of Systems Theory 73
meaning systems that all previous basic concepts of system theory are
carried over to the higher level. It is also the case, and this must be
emphasized, that all concepts developed at lower levels of emergent
order must be reinterpreted at the higher level, i.e., interpreted from the
code that is the organizing principle of that specific level.
This implies that the meaning system, like all systems – but in its own
particular way – can be understood from the functional perspective.
Meaning systems, like all systems, function to reduce complexity. As
with all systems, meaning does this by constructing a system/envi-
ronment difference. This constitutive difference allows the system
to construct its own elements, which are signs, and build up internal
complexity through functional differentiation. Like genetically orga-
nized systems, semiotically organized meaning systems are autopoi-
etic in that they generate their own elements and structures from the
network of their own elements. They are operationally closed in that
they respond to environmental stimuli based solely on their organizing
principle, i.e., the semiotic code. Concepts such as adaptation, viability,
and learning can also be reiterated at the semiotic level. Nevertheless,
the claim of systems theoretical discourse to provide an interdisci-
plinary terminology through which the difference between the natural
sciences and the humanities or between mind and matter can be over-
come remains problematic.86
Generally, concepts such as “observation,” “cognition,” “self-reference,”
“communication,” etc., are applied not only to systems of meaning but
to all systems. The “problem solving” activity is a case in point when
86 This is evident, for example, in Bateson’s (1987:113.) systems theory, where
there is a tendency to blur essential distinctions. Bateson lists six criteria of mind
that are explicitly indierent across mechanistic, biological, and semiotic levels
of emergent order: “1. A mind is an aggregate of interacting parts or components.
2. The interaction between parts of the mind is triggered by dierences.... 3. The
mental process requires collateral energy. 4. The mental process requires circular
(or even more complex) chains of determination. 5. In the mental process, the ef-
fects of dierences must be conceived as transformations (i.e., coded versions) of
preceding events. 6. the description and classication of these transformational
processes reveal a hierarchy of logical types intrinsic to phenomena.” We will
return to Bateson’s conception of mind and to this problem of the universality
claim of systems theory below.
From Systems to Actor-Networks
74
this is understood to be extended to machines in computer science,
cognitive science, artificial intelligence, and even down to single-cell
organisms. However, it must also be noted here that all systems theo-
retical concepts must be understood on the specific level of emergent
order that is being modeled by the theory. These concepts have different
meanings at different levels of emergent order, and they depend on the
model invoked to explain them. “Communication,” or “information,”
for example, mean something different at the physical level than at the
biological or semiotic level.87 We introduce these terms at the level of
systems of meaning to indicate that it is at this level that they have their
primary sense. “Cognition,” “observation,” “intelligence,” and “commu-
nication” are thus primarily terms applicable to systems of meaning, and
only in an analogical or derivative sense do they describe biological or
mechanical systems. The failure to take into account the different levels
of emergent order as the framework for interpreting systems theoretical
concepts has been a significant source of confusion and controversy.
87 Michael Levin is particularly liable to equivocate on such concepts as agency
and intelligence applying them as he does to physical and biological systems
from subatomic particles up to human cognition. These concepts run the risk of
becoming so general and abstract as to be useless. What do the “intelligence” of
subatomic particles, complex molecules, an amoeba and a human being have in
common?
Chapter 2
Meaning as a System: Limits and Possibilities
of the Systems Approach in the Social Sciences
The following considerations are conceived within the framework of a
general explication of systems-theoretical approaches to the problem of
the observer, meaning, and communication. Since Niklas Luhmann can
be considered the central figure in this area, we will focus on Luhmann’s
theory of social systems. Our guiding question will be: To what extent
and in what form can we speak of meaning as a system? As stated at
the outset, the universality claim of the systems paradigm stands or
falls with the extent to which systems theory can explain itself, that is,
explain the science of the social and thereby answer the question of
what society is. The question of whether meaning can be understood as
a system will, therefore, necessarily include a second question: Can the
science of society itself be understood as a system?
2.1 Observation
The observer and observation have already been discussed in various
ways and different contexts. This was done without a precise definition
of the term and to show how a framework for discussing the observer
has been inherited from systems models on lower levels of emergent
order. The previous discussion of biological reductionism is a case in
point. The upshot of that discussion was that the observer could not
observe itself as a biological system because observations presuppose
meaning, and meaning cannot be derived from organisms’ adaptation
or structural coupling. Whatever the observer might be, it cannot be
an organism. Now is the time to take a closer look at what or who the
observer is and what observation means within a broader framework of
possible system models.
As stated in the beginning, a system comes into being in the first place
From Systems to Actor-Networks
76
because of a difference. Where do differences come from? Who makes
the distinctions? This question is answered mainly in connection with the
term “observation.” The introduction of a distinction is an observation.
Systems exist only for an observer. Systems are “constructed” by obser-
vation. The observer of physical systems is not a physical system but the
physicist, just as the observer of biological systems is not an organism but
the biologist. This is so even if the physicist and biologist are personally
convinced that they are merely matter in motion or autopoiesis, respec-
tively. We wish to introduce the term “observation” formally and offi-
cially at the level of meaning systems. The specific construction of differ-
ences that constitutes any system takes place on the level of emergent
order of meaning, and this alone should rightly be called “observation.”88
In the most general sense, observation occurs whenever something is
distinguished from something else. Without distinction, there is no
cognition. Cognition, even in its most minimal or basal sense, implies the
distinction between system and environment and some form of self-ref-
erence, that is, the distinction between self and other. Only upon this
basis can the system “know” what information is for it and how it will
be constructed to enable operations. Observation, therefore, is always
cognition, and all cognition occurs only through observation. As we saw
above, life can only be thought of as cognition in a derivative sense, that
is, a sense derived from meaning. Organisms do not think of what they
are doing as “cognition,” only biologists do. Something can be known
“as” something only by being different from something else and from
everything else it could be known as. This is a fundamental principle of
meaning. Heidegger (Being and Time §§32-33) speaks of this principle as
the “hermeneutical ‘as’” and defines it as the fact that everything that
appears can only appear “as” this or that. The hermeneutical as means
88 Thus Luhmann (1990b:683): “In the medium of meaning...being dependent on
observational distinguishing and designating is always already built in.” Be-
cause it is not built in on the levels of physical and biological systems it must be
added, often inexplicitly, to the conditions of possibility of systemic order. There
arises because of this an ambiguity between ontological and epistemological sys-
tems theory. When it was claimed at the outset of our exposition that all systems
are constituted by certain principles this description omied the observation.
Now we can add to these general principles the constitutive function of observa-
tion since all systems are constituted on the level of meaning.
Meaning as a System 77
that Being is contingent; everything can be otherwise. If something can
only appear “as” this, it could also appear “as” that. Being is contingent
since it refers to what it is not and what could be. We know that some-
thing is and what something is only if we know what it is not or what it
could be, i.e., if we know what it is different from. Luhmann (1990:44-5)
speaks of a “difference technique” in this context:
Even within meaning-processing systems, as in living systems,
one’s own autopoiesis must be secured as a matter of priority.
That is: the system exists only if and as long as the meaning-pro-
cessing of information continues. The structural technique that
makes this possible can be called difference technique. The
system introduces its own distinctions and, with the help of
these distinctions, records states and events which then appear
as information for the system itself. Information is thus a purely
system-internal quality. There is no transfer of information
from the environment into the system.
Citing Maturana, Luhmann (1995:36) defines observation at the level
of general systems theory as an empirical/factual operation of a system
and defines it as the “handling of distinctions.” In this, Luhmann, like
Maturana, has been influenced by George Spencer Brown’s Laws of
Form (New York 1972).89 Spencer Brown places an arbitrary operation of
distinguishing – “draw a distinction” – at the beginning of everything.
The distinction, however, must then be coupled with an “indication.”
Who or what is drawing the distinction and making the indication is
not designated. There is no system reference carrying out the operation
of distinguishing. The operation can be attributed to an “observer.” It
can be understood as an operation of “observing” only because of other
distinctions, e.g., the distinction between operation and system and
between observing, observer, and observed. Of course, one can speak of
observing only if an observer and observed are thought together and if
observing as an operation is distinguished from what is operated upon,
and operations are distinguished from states, and so on. Again, it must
be kept in mind that these concepts are meant to be generally applicable
89 For the discussion surrounding Spencer Brown see Baecker (1993a, 1993b).
From Systems to Actor-Networks
78
to all systems and not only meaning systems. Nonetheless, only in a
derivative sense can observation be attributed to biological systems.
The differences that biological systems introduce and the acts of obser-
vation that construct these differences are only available on the level
of meaning. It is meaning that decides to whom or what these terms
should be applied. After the emergence of meaning, all biological and
physical phenomena are integrated into meaning. There is no particle
or field, no atom or molecule, no organism that is not contextualized
and constrained by meaning. The semiotic code is more complex and
variable than the physical or biological code and integrates these lower
levels of order into it. As Heidegger would say, Being is meaning.
It can now be asked what comes first, the distinctions or the operations
that are said to introduce the distinctions in the first place. Without
operations of observation, there are no distinctions, and without
distinctions, there are no operations. What comes first? Is there first
an observer or an observation? To answer this question, one could
argue that selection, relationing, and steering constitute any system
as a system and construct the operations and elements. Observation,
therefore, is an effect or result of system formation. First, the system is
there, and then it can perform operations of observation. On the other
hand, one could ask how the foundational system/environment differ-
ence could arise without observation. Nothing outside the system could
be thought of as a source or cause of observation before the system’s
self-organization through selection, relationing, and steering. According
to systems theory, complexity can only be reduced by these principles.
The moment self-reference is constructed using the distinction between
system and environment and the closure of system operations, the
notion of the observer who operatively introduces distinctions becomes
available. Identifying an observer is a product of distinctions, such as
that between observer and observed, operation and state, system and
environment. Therefore, the question arises of under what conditions
(i.e., distinctions!) are distinctions attributed to the operations of an
observer who itself is constructed on the basis of distinctions?
The fact that systems theory, in the form proposed by Heinz von
Meaning as a System 79
Foerster, Maturana, Varela, and Luhmann, conceives of distinctions as
operations and then defines operations as the “factually” accomplished
observations of an observer does not necessarily follow from the
assumption that the constitution of meaning occurs through construc-
tive articulation of differences. Instead, it represents a momentous deci-
sion to base specific differences on all others as “guiding differences.”
It is in the tendency of this decision to conceive of meaning as either
the product of an observer who has become a “subject,” standing at
once and consequently paradoxically outside and inside the boundary
of meaning, or as the product of an undifferentiated and therefore
unthinkable, anonymous operating. If systems theory, on the one hand,
does not want to hypostatize the meaning system into a “subject” in the
sense of modern epistemology and metaphysics, and/or, on the other
hand, fall prey to the irrationalism of a blind primal act in the sense
of Spencer-Brown’s “draw a distinction,” then it must find a plausible
model for the articulation of differences, or in other words, it must offer
a good explanation of the origin of meaning.
2.2 Self-Reference
If the system applies the difference technique, i.e., the distinguishing
operation of observation, to itself, then we have self-observation or self-ref-
erence. An autopoietic and, therefore, recursively operating meaning
system must do this; otherwise, it would not be able to know which
events are its own preceding operations and thus connect following
operations to them in a continuous chain of meaning. But what “self”
does self-reference refer to when everything, including the environment,
has meaning? There is no system that is meaningful on one side, a kind of
subject, and the world that is not meaningful on the outside as Kant was
forced to postulate to locate meaning within a knowing subject. Those
who criticized Kant’s dualism argued that the subject must distinguish
between itself and the environment (the thing in itself) within itself. For
meaning systems, the environment is constructed within the system
as the necessary reference to an “other,” enabling reference to “self.”
Self-reference and other-reference are mutually dependent. There is no
From Systems to Actor-Networks
80
self-reference without reference to an other, which is not the self. The
system must be able to observe itself, i.e., to distinguish itself. But it must,
therefore, differentiate itself from what it is not. Like all systems, it distin-
guishes itself from its environment and constructs itself by replicating or
modeling the system/environment difference within the system. Thus,
Luhmann can say (1995:36-7), “self-observation is the introduction of the
system/environment distinction within the system, which constitutes
itself with the help of that distinction.” In this way, a self-perception or
an “identity” is constructed.90
Luhmann (1995:33) defines the term “self-reference” as a form of
self-constructing, an operation that constructs unity by means of
distinction.
The concept of self-reference designates the unity that an
element, a process, a system is for itself. ‘For itself’ means inde-
pendent of observation by others. The concept not only defines,
but also contains a significant statement, for it maintains that
unity can only come about through a relationing operation, that
it must be produced and that it does not exist in advance as an
individual, as a substance, or an idea of its own operation.
There are three different forms of self-reference, depending on whether
it is elements, processes, or the system as a whole that is referred to as a
self in each case. At the level of system elements, there is a minimal form
of self-reference, which Luhmann (443) calls “basal self-reference:”
“Basal self-reference is the minimum form of self-reference without
which autopoietic reproduction of temporalized systems would be
impossible.” To produce itself autopoietically, the system must recur-
sively connect its operations to its own operations. Metabolic processes
in an organism, for example, must connect to other metabolic processes
such that the same processes are autopoietically re-produced and
homeostasis can be maintained. In this sense, all autopoietic, operation-
90 See Luhmann (1990b:482): “This designation (observation, description) of the
system by the system itself we wish to call reection, and to note the dierence
from a mere generation of the unity of the system (seen by an external observer),
in the case of reection we speak not of unity but of identity.”
Meaning as a System 81
ally closed systems can also be called self-referential:
One can call a system self-referential if it itself constitutes the
elements that compose it as functional units and runs reference
to this self-constitution through all the relations among these
elements continuously reproducing its self-constitution in this
way. In this sense, self-referential systems necessarily operate
by self-contact; they possess no other form of environmental
contact than this self-contact. (1995:33)
The difference between systems of meaning and biological or mechan-
ical systems is the way the system “refers” to itself. It is not merely that
operations are referred to system operations but that there are different
system “identities” to which operations are referred. The system builds
up internal complexity by distinguishing levels of self-reference. Basal
self-reference directs meaning to meaning. On a different level, agency
or ascription of meaning to a subject of some kind is another form of
self-reference. And finally, the system as a whole can be referred to. This
is the level at which the system-environment difference is thematized.
As soon as we are concerned with system processes as interactions
with an environment, that is, with the re-production of the system/
environment difference within the system, we are dealing with the
specific form of self-reference peculiar to meaning systems as totalities.
The environment of a meaning system, this was the critique of Kant’s
dualism, cannot be outside the system but must be a form of meaning
within the system.91 Meaning systems construct themselves by reintro-
ducing the system/environment difference into the system. Following
Spencer-Brown, Luhmann (1990b:83) speaks of “re-entry.” Thus, the
91 Homan (2019) for example buys into Kantian dualism and biological reduc-
tionism by insisting that internal information construction implies that “reality”
is an illusion. Because illusion, or “simulation,” is a comparative concept only
meaningful in distinction to reality or to something that is being simulated or
modeled, there must be a reality in itself, a noumenon, which we cannot know,
but can still somehow simulate and model. But if we can’t know the original,
what sense does it make to talk about a simulation or a model? How do we know
if it is a “good” simulation or a bad one? Homan can refer to natural selection
and claim that as long the system exists it must be viable and that’s all we need
to know. This is, however, a tautology, that asserts only that what exists exists.
From Systems to Actor-Networks
82
system refers to itself as a whole, which it can only do by simultaneously
“referring” to the environment, which must also have meaning and is,
therefore, inside the system and not outside. As a reference within the
system, the environment has meaning. Paradoxically, the environment
is included in the system and thought of as “belonging” to the system.
This unusual inclusion of the environment within the system causes
problems for the articulation of the system’s primary code. From a logical
perspective, pure self-reference is tautological: A is A. To articulate
further distinctions and not cause the meaning system to be eternally
locked into repeating A is A, the original tautology must be de-tautolo-
gized: A is non-A. De-tautologization, however, creates a paradox. This
paradox must be de-paradoxed to allow the construction of further
distinctions and thus allow further operations of the meaning system.
This process requires a special “semantics of de-paradoxification”
typical of articulating the primary code of a meaning system. We will
have occasion to discuss these issues and what a “primary code” means
in detail below. The point to be emphasized now is that self-reference
and other-reference (environmental reference) are simultaneously and
interdependently constructed within the system:
...unlike nervous systems, structures and processes that employ
meaning can include system boundaries and environments,
which take on meaning within the processes of a self-referential
system (not: in themselves!), so that such systems can operate
internally with the difference between system and environ-
ment. For all internal operations, meaning enables an ongoing
reference to the system itself and to a more or less elaborated
environment…. (1995:37).92
92 See Luhmann (1984:601): “We will speak of reection when the distinction be-
tween system and environment is at the basis. ... In this case, the self is the system
to which the self-referential operation ascribes itself. It takes place as an opera-
tion by which the system designates itself in distinction from its environment.
This occurs, for example, in all forms of self-representation....” For a formulation
that emphasizes external reference, see Luhmann (1990b:306): “But since sys-
tems that are operatively bound to meaning and closed on this basis cannot nd
the notion of a meaning-free environment – they would have to be able to do this
themselves – they can only refer to their environment internally, that is, only in
the medium of meaning.”
Meaning as a System 83
The reflection of the system is necessarily totalizing and universal
because observation cannot go beyond or behind the difference between
system and environment or subsume it under differences that somehow
lie further back. Thus, the meaning system constructs itself equiprimor-
dially with constructing a “world view.” Heidegger (Being and Time)
speaks in a similar sense of a non-subjective “understanding of being,”
which simultaneously discloses itself and the world as a whole. The
understanding of being forms the horizon of meaning of every self-ref-
erence, and one cannot go beyond it. As that being which is concerned
with itself, Dasein must be able to refer to itself and to what it is not at
the same time. There is always self-reference and other-reference within
the one all-embracing horizon of Being. However, to be able to refer
to oneself, one must already have interpreted oneself as something,
i.e., interpreted oneself as a being of a certain kind. Everything else is
interpreted as being of another kind. In this respect, the phenomeno-
logical descriptions of “world” that the early Heidegger employed are
not irrelevant to systems theory. Luhmann, interestingly, does not refer
to Heidegger to explain the concept of meaning but, as we shall see
in a moment, to Husserl. We will ask whether a certain distortion or
limitation of the concept of meaning in systems theoretical discourse
might not be due to this dependence on Husserl instead of Heidegger.
2.3 Meaning
A system of meaning constitutes itself, that is, internally differentiates
itself, including the system-environment difference, through successive
operations of distinction, i.e., observation and self-observation, in such
a way that a “world” is opened up, a world in which the system itself
occurs as something distinguished through self-reference/reflection.
This entire web of differences, distinctions, referents, and relations,
which from a systems theoretical point of view has the function of
reducing complexity in a certain way, is what we call meaning.
In previous discussions, we have spoken several times of meaning as
the organizational form or code of a specific level of emergent order.
From Systems to Actor-Networks
84
We have generally noted that meaning as a fundamental concept in
systems theory is not to be understood, as is usual in the philosophical
tradition, as a property or product of mind or of a subject – regardless of
how mind or subject might be conceived. Within systems theory, mind
is conceived of functionally as a reduction of complexity through a
system/environment difference constructed by internal differentiation
of the system. Subjects are one form of self-reference of the meaning
system. What is peculiar about meaning is that the meaningful distinc-
tion between self and other is drawn within the system as a sign. The
code, the organizing principle of meaning, thus maps the system semi-
otically.
There is no meaning without signs or symbols in the broadest sense
of these terms. Or if there should be such, we could neither talk about
it, think about it, nor know it in any way. As the idealist critique of
Kant argued, we can talk about what we cannot know and, therefore,
it is known. The sign, therefore, has no other, no unsignified, no other
from which it could be distinguished and in terms of which it could be
defined. The function of the sign consists precisely in determining the
“what” of possible thinking and speaking. Meaning differs from what is
often referred to as “sentience” or what is typical of animal “conscious-
ness” or “cognition” as described, for example, by Maturana, Levin,
Hoffman, and many others. If biological systems theory equates cogni-
tion with life, then signs are regarded as mere signals or triggers without
semantic reference.93 A single-celled organism may be processing
information from its environment to sustain its autopoiesis, but it is not
constructing meaning. If we do not want to blur the difference between
semiotic and genetic coding, and thus the difference between signs and
signals, we must hold that at the level of meaning systems, observing
or distinguishing is necessarily accompanied by “signifying.” Distinc-
tions that are not or could not be named, or as Spencer-Brown would
say, “indicated,” are not meaningful distinctions. Naming is mean-
ing-making. In terms of systems theory, we can say that “language”
reduces or organizes the infinite complexity of the primal substrate into
93 It is at this point that much current discussion on “free will” is generated.
Meaning as a System 85
meaningful contingency. The “hermeneutical ‘as’,” which Heidegger
understood to be how Being appears as meaning, implies that what-
ever appears can appear otherwise; that is, it appears against an infinite
horizon of possible interpretations but cannot appear as anything other
than meaning.
The “subject” or bearer of the operation of introducing differences and
constructing relations is the system itself, which organizes itself “auto-
catalytically” as a system of meaning. It articulates a primary code and
opens up a symbolic order. The opening up of a symbolic order – called
“world” for short – is done by communication. We thus equate semiotic
complexity reduction with communication. Meaning and communica-
tion are closely connected. Communication “makes” meaning. Without
communication, there is no meaning. Accordingly, the conditions of
meaning will have to be sought in the conditions of communication,
not the operations of the brain or some mysterious entity called “mind.”
Systems theory is not a theory of mind in any sense in which this word
has been interpreted throughout Western philosophy and especially in
European modernity since Descartes and, above all, in contemporary
cognitive science. Systems theory does not talk about subjects and
objects and wonder how they epistemologically relate to each other and
what they might ontologically be as entities.
These definitions and assumptions mark our position regarding the
status of systems theory in the social sciences. They allow us to conceive
of systems of meaning as systems of communication and not merely as
cognitive systems or psychological systems, which inevitably brings
with it traditional assumptions about human individuals that do the
thinking, know the world, speak, act (freely or not), are morally respon-
sible (or not), are deluded (or not), are living in a simulation (or not),
etc. Systems theory, we propose, is not bound to the assumptions of
humanism, Western philosophy of substance, or cognitive science. Its
potential goes far beyond the predominantly modern Western frame-
work in which it is mainly used. This opens up an approach to analyzing
systems of meaning free from the presuppositions of Western moder-
nity. However, this thesis is not uncontroversial. Luhmann himself
From Systems to Actor-Networks
86
seems to uphold traditional views by attempting to distinguish between
meaning and communication, with communication representing only
one particular form of meaning. For Luhmann, there are two different
and mutually exclusive kinds of systems on the level of emergent order
of meaning: psychic systems or human individuals and their brains,
and social systems constituted by communication and the exclusion of
human individuals and brains. In what follows, we will attempt to argue
for a systems theoretical analysis of meaning that places itself squarely
on the side of language and communication and leaves non– or pre-lin-
guistic psychological systems to the neuro and cognitive sciences, that is,
to lower levels of emergent order than meaning.
2.4 A Critique of Luhmann’s Theory of Meaning
Luhmann regards meaning as the basic concept of a systems theoretical
approach to sociology. Despite the sociological perspective of his theory,
he defines the concept of meaning so broadly that meaning also appears
to be constitutive for non-social and non-communicative systems,
that is, for “psychic” systems. Luhmann is to be credited with clearly
assigning meaning to a level of emergent order in its own right beyond
life. His theory of meaning as a level of emergent order beyond the
physical and biological levels thus claims general epistemological and
philosophical relevance. Meaning is not something that living systems
do. Brains don’t make meaning. Meaning alone makes meaning, and it
makes nothing else than meaning, just as life operates to maintain life
and nothing else. Everything that has long been considered under the
term “mind,” that is, rationality, thought, consciousness, experience,
emotions, beliefs, etc., in short, what traditionally characterizes the
human as human is now to be conceptualized as a form of systemic
order. Nevertheless, systems theory inherits the legacy of philosophy,
its questions, assumptions, and even its methods to a certain extent.
Therefore, there is an ambiguity in the theory, which results from the
fact that traditional ways of conceptualizing meaning derive from and
are unavoidably associated with the subjectivist ontology and episte-
mology of modern philosophy of mind.
Meaning as a System 87
From Descartes to Husserl, to talk about meaning meant to talk about
a conscious subject. Even if this subject, as in the case of Hegel, is no
longer the human being, but Geist, the world spirit, and as in Kant, no
longer empirical but transcendental, that is, no longer merely individual
but universal. It was generally accepted and, for the most part, still goes
without saying that whatever is referred to by language is an object of
subjective intentionality, an act of knowing by a cognitive agent. Within
this frame of thought, the idea of a “system” of meaning almost ines-
capably implies a subject, literally as that which “underlies” all thought
and meaning. In modern thought, this subject was “transcendental”
if it preceded experience as a condition of its possibility. Or it was
“empirical” if it was the object of an experience of self, a self-reference, a
person with their own inalienable experiences. Starting from Descartes’
mind/body dualism, modern philosophy understood mind to be either
an individual human being or, as in cognitive science, a brain. In its
transcendental form, mind was a universal principle of reason, or a
collective intelligence, a super-self like God, which grounded all beings.
If it should happen that individuals or brains communicate with each
other by means of language, then language was conceived of as a cogni-
tive tool. Subjects could use language to express their mental states
or choose not to communicate and remain within a purely subjective
world of inner experience.
Communication, therefore, was almost inescapably understood to be
the act of a subject and attributed to a subject who is not only a subject of
free will (someone who makes choices, for example, the choice to speak
or not) but who is also in some way the source of meaning, the instance
that, as Husserl puts it, “constitutes” meaning in intentional acts. The
consequence of this tradition is that meaning is inseparably associated
with consciousness, which in turn is inseparably associated with an
“I,” an “identity,” a subject of meaning. The materialists claimed that
the I is an epiphenomenon of neural activity and, therefore, a kind of
illusion. In contrast, the idealists struggled to maintain and distinguish
the universal consciousness of reason from the particular consciousness
and identity of unique and often irrational human individuals. When
systems theory proposes to conceptualize meaning in its own right, it
From Systems to Actor-Networks
88
must carefully maneuver through the obstacle course of the modern
philosophy of the subject. This is exactly what Luhmann attempts to
do. Below, we will examine whether he succeeds or fails and what the
consequences of this attempt mean for the systems paradigm.
The dependence of systems theory on modern philosophy of mind can
perhaps most convincingly be illustrated by the fact that Luhmann
(1995:60ff) bases his theory of meaning on the phenomenological
analysis of meaning found in Husserl’s theory of consciousness. This
momentous decision arises not from any internal necessity of the
systems model but from the above-cited fact that modern philosophy
of mind is inseparably associated with consciousness and the idea of
the subject. For modern philosophy, there is no meaning that is not the
meaning of a subject. All consciousness is consciousness of something
by someone. However, neither consciousness nor the subject are funda-
mental concepts of systemic order. Despite this, Luhmann does not
attempt a theory of meaning based on systems theory. Instead, he tries
to exploit an ambiguity concerning the status of the subject in Husserl’s
thought in order to propose two different kinds of meaning system.
Husserl’s phenomenology distinguishes between the empirical and
the transcendental ego, or what could be understood as the individual
mind, and the collective or universal mind. Luhmann transforms the
transcendental ego into society and communication, while leaving the
empirical ego to psychology. He then bans empirical individuals into
the environment of the social system, leaving society to consist exclu-
sively of communication and not consciousness. Let us take a closer
look at this theoretical maneuver.
Luhmann (1995:60) begins by pointing out that Husserl describes
meaning as a totality or “world” of interconnected references.
The phenomenon of meaning appears in the form of a surplus
of references to further possibilities of experiencing and acting.
Something is in the focal point, in the center of intention, and all
else is indicated marginally as a horizon for an ‘and-so-forth’ of
experience and action. In this form, everything that is intended
Meaning as a System 89
holds open to itself the world as a whole, thus guaranteeing
the actuality of the world in the form of accessibility. Reference
actualizes itself as the standpoint of reality. It refers, however,
not only to what is real (or presumably real), but also to what
is possible (conditionally real) and what is negative (unreal,
impossible).
Overlooking for the moment the reference in this citation to “experience
and action” and “intention” to which we will return, if reference can
be said to “actualize itself,” then it is not necessarily something that is
constituted by a subject, that is, by something other than the system of
references as such. In other words, the system is the actor. If the system
is the actor, this implies that consciousness and meaning are not neces-
sarily the same thing, at least, as long as we are talking about a personal
consciousness and not some kind of world-consciousness. Conscious-
ness seems necessarily to imply a subject that is conscious of itself. This
is the subject of intentionality, the subject that knows that it is aware of
an object. The object is the reference, which appears against the horizon
of further possible references. Meaning is not in the mere fact of inten-
tionality. Intentionality means that all consciousness is consciousness of
some object that is necessarily “associated” with other possible objects
in a world or against the horizon of a world. It is these associations
or references that constitute meaning. Meaning is not some particular
thing of which we may be aware of, but necessarily the horizon of other
possible things which could become “thematic” for us, that is, objects of
other intentional acts.
Luhmann cites Husserl’s Ideas of Pure Phenomenology and Phenomenolog-
ical Philosophy I and Experience and Judgment. The upshot is that any anal-
ysis of meaning always presupposes meaning and thus must articulate
itself as self-description. For this reason, Luhmann thinks he must start
from phenomenology, an unprejudiced description of experience. But
is the experience of meaning the experience of a self-conscious subject?
Or does it merely show an open horizon of references, a world in which
subjects may also appear?
From Systems to Actor-Networks
90
Meaning refers to further meaning. The circular closure of these
references appears in its unity as the ultimate horizon of all
meaning: as the world. ... Husserl outlined this situation in the
metaphor of the ‘horizon,’, without thoroughly analyzing the
self-reference of all meaning. All proofs of this statement must
already presuppose it; they have no other way of operating
than by reflecting on the world within the world. We begin by
phenomenologically describing the experience of mean and
the nexus of meaning and world constituted simultaneously,
not basing this description on the underlying existence of an
extramundane subject (which everyone would know existing
in oneself as consciousness) but understanding it as the self-de-
scription of the world within the world. (Luhmann 1995:69)
The “subject” of this description is what Husserl called the “transcen-
dental Ego.” It is a pure consciousness that is somehow, as already for
Kant, related to the empirical ego, the self that I know myself as an
individual to be, but is nonetheless not a person, an individual and thus
coterminous with the entire world. There is no necessary implication
that we can have meaning without conscious subjects, only that when
meaning emerges, it also has consciousness within it. Meaning comes
first and is foundational, whereas subjectivity arises within meaning, as
we shall see, as a form of self-reference. So Luhmann can write (1971:28):
The concept of meaning is to be defined primarily, that is,
without reference to the subject, because the latter, as meaning-
fully constituted identity, already presupposes the concept of
meaning.
In this definition of meaning as a system of mutual references against a
horizon of further possible references, if there is a subject, then it is one
reference among all the others. In this way, Luhmann attempts to make
the phenomenological concept of meaning fruitful for non-subjectivist
oriented systems theory. His program is to reinterpret the concept of
meaning and, thereby, the entire modern subjectivist philosophical
tradition in terms of five fundamental concepts of systems theory:
Meaning as a System 91
1) reduction of complexity
2) system/environment dierence
3) self-reference
4) autopoiesis
5) operational closure
Reduction of Complexity: Functionally, as with all systems, meaning
is concerned with reducing complexity. The event or operation of
meaning selects something as an intentional theme or object from a
horizon of possibilities. When we talk about something, we mean this
or that within a specific context or frame of relevance. We thereby select
something particular out of all possibilities. The phenomenological
concept of intentionality, which is the act of a conscious subject, is here
reinterpreted in terms of the systems theoretical concept of selection,
which is an organizing process of a system. When something is selected
as a content or object of intention by observation and distinction, all
that the object is not is also simultaneously co-thematized. If a hammer,
to use an example from Heidegger, is selected/thematized, then at the
same time, all that is not a hammer becomes a co-thematized horizon,
i.e., there arises a relationing of references or possibilities of selection
referring to all the other tools, to the workshop, to the various mate-
rials that are to be worked upon, to construction plans, to the people or
machines doing the work, even to the possibility of not having to work at
all, and so on indefinitely. These relations are directed or steered toward
possible goals. Every event of meaning involves, therefore, selecting,
relationing, and steering. These system-building processes reduce
complexity in a certain way. World, that is, the meaningful world, as a
system, is always a form of reduced complexity. We encounter and deal
with things like hammers in a world where hammers, tools, workshops,
tables, etc., all have their place. Without the totality of these references,
a hammer could not appear “as” a hammer at all, and no meaning
would be possible.
From Systems to Actor-Networks
92
The distinction between actual and possible and the accompanying
associations and relations can be understood as a reduction of complexity
in two ways. First, specific contents are selected and explicitly thema-
tized, and second, specific other contents are accordingly related as
possibilities for further references. Beyond this, many other possibili-
ties are excluded as irrelevant.94 Not everything possible in the world
is included in the horizon of a typical workshop where things like
hammers have their natural home. Unless I am a physicist, when using
the hammer, I do not consider the electromagnetic field and the strong
and weak forces within the nuclei of atoms making up the hammer.
Not everything in the world appears relevant for further distinctions
and observations. Only those objects and references compatible with a
world in which hammers exist in certain ways and for certain purposes
appear. Thus, the thematization or observation of an object, such as the
hammer, takes place within a certain horizon of possible distinctions
and observations. This horizon and all its possibilities is not something
an individual subject creates. It is what the system does, and this horizon
appears as historically given. A carpenter’s workshop in a medieval
village has a different horizon with different references and possibili-
ties of action than a modern furniture factory. Indeed, the language that
at any time is spoken and the entire social order in which this language
is spoken is not created by the individuals who speak it and act within
the possibilities that the particular world they live in offers. Heidegger
called this fact of human existence “being-in-the-world,” whereby
the “world” is always a historically given world into which Dasein is
“thrown.” Luhmann interprets this as the result of systemic operations
of selection, relationing, and steering, which reduce complexity. The at
any time relevant horizon of meaning arises from the transformation or
reduction of indeterminate complexity, wherein anything and every-
thing is possible, into the determinate complexity of a historical world.
Out of the chaos of all possible observations and distinctions, only some
are selected:
With each and every meaning, incomprehensibly great
94 This is known as the “framing problem” in AI, cognitive science, and philosophy
of mind. See for an overview hps://plato.stanford.edu/entries/frame-problem/.
Meaning as a System 93
complexity (world complexity) is appresented and kept avail-
able for the operations of psychic and social systems. ... ...every
specific meaning qualifies itself by suggesting specific possi-
bilities of connection and making others improbable, difficult,
remote or (temporarily) excluded. (Luhmann 1995: 60/61)95
System/Environment Difference: If meaning reduces complexity, it
can be seen as a particular way of drawing the system/environment
difference. The system-constitutive difference between indeterminate
complexity/contingency and determinate complexity/contingency is
accomplished in systems of meaning by means of those special distinc-
tions that constitute observations of others and observations of self. As
we have seen, all autopoietic, operationally closed systems must refer
their operations to themselves to ensure that what they do is directed
toward ensuring their autopoiesis. Autopoiesis means that a system
must operate to ensure its own operations. Unlike genetically coded
systems, where the constitutive difference between system and envi-
ronment is an actual physical boundary such as a membrane, skin, etc.,
and the internal relationing of the elements determine system opera-
tions which aim at maintaining and reproducing these structures via
autopoiesis and homeostasis, semiotically coded meaning systems do
not have “real” boundaries. Just as the boundary of a living system
must be something living, the boundary of a meaning system must be
meaningful. Contrary to living systems, however, where the boundary
is physical and physically separates the system from an external envi-
ronment, the boundary of a meaning system is inside the system since it
is meaningful, and therefore, distinguishes the system from an equally
meaningful environment, that is, an “internal” environment. For
meaning systems, the system/environment difference is itself something
meaningful, and thus within the system and not a real boundary sepa-
rating the system from an external environment. Outside the meaning
boundary, there is nothing, not even nonsense, because even nonsense
has some meaning; otherwise, we couldn’t talk about it. Nothing, and
95 Again within cognitive science and AI this is discussed in terms of what is called
the “framing problem.” See Anderson et. al. (2022), who argue that cognition can
be understood as a solution to the framing problem.
From Systems to Actor-Networks
94
not even “nothing,” is excluded from a meaning system. For meaning
systems, it can be said that the outside is necessarily inside. The world
horizon of possible references is infinite. Who can say what is not in
some way possible or thinkable? The very idea of the horizon suggests
endless extension and not bounded totality. We cannot see beyond the
horizon. Nevertheless, the world of meaning is not infinite complexity.
Every system must construct its relevant environment. Our air condi-
tioner existed in a world of only a few relevant temperature differences.
For the frog, only quickly moving black patches across the visual field
are relevant information, not the concert in the nearby park. Meaning
systems always emerge together with an environment that is a struc-
tured complexity, if for no other reason than that the environment must
also have meaning and thus be “contained” within the system. Every
religion, worldview, or philosophy claims to circumscribe all that exists
conceptually and to understand the world in its totality.
Meaning becomes [for meaning systems AB/DK] the form of
the world and consequently overlaps the difference between
system and environment. Even the environment is given to
them in the form of meaning, and their boundaries with the
environment are boundaries constituted in meaning, thus refer-
ring within as well as without. ... The system’s differentiation
with the help of particular boundaries constituted in meaning
articulates a world-encompassing referential nexus, with the
result that it becomes possible to ascertain when the system
intends itself and when its environment. But the boundary itself
is conditioned by the system, so that the difference between the
system and its environment can be reflected as a performance
by the system, that is, thematized in self-referential processes.
(1995: 61/62)
It is thus a unique characteristic of meaning boundaries that they
include in themselves all that is outside. “Nonsense” also makes
sense, for otherwise, we could not think or speak about it.96 Meaning,
96 It is interesting to note that all religions and cultures have a problem with “oth-
er” Gods, beliefs, ways of life, etc. They are usually judged negatively as “unbe-
Meaning as a System 95
according to Luhmann (1995:60), is a concept “without difference,” that
is, a concept “which includes itself” and therefore cannot be defined
by opposition since it contains all differences and negations within
itself. This generates paradoxes. How can something be A and non-A,
meaning and non-meaning? The paradox, however, and the resulting
need for “de-paradoxing” are typical of systems of meaning.
Self-Reference: These considerations lead to the third way in which
systems theoretical foundational concepts apply to systems of meaning:
self-reference. Not only does the emergence of meaning imply reducing
complexity and system constitution, but meaning is also self-referential
in various ways. First, every actual unit of meaning, i.e., every thema-
tized content, in turn refers to itself as a possibility, i.e., by thematizing
something, all other possibilities, as well as it itself, remain in view. The
hammer, for example, can be thematized only if it can become thema-
tized again with all the associated references to nails, wood, etc. If one
turns away from the hammer to get some nails, the hammer doesn’t just
disappear. The meaning unit is always theme and horizon together. In
short, we can talk about the hammer only if we can talk about it again
and in many different ways. Translated in terms of systems theory, this
means that every distinction occurs as a possibility within the horizon
in which it itself is contained:
Every intention of meaning is self-referential insofar as it also
provides for its own reactualization by including itself in its
own referential structure as one among many possibilities of
further experience and action. (Luhmann 1995:61)
The possibility of mere reactualization is one form of self-reference,
which Luhmann calls basil self-reference. Meaning, however, is also
self-referential in the sense of being a thematic content itself. Self-refer-
ence of this kind can be called “reflection” since it reflects upon itself as
in a mirror. The system “sees” itself and thus constructs an identity. The
hammer, of course, is not conscious of itself. But the meaning system
lievers,” “ignorant of the true God,” “abandoned by God,” “primitive,” or their
followers are considered not to be human at all. See Krieger (1991b/2006) for a
discussion.
From Systems to Actor-Networks
96
that selects the hammer becomes, in a certain sense, the subject of this
action, and thus something that the system also thematizes. After all,
when hammers are being used, someone, if even a robot, is using them.
When the system selects the hammer, it simultaneously selects itself,
but not as in basil self-reference, when it refers to the hammer, but
in the ascription of agency. This fact has generated much theoretical
effort in modern philosophy. For some, such as Descartes and Kant,
this reference was interpreted as a reference to the ego or the res cogi-
tans in Descartes’ terminology. But what was this thinking subject? Was
it the personal ego or something greater, something universal just as
reason is universal and not a property of me? The “I” that, according
to Kant, must necessarily accompany all perception and knowing was
the transcendental ego and, in some sense, the empirical ego as well.
Whatever is known, it is simultaneously known that “I” know it. Who
else? But which ego is it? The empirical ego of me as an individual
human being, this particular carpenter, or the universal reason in which
I somehow participate? The individual mind or the collective mind?
From the systems theoretical point of view, the subject, whether empir-
ical or transcendental, whether individual or collective and universal,
is a construction of systemic self-reference on the level of meaning. The
subject must, after all, mean something. Meaning systems are self-ob-
serving and thus distinguish themselves from the environment by a
self-description which, as all information, is constructed within the
system as an image or model of the system:
...the self-description of a social system [is] a semantic figure
in which is expressed, in reduced and simplified form, the
specificity of the system. The reduction and simplification is
necessary because an operation of the system cannot refer to
the real system as a whole; for then this operation would in turn
belong to the system, thus would have to refer to itself at the
same time, and so on in an infinite regress. A self-description is,
so to speak, the core of an identity, an inner model or self-model
of the system in the system, which is not directed at the most
comprehensive possible representation, but at an adequate
version of the basic operative structure. (Wilke 1991:135)
Meaning as a System 97
What Willke calls “an adequate version of the basic operative structure”
of a meaning system is what traditionally has been called “conscious-
ness,” which was always a mystery and today has become even the
“hard problem” of cognitive science. The self-referentiality of the
meaning system produces particular logical and semantic problems,
which we will discuss in detail below. Logically, these are the prob-
lems of “de-paradoxing,” for the system cannot be reproduced within
itself without contradiction. If the system observes itself as “I,” who is
the observer of the “I”? If the observer is also “I,” we are trapped in a
tautology. “I” is “I”. If we de-tautologize the I, we get a paradox, since
I becomes “not-I.” The paradox must be de-paradoxed for system oper-
ations to continue and not be blocked by the contradiction. The system
must find a way out of the tautologies and paradoxes of self-reference
to enable connecting operations. This is at least a logical requirement of
the autopoiesis of the system.97
Finally, there is a third form of self-reference insofar as the system as a
whole is considered the agent. This is the place traditionally occupied by
universal reason, the transcendental subject, or God. It could be called
the metaphysical self-reference of the system. If you ask meaning what
it is, it will answer Being, and if you ask Being what it is, it will answer
meaning. At least, this is what Heidegger heard it say. The question of
agency and ascription of observing to an observer is answered by refer-
ence to the system as a whole and not to any instance, such as a subject
or a cognitive agent within the system. On this level of self-observa-
tion, there is no subject. The system operates by constructing meaning.
The system is the actor. Heidegger will say, “language speaks,” and
Luhmann will repeat, “communication communicates.” There is no one
doing this. Neither language nor communication are in any sense of the
term subjects. On the contrary, subjects, persons, cognitive agents, egos,
etc. are not only particular distinctions within the system, but they know
that they are not the creators of the “world” in which, as Heidegger
97 The fact that semantic and pragmatic requirements of autopoiesis do not let
themselves be blocked by purely logical problems goes, of course, without say-
ing. No one goes through life without contradictions and no religion or world-
view is logically consistent.
From Systems to Actor-Networks
98
puts it, they are “thrown.” No inner-worldly subject ever created the
language it uses, the social order in which it lives, the institutions that
govern its life, and the culture and religion that determines its horizon
of possibilities. Let us recall that Descartes’ systematic doubt and the
wish to overthrow all tradition and begin anew based on certainty
could not get the job done without the help of God. The system is the
agent, and the system is not a person, an individual, or an empirical
ego. It may, however, be thought of as a kind of super subject or as God.
Indeed, this was what theology and much modern philosophy does.
Kant’s “transcendental ego” was the function of systemic self-reference
previously occupied by God and was then replaced by pure reason
and, finally, in the theory of social systems, by society. Durkheim, as
is well known, referred to society as a “secular god.” Since Durkheim,
systemic, or metaphysical self-reference has become the historical
and empirical a priori of society about which post-structuralism and
post-modernism happily theorize. On this third level of self-reference,
we have anonymous processes of making distinctions, constructing
information, and nothing else. This type of self-reference is important
in Luhmann’s theory of social systems because, as we shall see, it will
be the social system in the form of communication that is the primary
reference for meaning and not psychical or cognitive systems, as is the
case even today for philosophy of mind and cognitive science.98
The status of the second form of self-reference, that is, the self as a human
individual or person, the ego, remains ambiguous in Luhmann’s theory
since, on the one hand, Luhmann defines psychic systems, that is,
brains and individual human beings, as a kind of meaning system, but
on the other hand, must exclude them from the social system. Luhmann
98 Cognitive science, with the exception of “non-Cartesian”, or “4E” cognitive sci-
ence, as well as philosophy of mind largely ignore sociology and this to their
own detriment. They bar themselves from exploiting the theoretical advances of
sociological systems theory and are locked into biological or machine models.
When they speak of “mind” they mean either brains or computers, but acknowl-
edge nonetheless that minds must be somehow connected and that cognition is
somehow distributed and collective and that there must be some kind of mind,
what kind?, where?, beyond individual brains or computers. Many refer at this
point to the pure consciousness of Eastern spirituality. One could get the impres-
sion that many cognitive scientists are Buddhists.
Meaning as a System 99
seems unable to do away with human beings as producers of meaning
but cannot explain what meaning they produce. The moment they open
their mouths and say something, they become socially constructed
“persons” (persona) through which something else speaks and are no
longer a-social individuals. It is difficult not to think of Hobbes’ descrip-
tion of the state of nature in which isolated humans roamed the Earth
continually at war with each other until they entered into the social
contract and became citizens. We will return to this typically modern
problem and its many ramifications later.
Autopoiesis: According to Luhmann, meaning is autopoiesis par excel-
lence because the processing of meaning via the distinction between
actuality and possibility is a constant selecting, relationing, and
steering, which is nothing other than self-organization or autopoiesis.
If one is working with a hammer, one is looking for the nails, posi-
tioning the wood, following the blueprint, etc. As soon as one thing
is in the spotlight, many others are waiting just off stage. As Husserl
and Heidegger well knew, meaning is the temporal processing of actu-
ality and possibility. Luhmann reinterprets the dynamic restlessness of
human knowing and consciousness in systemic terms as autopoiesis. In
the operations of the meaning system, each thematic content decays in
time, thus forcing further selections that run along the referential lines
disclosed by the relevant horizon:
…meaning processing constantly shapes anew the mean-
ing-constitutive difference between actuality and potentiality.
Meaning is the continual actualization of potentialities. But
because meaning can be meaning only as the difference between
what is actual at any moment and a horizon of possibilities,
every actualization always also leads to a virtualization of the
potentialities that could be connected up with it. The instability
of meaning resides in the untenability of its core of actuality,
the ability to restabilize is a provided by the fact that every-
thing actual has meaning only within a horizon of possibilities
indicated along with. And to have meaning means that one of
the possibilities that could be connected up can and must be
From Systems to Actor-Networks
100
selected as the next actuality, as soon as what is actual at the
moment has faded away, transpired, and given up its actuality
out of its own instability. Thus, one can treat the difference
between actuality and possibility in terms of temporal displace-
ment and thereby process indications of possibility with every
(new) actuality. Meaning is the unity of actualization and virtu-
alization, of re-actualization and re-virtualization, as a self-pro-
pelling process (which can only be conditioned by systems).
… The auto-agility of meaning occurrences is autopoiesis par
excellence. (Luhmann 1995:65/66)
As we shall see, the autopoiesis of the social system is communication.
One communication must lead to other communications, or the auto-
poiesis of the social system stops, and society disappears.
Operational Closure: The fifth systems theoretical resource that
Luhmann calls upon to conceptualize meaning is operational closure.
Meaning becomes a fundamental concept in systems theory insofar as it
is understood as the operational closure of a particular kind of system:
Meaning always refers to meaning and never reaches out of
itself to something else. Systems bound to meaning can therefore
never experience or act in a manner that is free from meaning.
They can never break open the reference from meaning to
meaning in which they themselves are inescapably implicated.
(1995:62)
If a meaning system were to experience “something” meaningless,
which often happens in life, then its autopoiesis would be interrupted,
and the system would disappear. We know this is a possibility from
various forms of psychopathology and “crises” in life. Precisely for this
reason, however, the meaning system directs its operations toward itself
and makes meaning out of whatever occurs. This is what psychiatry
and psychotherapy do. Everything can be interpreted as a “symptom,”
a sign of some psychopathology. No operation of the meaning system
can produce what is meaningless; it codes everything as meaning. It
has always been remarked that whenever a natural catastrophe occurs,
Meaning as a System 101
one seeks for reasons, whether it is God’s wrath, the authorities’ negli-
gence, imperfect warning systems, or irresponsible behavior. Whatever
disaster could be the end or limit of the system becomes a possibility
within the system. The contingency of the system is transformed into an
operation of the system and absorbed into the system. The meaningless
appears “as” (see the hermeneutical “as”) a particular symbolism of
the limit of the system. For a system of meaning, the meaningless has
indeed a meaning, namely, the meaning of a limit, against which one
can only push but not get beyond. As Luhmann (1995:62) puts it: “Any
attempt to negate meaning on the whole would presuppose meaning,
would have to occur within the world. Thus, meaning is an unnegatable
category, a category without difference.” This is the operational and
informational closure of the meaning system.
Operational and informational closure shows what “world-complexity”
means. It appears as a boundary of meaning, as “unmarked state”
(Spencer Brown) of all possible world-orders, or what we have previ-
ously called the absolute complexity of environment-1, often symbol-
ized as “chaos.” But it is also the system-relative environment (environ-
ment-2), the world as it is accessed within any given meaning-boundary.
This is “my” world, “our” world.99 In this world, the world we at any
time live in, specifically and structurally reduced complexity emerges.
Whatever world we live in, it is always and necessarily, in many ways, a
familiar world, a world we know and understand despite all the absur-
dities and injustices of fate. The absolute complexity, or contingency of
environment-1 from which all possible worlds can emerge, raises the
question of why there is something rather than nothing and the utopian
hope that the world will be completely different in the future. Relative
contingency is concerned with how this or that particular event or idea
99 On the concept of world in general in Luhmann, see Thomas (1992). Specically
on the distinction between environment-1 and environment-2 (Thomas speaks of
world-1 and world-2), see 346: “In the world (W1) of indeterminable complexity,
the introduction of system/environment dierences leads to the emergence of
system/environment-relative worlds of (system)relative determination. ... This
cosmic event of meaning unfolds meaning as determination of world (W1) by a
multiplicity of diering, multidimensional, but determined and system-relative
actualizations of sense.
From Systems to Actor-Networks
102
can be integrated into the world we live in and not with the contingency
of the whole world. Both forms of contingency are “dealt” with by the
operational closure of the system. If God should suddenly appear and
announce the end of the world, prophets would arise everywhere,
declaring, “I told you so.” And God or the prophets would be obliged
to say something about what comes next and what we should do in
this situation. Even God, at least if He wants to be understood, cannot
escape the operational closure of the system; there is no way out.
To summarize, a meaning system, according to Luhmann, is an opera-
tionally and informationally closed, autopoietic, self-referential system
constructed by a meaning-specific system/environment boundary
whose function – like all systems – is to reduce complexity. However,
for meaning systems, this is done via observations and self-observa-
tions, that is, by operations on its specific level of emergent order.
2.5 SubsystemicDierentiationorInternalComplexityof
the Meaning System
Based on a systems-theoretical reception of Husserl’s phenomenolog-
ical analysis of meaning, Luhmann specifies how the system of meaning
constructs internal or subsystemic differentiation, that is, internal
complexity. Internal complexity, as we saw, determines the setpoints or
goals of the system and possible states of the system, which can become
responses to events in the environment influencing the attainment of
these goals. When dealing with meaning systems, this raises important
issues concerning the commitment of systems theory to a specific
understanding of society and human existence, that is, to a particular
worldview. We will ask if the systems theoretical model of society, at
least in Luhmann’s version, is not in fact a model of modern Western
industrial society, that is, a culturally and historically limited perspec-
tive. And if the systems model of meaning turns out to be bound to a
description of Western modernity, what consequences does this have
for the universality claim of the theory? Could another kind of society
than Western industrial society, a society which Luhmann describes as
Meaning as a System 103
functionally differentiated, also be modeled as a system? Is meaning,
therefore, only something that has emerged in the modern period of the
West? What about possible forms of social order that are not adequately
described as systemic? We will leave these questions unanswered for
the moment and give an overview of Luhmann’s attempt to theorize
how meaning differentiates and specifies itself within “the” world.
For Luhmann, the “self-processing” of meaning occurs via mean-
ing-specific information. According to Luhmann, information at the
level of the emergent order of meaning is an event that changes states
in a meaning system. The states of a system are changed according to
possibilities that are predetermined in the system’s structures.100 The
weather report, for example, predicts a storm. Based on this informa-
tion, the school authorities declare that the school will be closed that
day. The weather report is information because it makes a difference;
it leads to the action of closing the school. Suppose everyone already
knows the storm is coming and all preparatory measures have been
taken. In that case, the report is a difference; a storm differs from sunny
weather, but it makes no difference. Therefore, it is not information. For
Bateson (1981:582), whom Luhmann likes to quote in this context, infor-
mation is a difference (distinction, observation) that makes a difference.101
How does information arise, and what does it do? What happens in a
meaning system when information “happens”?
According to Luhmann, the structures by which system states can be
transformed by information are, in the first place, meaning dimensions.
Luhmann finds three such dimensions of meaning: There is the “factual
dimension,” which is structured according to the double horizon of
inside and outside; there is secondly a “temporal dimension,” which is
structured by the double horizon of before and after; and third, there
is the “social dimension,” which is structured according to the double
100 See Luhmann (1995:67.)
101 We will return to the notion of information in the discussion of communication
below. In general, we may assume that the unit of meaning which Luhmann
regards as “symbolic generalization” is information, although it must be kept
in mind that under certain conditions and at certain levels of emergent order
information can be distinguished from meaning.
From Systems to Actor-Networks
104
horizon of ego and alter ego. Each unit of meaning or information is
conditioned in all three dimensions simultaneously:
This frame of reference urges every operation to locate its
intended meaning within the structure of the dimensions and
their horizons. Operations must carry out determinations corre-
sponding to the dimensions – not for the sake of the operation’s
own determinacy, but because otherwise they could not connect
to any other operations. (1995:52)
When Luhmann speaks of “connecting to other operations” he refers
to the system’s autopoiesis. What do these dimensions do to enable
the autopoiesis of the meaning system? Whatever we can talk about,
these are things that happen either in me or outside of me. This is the
“factual” dimension with its double horizon of experience and action.
Experience is what happens to me; it is outside me, whereas action is
what I do. Whatever is said, whether I experience it as a fact or as some-
thing that I do, it is further conditioned by the social dimension since
there must always be “others” who can respond in some way, agree or
disagree with what I say, in short, talk about it. And finally, whatever is
said, whether it is a fact or an action, and whoever says it, ego or alter
ego, takes place necessarily in time, in a sequence of before and after.
We must always know what happened first, who spoke first and what
was said afterward, and whose turn to speak is now. Without applying
these three dimensions to information, meaning cannot be processed by
the communication system. No one would know what was being talked
about, how to respond, or in which way to proceed. The system’s oper-
ations could not autopoietically connect to further operations. Commu-
nication would come to a halt, and the system would disappear. It is
precisely to ensure that this does not happen that the three dimensions
necessarily come into play; that is, they are solutions to problems that
the system encounters and not structures of the world either given by
God or otherwise created by some agency.
The dimensions come into play through “attribution” schemes. To reduce
dangerous over-complexity and thus effectively manage the production
Meaning as a System 105
of possibilities of references by building up internal complexity, each
of these three dimensions is “schematized.”102 Schematism operates via
“attribution” and “provides assistance for understanding and simpli-
fication in processing complexes with open meaning that are indis-
pensable for preserving complex systems” (Luhmann 1995:85). The
factual dimension is structured according to the scheme of external and
internal attribution. According to this attribution scheme, internal and
external horizons are determined to be domains of either experience or
action. If information is attributed to the external horizon, it takes place
within the horizon of what is perceived as outside of us. We attribute
the event to the world outside. It is raining. The environment is active;
I am passive. This is what experience means. To experience something
implies that something happens in the world, and I experience it. I see,
for example, the book in front of me on the table. I attribute the book
to the external horizon. It is not part of me. It exists in the world and
presents itself to me as soon as I open my eyes. Whatever comes in from
the outside is the world of experience. On the other hand, if information
is attributed to the inner horizon, then we are dealing with action. I go
over and pick up the book. I do something. I may be upset by the fact
that it is raining or be happy. The system is the agent; the environment
is passive. For example, I close the book and take a sip of coffee. This
was my decision and my action. Or if I close my eyes and imagine the
book lying before me on the table, the book isn’t doing this; imagining
something is my action. Whether drinking coffee or imagining things,
the system is the actor. The imagined book is attributed to the inner
horizon; it is an image, a product of the imagination. Drinking coffee,
on the other hand, is something I do, and neither the cop nor the coffee
experience this or act on their own.
102 The reader is inevitably reminded of Kant’s notion of the disordered manifold of
sense impressions which must be ordered by the schemata and categories of the
mind. There are many parallels between systems theory and Kantianism since infor-
mation is constructed within the system out of undierentiated perturbations com-
ing from the environment. Similar to Kant’s unbridgeable gap between what can be
known and the thing in itself, in systems theory there can be no communication, that
is, ow of information between system and environment. Whether this model can
be used to describe meaning systems for whom the environment only exists within
the system is a question which will occupy us at length later in this book.
From Systems to Actor-Networks
106
Second, there is the temporal dimension. The time dimension is sche-
matized according to the attribution of either stability or variability.
Without change, there would be no perception of time. Information
comes and goes, stays a while, or immediately disappears. No matter
what the information may be, whether experience or action, it happens
in time and is ordered sequentially along a timeline. First, it started
raining, and then I got upset. This becomes apparent when we consider
how important narrative is in creating order of all kinds. Our world, our
lives, indeed practically everything, is information ordered narratively.
A story is a series of events along a timeline, and everything occurs
along some timeline.
And finally, the social dimension is schematized similarly to the factual
dimension, that is, by attribution, but in this case, not to causes, internal
or external, but to persons. Whatever is said is said by someone. Infor-
mation, such as the weather report, comes from people, speakers, and
communicative actions; even if I am the only one who says it, at least I
have to say it or be able to say it, or it doesn’t exist. The social dimen-
sion has a troublesome ambiguity because it may be God, Reason,
or Language and not any person doing the “speaking.” Recalling
Heidegger, “language speaks,” and Luhmann himself will say that
“communication communicates.” As we mentioned above, this is the
metaphysical self-reference of the meaning system. Does metaphysical
self-reference follow the attribution schematization of the social dimen-
sion? Do we need the attribution of a speaker? Perhaps we don’t when
the system self-reference is to itself as a whole and not to a subject or
speaker within the system. But when language speaks, the problem
arises of whether anyone should respond or how the conversation
should continue. As Job in the Bible well knew, one should not argue
with God. This problem is solved by avoiding metaphysical self-refer-
ence and attributing information to a speaker in the social dimension,
thereby accepting God’s personalization and all the theological issues
that come with this.
In summary, it is important to emphasize that the differentiation of the
three meaning dimensions listed above and the attributions of informa-
Meaning as a System 107
tion to either experience or action, a temporal sequence, and to different
speakers are required by the fact that the system must somehow punc-
tuate, control, and structure the otherwise unstoppable and diffuse
flow of references. These three dimensions are not the only internal
differentiations of the meaning system. There are also fundamental
distinctions between psychic and social systems, which we will discuss
immediately below, as well as the differentiation of semi-autonomous
social subsystems we will deal with later. For now, let us note that
meaning must be ordered by the various dimensions, schemata, and
attributions to create islands of unity, identity, and information. It is
within the context of this problem that Luhmann speaks of “symbols.”
And it is with regard to the distinction between symbol and sign that
the problem of the semiotic coding of the meaning system arises, as well
as justification for a non-linguistic, psychic system of meaning that is
distinguished from communication.
2.6 Psychic Systems and Social Systems
Apart from the internal differentiation of meaning into the factual, social,
and temporal dimensions, Luhmann proposes a further fundamental
differentiation of the meaning system. This is the distinction between
psychic and social systems. This distinction is central to Luhmann’s
theory of society as a communication system and raises the more basic
question about what meaning as such is. Let us recall that for Luhmann,
the substrate of all system building is a “medium,” and systems emerge
in this medium as “form.” Our supposed primal substrate, the medium,
is pure diversity, or as Luhmann puts it, “loosely coupled” elements.
Form, however, happens when elements somehow become “closely
coupled,” when the elements are selected and related to each other in
specific ways that can then be steered toward goals. Once form appears
in a medium, not everything is possible. Once systems are constituted,
order appears where before there was disorder. To refer to meaning as
a medium in which systems arise is simply another way of speaking
about the specific complexity that system construction on the level
of meaning reduces. Every level of emergent order, the physical, the
From Systems to Actor-Networks
108
biological, and the semiotic, has its specific complexity. The chaos
of all possible relations and references, world complexity, which is
the problem that meaning systems must solve, is reduced by closely
coupling otherwise loosely coupled elements by selection, relationing,
and steering processes. Introducing form in a medium is nothing other
than the reduction of complexity. First, this is accomplished by intro-
ducing the system/environment difference. The next step is building up
internal complexity. For systems of meaning, this is done by observa-
tions, that is, by introducing distinctions within the system.
Of course, the meaning system can differentiate itself however it
wishes. Saussure pointed out that systems of signs and languages are
“arbitrary.” A system of meaning can differentiate itself into expe-
riences and actions, temporal sequences, and social interactions, as
Luhmann proposes. Or it can differentiate itself into the holy and the
profane, domains of the ritually pure and impure, realms of spirits
and demons, ancestors or aliens, animals, plants, humans, and things,
etc. A system of meaning can differentiate itself internally in unlim-
ited ways, as demonstrated by the surrealist classification of animals
that Jorge Luis Borges supposedly finds in a Chinese encyclopedia,
which inspired Foucault’s (Words and Things) archeology of knowl-
edge. According to this source, which bears being cited, animals
are classified as “those that belong to the Emperor, embalmed ones,
those that are trained, suckling pigs, mermaids, fabulous ones, stray
dogs, those included in the present classification, those that tremble
as if they were mad, innumerable ones, those drawn with a very fine
camelhair brush, others, those that have just broken a flower vase,
and those that from a long way off look like flies.” A perusal of the
philosophical literature dealing with mind and current cognitive
science would find a no less surprising and confusing list of mental
states that supposedly make up the realm of meaning, from qualia to
thoughts, perceptions, and judgments on to attitudes, beliefs, percep-
tion, memory, reasoning, and much more. As social and cultural
history shows, societies have been differentiated segmentally or into a
stratified hierarchy, or, as Luhmann for Western modernity proposes,
in semi-autonomous functional subsystems. All these and many other
Meaning as a System 109
possibilities offer themselves when differentiating a meaning system.
With this general caveat in mind, let us follow Luhmann’s description
of the internal differentiation of the system of meaning. The distinction
between psychic and social systems raises the question of whether all
meaning is semiotically coded or only that meaning communicated in
language. And if meaning as such is semiotically coded, what basis is
there for distinguishing psychic systems from communication systems?
Or, if meaning is not semiotically coded, how can we talk about this
supposedly non-linguistic meaning? If, as Wittgenstein said, the limits
of our language are the limits of our world, and we cannot go beyond
these limits, what sense does it make to speak of an extra-linguistic,
purely psychic domain of meaning? For Luhmann, as we shall see, it is
sufficient to ban human individual psychic systems from society. Can
this strategy adequately deal with the problems arising from assigning
meaning at once to consciousness and society?
Contrary to the relativistic perspective that post-structuralist decon-
struction takes toward systems of meaning, the differentiation between
human individuals as psychic systems and society as an all-encom-
passing domain of communication that Luhmann proposes does not
appear as a historical and cultural accident. Luhmann claims that this
distinction is necessary and generally valid for all meaning. We will have
occasion to ask whether Luhmann’s preference for the differentiation
of psychic and social systems expresses more a preference for modern
Western convictions about free individuals over against society than an
imperative of systems theory itself. To approach this complicated issue,
we will briefly examine Luhmann’s discussion of symbols and signs.
For Luhmann, meaning arises from the formation of unity in the
undifferentiated flow of references. This is the most basic operation
of meaning construction. Luhmann calls this the “symbol.” Symbols
account for the ordering or unity of references that together constitute
a unit of meaning:
The concept symbol/symbolic is meant to denote the medium
in which units are formed, the term generalization the units’
From Systems to Actor-Networks
110
function – to handle multiplicity operatively. In very rough
outlines, it is a matter of a plurality being related to a unity and
being symbolized by it. (Luhmann 1995:93)
The plurality in question here seems to be the undifferentiated flow
of sense impressions. If something is to become meaningful, it cannot
remain a momentary perception or unnamed impression, but many
such perceptions must be grouped and generalized to become a
“unity.” This unity is created by a symbol, a name. Out of the flow of
many individual sense impressions of a certain four-legged, barking,
tail-wagging animal, there arises the symbol “dog.” I see a dog. Before
the formation of such units of meaning, one cannot speak of meaning at
all because one has nothing “fixed,” nothing lasting long enough in the
mind’s eye to be attributed to a horizon, whether factual, temporal, or
social. “Symbolic” means that an undifferentiated multiplicity has been
transformed into something meaningful; that is, it has become a distinc-
tion, a difference, a punctuation, and stability in the flow of references.
Symbols are, therefore, always “generalizations” because the manifold
(to use a term from Kant) is thereby given enough stability to be ordered
into a system of references, a “world.” Despite similarities, it must be
emphasized that symbolic generalizations are not to be equated to the
synthesis of an undifferentiated manifold of sensations as in Kant, but
the construction of repeatability and regularity which underlies all
object permanence and therefore all intentionality as consciousness of
something:
Every meaningfully grasped given must not only be fully
present at a moment and thereby ‘fulfill’ experience or action, it
must also organize self-reference, thus, ensuring that, if neces-
sary, it can be made available again in (more or less) different
kinds of situations, at other points in time, with other possible
partners of social communication. This re-availability is built
into concrete experience and action by symbolic generalization.
(Luhmann 1995:93)
The operations of the meaning system construct redundancy, repeat-
Meaning as a System 111
ability, or regularity. Symbolic generalizations thus have the function of
acting as structures of expectation. “Symbolic generalizations condense
the referential structure of every meaning into expectations, which
indicate what a given meaning situation foresees” (Luhmann 1995:96).103
The dog I see, for example, is a unit of meaning that allows and even
imposes certain expectations upon me. I expect, for instance, that the
dog will not fly away, melt into water, or turn into a table, that it is
something I can feed, pet, take for walks, and much more. But with
this, it becomes clear that structures of expectation have the function
of restricting possibilities while at the same time opening up other
possibilities and making them available for connecting operations. I
know at the same time how it “goes on” with the dog and also what
“does not fit.” Stones, for example, are usually expected to be hard,
heavy, opaque, usable as building material etc., but they are not suit-
able for eating, and so on. In the sense of a consistent “systems-the-
oretical semantics,” symbolic generalizations thus reduce external
complexity by constructing internal complexity in the form of symbolic
generalizations as sets of expected references. Because of symbolic
generalizations, not everything is possible anymore. Through symbolic
generalizations, things remember themselves or refer to themselves, as
it were, and remain more or less bound to their specific “meaning.”
Indeed, meaning is constructed by this self-reference. This is the “basal
self-reference” of the meaning system, which we mentioned above as
the first form of self-reference distinct from the subject-reference or the
system-reference.
Because of this basal self-reference of meaning, Luhmann (1995:96)
thinks he has to “replace” the concept of sign with the concept of
symbolic generalization. This is a momentous decision because it
lays the foundation for distinguishing between two kinds of meaning
systems: psychic systems and social systems. The systems theory of
meaning distances itself thereby from semiotics, which equates meaning
with the sign and thus with semiotic coding. According to most semi-
otic theories (Saussure and Peirce), no meaning can be without a sign.
103 See for a discussion of expectations Luhmann (1995:267).
From Systems to Actor-Networks
112
According to Luhmann, however, meaning need not be semiotically
coded. This is because signs, by definition, stand for something else,
whereas meaning is related to itself through symbolic generalizations
and stands, therefore, only for itself.
The reference to the world that is immanent in all meaning
prevents one from defining meaning as a sign. One must care-
fully distinguish between the structure of reference and the
structure of signs. The function of a sign always requires refer-
ence to something specific, while excluding self-reference. It
requires asymmetrization of a basal, recursive self-reference. In
other words, there is neither a sign for the world nor a sign that
indicates itself. But both of these – universality and self-reference
– are indispensable properties of meaning. That is why meaning
is the foundational matter: a sign must have meaning to be able
fulfill its function, but meaning is not a sign. Meaning forms the
context in which all signs are determined, it is the conditio sine
qua non of their asymmetrization. But taken as a sign, meaning
would be able to stand only as a sign for itself, that is, as a sign
for the non-fulfillment of the sign’s function. (Luhmann 1995:71)
If the function of the sign is always to refer to another, to that which the
sign signifies, that of which it is a sign, and on the contrary, meaning
can only refer to itself, then it would seem that meaning cannot be a
sign, or constituted by signs. This implies that the semiotic coding of
meaning is something derivative from a primary form of meaning. But
what, we may ask, is this primary form of meaning if not something we
can talk about, something that is referenced by signs and thus semioti-
cally coded? The attempt to distinguish systems theory from semiotics
is counter-intuitive since it is obvious that signs also refer to themselves
and that language can talk about itself. After all, Luhmann himself is
talking at length about it. Meaning it would seem, can and must have a
sign structure; otherwise, we couldn’t be having this discussion.
In his later work, Luhmann (1993:61-62) proposed a difference-theoretic
or form-analytic approach to meaning, which allows the symbol to be
Meaning as a System 113
understood as a sign:
In cases where the sign itself designates its own function of
unifying what is separated, one could speak of symbols. A
symbol would then be the self-signification of a sign. Therefore,
symbolic signs are not simply guideposts that point to some-
thing else. They are not just carriers of a signifying reference
and, therefore, not only materializations of the signifier. Rather
they also contain a hint of the function they fulfill, that is, a hint
of the very unity-constituting meaning of the sign.
Following this reasoning, there is no fundamental difference between
symbols and signs. No non-linguistic or pre-linguistic symbols could
be available as carriers of meaning before or outside of semiotic coding.
Both symbols and signs must therefore be understood as necessarily
semiotic. If they are to be distinguished at all, both fulfill the criteria
of universality and self-reference that characterize meaning. Contrary
to his earlier view, Luhmann grounds this possibility in the fact that
the “form” (Spencer Brown) of meaning and sign are the same.104 Both
sign and meaning are the unity of a fundamental, non-hierarchical
distinction. Meaning, as we saw, is understood as the unity of the
difference between actuality and potentiality since every reference
refers necessarily to other possible references, and the sign is the unity
of the difference between signifier and signified.105 Saussure (1967)
expressed this unity between signifier and signified as an inseparable
104 On Luhmann’s concept of form, see (1993:49): “The distinction itself, insofar as
it is distinguished from that which is distinguished by it, is then the form.” Luh-
mann is aware that here he goes beyond what can be found in Spencer Brown.
105 See Luhmann (1993:63): “The concept of the two-sided form, however, beer
expresses that the inside of the form, namely the respectively actualized sense,
only makes sense with regard to the possibility of actualizing other possibilities,
and that this presupposes dynamic systems consisting of operations (events).
Meaning is consequently form as boundary, which is always co-observed, but
can never be exceeded operatively, since every operation remains on the inside
of form, namely actualizes meaning. We thus nd in the form of meaning exactly
the same state of aairs that the analysis of the form of signs has revealed: The
operations always remain on the inside of the form. ... The operations which use
the form of the sign always stay on the side of the signier. But they need, to be
able to mark this side, another side. This other side is always present as an un-
marked state.”
From Systems to Actor-Networks
114
bond. So inseparable is this bond that there is no signified without the
signifier, which implies that the order of things, which was supposed
to lie outside the order of signs and give meaning to signs, is merged
with the order of signs. Meaning cannot be derived from things that
are somehow outside signification but consists in the relations of signs
among themselves, that is, the system of references within language.
In the same way, meaning for systems theory is the system of refer-
ences constituting a world. Here, Luhmann, just like Saussure, wants
to overcome the long outdated “realist” conception of the sign, which
assigns the source of meaning to the thing outside of language that is
then supposed to be signified by language.
An operationally closed system using language cannot relate
to the environment at the level of its operations. In this respect,
there is no referent. (Luhmann 1993:51)
If one follows realism and derives the meaning of signs (semantics)
from the things they refer to, then one runs the risk of reifying the
sign as isolated from its meaning-constitutive references to other signs
(syntax) and to the respective uses of signs (pragmatics). Semiotically
and in terms of philosophy of language, such a concept of the sign is
highly questionable because the “suppression of the referent,” i.e., the
subsuming of extra-linguistic things to language, is precisely what
modern semiotics does. According to Saussure (1967), a sign is an
inseparable unity of signifier and signified, whereby the signified, i.e.,
the thing signified, cannot be a real thing outside the differential struc-
ture of language. Signs are said to be “arbitrary” and “unmotivated”
precisely because they do not read their meaning from things. But they
are not arbitrary because their meaning arises from the differential
relations of signs to each other. Saussure is not alone. For Peirce (1983),
the object, or referent, is not something outside the sign or something
other than the sign but one of the three constitutive relations of any
sign: object, representamen, and interpretant; all three together make
up the sign. The object belongs essentially to the sign; it is internal to
the sign and cannot lie around somewhere outside it. From a systems
theoretical perspective, this fundamental assumption of semiotics is
Meaning as a System 115
accounted for by the fact that the environment of a meaning system is
within the system and that the system is operationally closed.
If there can be no meaning without signs or language, then it becomes
difficult for Luhmann to maintain that there are two kinds of meaning
systems: psychic and social systems. If one overlooks the fundamental
connection between sign and meaning, then the consequences for
system-theoretical semiotics are problematic. Meaning, according to
this view, cannot be equated with language since language consists of
signs that are only subsequently connected with symbolic generaliza-
tions that, according to Luhmann, are already developed “in concrete
dealings with objects and events” (1993:90). The distinction between
“symbolic generalizations” on the one hand, which Luhmann wishes
to define as the basic units of meaning, and which can be ascribed to
psychic systems, and signs on the other hand, which Luhmann ascribes
to communication and thus social systems grounds the distinction
between psychic systems and social systems. But if there is no difference
between the meaning that psychic systems construct and the meaning
that social systems construct since both are semiotically coded, how
are we to understand this fundamental differentiation of meaning into
psychic systems and social systems? If there are psychic systems without
language, signs, and communication that also construct meaning, how
do they do this? What kind of meaning are we talking about, assuming,
of course, that we can talk about it at all.
We probably have to think of visual perception here since Luhmann
often refers to the distinction between form and medium from Heider’s
psychology of perception (e.g., in Luhmann 1990b:53ff.) when he wants
to give an example of non-linguistically constituted meaning.106 Thus,
meaning can be distinguished from communication, which is, after all,
language-dependent. Language and communication are, therefore, not
the forms par excellence in which meaning is constructed. According
106 See Luhmann (19995:412): “Perception is a less demanding form of acquiring
information than communication. It makes possible information that does not
depend on being selected and communicated as such.” But, of course, it could be
communicated because it is semiotically coded. Whether it is less demanding not
to say what one sees or is thinking than just blurting it out is an open question.
From Systems to Actor-Networks
116
to Luhmann, communication becomes a particular form of processing
meaning among other forms. This assumption does not change in his
later difference-theoretical semiotics.107 Despite his ambivalence on the
nature of the sign, Luhmann consequently adheres to the distinction
between psychic and social systems, that is, to a fundamental differ-
ence between individuals and society. But we much ask if this typically
modern Western view of free, autonomous individuals existing in a
Hobbesian “state of nature” somehow before and outside of society is
compatible with a systems theory of meaning or whether we are dealing
with ideological commitments to Western modernity that systems
theory does not need. Must the meaning system differentiate itself into
individuals over against society?
Luhmann’s insistence on the typically modern distinction between
individuals and society leads him to assert that semiotics and herme-
neutics cannot provide an adequate theory of meaning. Signs have to be
interpreted, i.e., they require a hermeneutic understanding. But herme-
neutic interpretation presupposes that a unit of meaning is embedded
in another horizon. This is the effect of the social and temporal schema-
tization of meaning. It appears as consensus or disagreement about the
past, present, and future. Gadamer (1989) referred to this basic foun-
dation of hermeneutics as “fusion of horizons” (Horizontvechmelzung).
Because other people and other times have their own horizons of
meaning, this difference must be bridged in present-day understanding
of others and of tradition. This fusion of horizons is what makes up
hermeneutical understanding. Since the social horizon is not available
to isolated psychic systems, Luhmann must claim that the schematiza-
tion of meaning in a social horizon is unnecessary for meaning. This
implies that there is information that supposedly does not need to be
understood. What is this information? According to Luhmann, these are
the mental states of psychic systems. Psychic systems process meaning
in a purely inner, private, signless flow of perceptions and thoughts.
107 See Luhmann (1993:64): “The sign-form is one (emphasis AB/DK) among several
possibilities to translate the paradox of meaning, the indeterminacy of the deter-
minate, into a distinction and thus to unfold it.” What are the other possibilities if
we cannot designate them? Here again, the assumption seems to be that meaning
without signs is possible.
Meaning as a System 117
Meaning is constituted unmediated and directly rather than mediated
through sign use. This idea is evident in Luhmann’s (1993:53) “substitu-
tion” of Peirce’s concept of “thirdness” for the concept of “observation.”
According to Peirce, thirdness underlies the interpretant; it is the sign’s
meaning or how it is finally interpreted. For Peirce, the interpretant
denotes the application of a rule or convention that governs the associ-
ations a sign might bring with it. Observing, however, is not an opera-
tion that applies a rule or convention. Herein lies, we may assume, the
reason why pragmatics, the study of how signs are used according to
rules, is neglected in systems theory.
This foreshortening of the concept of sign in Luhmann’s systems theory
has serious consequences in that it allows Luhmann to postulate two
kinds of systems. Luhmann’s decision to take Husserl’s phenom-
enology with its attachment to the ego/consciousness as a starting
point inherits Husserl’s solipsism and leads his theory of meaning to
distinguish between systems of consciousness, which process meaning
in the form of a purely internal, pre-communicative (and consequently
pre-semiotic) flows of thought and perception, and social systems,
which process meaning in the form of communications.108 According
to Luhmann (1995:98), psychic systems are distinguished from social
systems in that meaning is reproduced in different forms in each case:
Meaning can insert itself into a sequence that is attached to
bodily feelings, then it appears as consciousness. But meaning
can also insert itself into a sequence involving the understanding
of others and then appear as communication.109
108 It is interesting to note that this assumption is foundational for cognitive science
in its various Cartesian forms.
109 Luhmann insists that systems of consciousness and systems of communication
must be strictly distinguished. Indeed, as systems, they are constituted by mutu-
al exclusion, that is, by the system/environment dierence. Thus Luhmann can
say (1990b:32): “Systems of consciousness and systems of communication exist
separately.” Even more precisely (45,46): “Systems of consciousness are opera-
tionally closed, self-dynamic, i.e., restless, mercurial systems of the reproduction
of their own thoughts by their own thoughts. ... ...a communication system is
a system coupled to consciousness, irritated by consciousness, but which can
determine its own operations only by its own structures and its own structures
only by its own operations.”
From Systems to Actor-Networks
118
There are thus two kinds of meaning systems, the psychic and the social,
with the psychic system consisting of a sequence of mental states that
are not semiotically coded and the social system consisting of a series of
linguistic communications.
Meaning can thus be understood independently of language, commu-
nication, and action, as in the Western metaphysics of the subject from
Descartes to Husserl. In this view, meaning theory remains committed
to Western modernity. However, one of Luhmann’s theoretical inno-
vations is to strictly separate the two types of systems and to empha-
size, completely contra-intuitively, that society does not consist of
human beings but of communications. Human beings are banned from
society for the first time in Western thought.110 Nonetheless, there are
no communications without mental states, no social system without
consciousness. Luhmann conceptualizes the relation between indi-
vidual and society according to the scheme of system and environment
and appeals to a biological metaphor.
This is precisely how the chemical system of cells is environment
for the brain and how a person’s consciousness is environment
for the social system. No decomposition of neurophysiolog-
ical processes could ever reach individual cells as its ultimate
elements and no decomposition of social process could ever
arrive at consciousness. (1995:179)
Is this distinction between two forms of meaning systems tenable?
Is meaning conceivable without semiotic organization? Can a purely
private, non-semiotically coded stream of consciousness be conceived
as a meaning system? Is purely “private” observation as the handling of
110 Thus, according to Luhmann (1990b:30-1) in an extreme formulation, “There are
thus no ‘conscious communications’, just as there is no ‘communicative think-
ing’ (sensing, perceiving). Or, to put it another way, human beings cannot com-
municate....” Luhmann (1995:210) sees the main dierence with tradition in that
for “the humanist tradition human beings stand within the social order and not
outside it. ... If one views human beings as part of the environment of society
(instead of as part of society itself), this changes the premises of all the traditional
questions, including those of classical humanism” (212). With these statements
Luhmann clearly takes a “post-humanist” position.
Meaning as a System 119
differences possible? Or can differences only be handled based on semi-
otic organization, which implies communication? Is the specific system/
environment difference that constitutes social systems a difference
between consciousness and communication, or would it not have to be
understood as a difference between external and internal communica-
tive contingency? Even if I don’t talk to others, I am constantly talking
to myself, and for this very reason, I am able to talk to others and they
to me. Even if my feelings and thoughts are my own, they are so consti-
tuted that I could talk about them if I wanted. Is it not so that all systems
of meaning are semiotic systems whose elements are selected, related,
and controlled by semiotic organization? Must we not conclude that
insofar as the human being or consciousness functions as an element
of the meaning system, it functions only, as Peirce said, as a sign or as
Luhmann himself will say, as a socially constructed “person?”
These questions are at the basis of our critique of Luhmann’s theory.
Therefore, in the following discussion, we will attempt to question
the distinction between the two types of meaning systems, and this
must be emphasized, without denying the existence of individuals or
of consciousness or, on the other hand, advocating any form of collec-
tivism in opposition to individualism. Instead, it is a question of the
importance of communication for the constitution of meaning and,
consequently, for the construction of individuality, consciousness, and
society in communication. Contrary to Luhmann, we ask: Can meaning
be adequately described without reference to signs, language, commu-
nication, and understanding? Does not all information, insofar as it is
conceived as distinction and designation, necessarily imply coding, and
does not coding at least semiotic coding – in turn, necessarily imply
communication and understanding? Is not a system of meaning neces-
sarily semiotically organized?
If it turns out that individual consciousness is not even possible outside
the framework of the collectively and socially constituted communi-
cation community, there would be no evidence in systems theory to
postulate two kinds of meaning systems. The two domains would
coincide, meaning that meaning could not be considered pre-social and
From Systems to Actor-Networks
120
pre-communicative. In other words, if systems of consciousness could
not be distinguished from social systems, at least at the level of meaning
constitution, there would be no psychogenesis of meaning indepen-
dent of sociogenesis, no individual without society, no brain without
other brains, and constitutive relations to the world.111 Consciousness,
whatever it might be, becomes much less important for understanding
meaning than it currently is in most discussions of the philosophy of
mind and cognitive science. If it turns out that cognition – also cognition
about one’s own “bodily feelings” – is only possible if it is semiotically
mediated and intersubjectively accessible and controllable, systems
of meaning cannot be divided into two species – one psychic and one
social – which would then somehow be “adapted,” or “structurally
coupled” to each other in a kind of symbiosis.112
The upshot of this discussion is that meaning systems can handle differ-
ences only under the conditions of communication. Communication can
differentiate its ascription of agency however it wishes. The meaning
system can construct any kind of subjects or identities that it wants.
There is no necessity that we must concern ourselves with individuals
and society, with psychic and social actors. At the level of emergent
order that we call meaning, we are never dealing with single individ-
uals and their private, purely internal cognitive states or observational
operations, but always, directly or indirectly, with a system of commu-
nication as an intersubjective whole within which such distinctions as
those between individual and group are used for specific purposes in
specific historical circumstances.
Thus, as a counter-thesis to Luhmann, we claim that meaning must
be conceived of at the communicative level. Communication must be
understood as a foundational condition of meaning, not vice versa.
This does not prejudge whether one can continue to speak of social
111 According to 4E cognitive science, mind is embodied, embedded, enacted, and
extended into the environment. See Rowlands (2010).
112 Luhmann doesn’t speak of symbiosis, but of “interpenetration” (1995:213.) to
describe the relation between psychic systems and social systems. We will return
to the term “interpenetration,” which Luhmann following Parsons takes to refer
to the relationship between humans and society below.
Meaning as a System 121
systems and mental systems as particular internal differentiations
within communication systems. The thesis claims only that systems of
meaning are necessarily semiotically coded systems of communication,
no matter how they may be further differentiated and specified.
To make this thesis plausible against Luhmann’s view, we have to
resort to the linguistic turn in the philosophy of language and episte-
mology. In short, the meaning-constitutive role of communication is
shown by the fact that internal contingency, i.e., the non-arbitrariness
of structures of expectation, is not conceivable outside or before inter-
subjective corrigibility, i.e., through communication. No distinction,
observation, designation, or symbolic generalization is conceivable as
a pre-communicative product of a not-yet-social system. Thus, in place
of the transcendental subject, there are not two structurally coupled
kinds of meaning systems but the linguistically constituted community
of communication.
We assume that units of meaning reduce contingency by creating redun-
dancy. According to Wittgenstein, who is close to Peirce’s pragmatic
semiotics in this respect, symbolic generalizations are to be understood
as “rules” (Peirce speaks of conventions) because a rule cannot be
followed only once. We quote this passage (Wittgenstein 1953:§199)
again:
It cannot be only one time that only one person has followed a
rule. It cannot have been only once that a communication was
made, a command given, or understood, etc. – To follow a rule,
to make a communication, to give a command, to play a game
of chess are customs (usages, institutions). To understand a
sentence is to understand a language. To understand a language
is to master a technique.
According to Bateson, a unit of meaning is constituted by communi-
cation because only information can be perceived. Information is a
difference that makes a difference. Such a difference must be commu-
nicative because it could not make differences without selection, i.e.,
encoding and communication, without which it could have no state-
From Systems to Actor-Networks
122
changing effect for any system. A difference that makes no difference is
not information. For example, if we know what “smoke” is, we expect
that there is fire where there is smoke, and we can respond accord-
ingly. Not everything is possible in a coded world where rules of asso-
ciation and reference apply. The complexity of the world is reduced
to a certain regularity. For Bateson, as for Wittgenstein, the reference/
horizon relation constituting all meaning is constituted by redundancy.
Unlike Luhmann, however, for both Bateson and Wittgenstein, there
is redundancy only through linguistic and semiotic organization. As
Bateson (1972:421-2) says:
I would argue, however, that the concept ‘redundancy’ is at
least a partial synonym of ‘meaning.’ As I see it, if the receiver
can guess at missing parts of the message, then those parts
which are received must, in fact, carry a meaning which refers
to the missing parts and is information about those parts. … In
sum, ‘redundancy’ and ‘meaning’ become synonymous when-
ever both words are applied to the same universe of discourse.
How else do we imagine the possibility of reduction of contingency if
not through language? How could the psychic system organize, condi-
tion, and relate many thoughts, feelings, imaginings, intuitions, sensory
and bodily perceptions, fantasy images, moods, desires, impulses,
etc., if not through a semiotic code? Without semiotic coding, that is,
communication, we have only autism. To disrupt the autopoiesis of
a psychic system, the organism doesn’t need to die. All that needs to
happen is that the operations of the system must become disorganized,
and the system will dissolve into madness. The specific contingency of
consciousness must be reduced and organized by some principle. This
principle can be only language, for no other principles of meaningful
organization remain. As Wittgenstein showed, there can be no private
language; the idea of a self-contained, isolated psychic system is absurd.
Below the language boundary, no consciousness is possible. And above
the language boundary, beyond language, no isolated, purely private,
a-semiotic consciousness is possible. Even the mystics only know about
their silence because they can talk about it, and this is when they say
Meaning as a System 123
it is ineffable. The idea of a pre-communicative, pre-linguistic system
of consciousness is neither empirically nor logically possible. If such
a system of pure, non-communicative consciousness were empirically
present, it would have to be able to make itself known, for which it
would then have to make use of language and communication. If it
could not do this, it would dissolve in the senselessness of an arbitrary
jumble of states of consciousness. Finally, as Habermas points out,
meaningful complexity reduction need not be explained on the basis of
the model of intentional consciousness:
If we now abandon the basic consciousness-philosophical
concepts in which Husserl treats the lifeworld problem, we can
think of the lifeworld as represented by a culturally transmitted
and linguistically organized stock of interpretive patterns. Then
the talk of a referential context that connects the situational
components with each other and the situation with the life-
world no longer needs to be explained within the framework
of a phenomenology and psychology of perception. (Habermas
1981b:189-90)
Let it be said again, we do not wish to advocate collectivism of any
kind by rejecting Luhmann’s’ notion of individual psychic systems. The
dichotomy of individualism and collectivism is a typical structure of
modern Western culture and society. Not having one does not imply
you are automatically stuck with the other. The solution is to have
neither. Our critical remarks on Luhmann’s theory of meaning are not
intended to assert that there are no individuals or that subjectivity in
the sense of a private, emotional, and thoughtful interiority is illusory.
We merely advocate that individuals and society be understood from
an adequate theory of meaning. Such a theory must distinguish levels
of emergent order. It must distinguish between systems of meaning, on
the one hand, and mechanical and biological systems, on the other. It
must be able to account for the difference between genetic and semiotic
coding theoretically. We claim that there is no other form of organiza-
tion for meaning systems than semiotic coding.
From Systems to Actor-Networks
124
If talk of mental and social systems is to have any meaning at all, and
even if only to serve as a fruitful guiding idea for different research
approaches (e.g., in cognitive science, psychology, and sociology), then
both types of systems should be considered as constructs of the one
semiotically constituted meaning system. Accordingly, we emphasize
that distinctions such as those between the psyche and society do
indeed have meaning. They are common and useful distinctions in
everyday understanding and as demarcations between scientific disci-
plines. It is, interestingly, the everyday understanding that Luhmann
draws on when he wants to make his distinction between thinking and
talking plausible: we say something and think something else at the
same time; we cannot perceive another person’s perceptions directly;
we think associatively and only in fragments of words and sentences
and not according to the regular flow of linguistic communication, etc.
All this is just as true as it is irrelevant. Such distinctions have been
conceived for purposes other than to ground a theory of meaning.
They are far too undifferentiated or, as Luhmann says, have too little
“resolving power” to serve as the foundations of a solid theory of
meaning. They presuppose meaning as already given. Accordingly,
they cannot be applied without fundamental reinterpretations at the
level of a theory.
With these arguments in the background, we now turn to the specifi-
cally social form of meaning that Luhmann calls “society.” So far, we
have critically examined the form of meaning-processing that Luhmann
attributes to psychic systems. We have pursued the question of whether
the distinction between two kinds of meaning systems can itself have
meaning. The argument was directed against the idea of an a-semiotic,
pre-communicative meaning and thus against the idea of two kinds of
meaning systems. It will be seen in the discussion below that even in
the specifically social realm, the preliminary decisions of Luhmann’s
theory of meaning lead to problems. If psychic and social systems are so
different, it has to be explained how they relate to each other. How do
they condition each other? We begin by asking how social systems can
arise from psychic systems.
Meaning as a System 125
2.7 The Social System
We saw that one of the first major internal differentiations of the
meaning system was the introduction of three fundamental dimen-
sions of all meaning: factual, temporal, and social. Each dimension is
“schematized” into two mutually dependent horizons. In the social
dimension, all meaning is attributed to either ego or alter-ego. The
social dimension conditions information such that it flows between ego
and alter ego. This is communication. This is also, for Luhmann, the
specific social system in distinction to individual, non-communicative
psychic systems. We have argued at length above against the idea that
there can be individual, non-communicative psychic systems. We need
not repeat these arguments here. Nonetheless, since, for Luhmann, the
emergence of meaning is not equivalent to the emergence of society, the
theory of social systems must explain how communication and thus
society come into being on the basis of non-communicative psychic
systems. He attempts to answer this question in two ways. First, he
refers to Parsons’ sociological theory of double contingency. Secondly,
he proposes a systems theory explanation of how individuals and
society relate to each other in terms of what he calls “interpenetration.”
We will briefly look at both these attempts to explain society beginning
from individuals even though, as argued above, we view this entire
enterprise as misguided and unnecessary.
To explain the specific social processing of meaning, Luhmann first
introduces the model of double contingency.113 Double contingency
is intended to explain how the social as a realm of commonly recog-
nized, i.e., stable, and repeatable structures of expectation emerges
113 In the original version quoted by Luhmann (1995:103.) from Parsons and Shils’
Towards a General Theory of Action, “There is a double contingency inherent in
interaction. On the one hand, ego’s gratications are contingent on his selection
among available alternatives. But in turn, alter’s reaction will be contingent on
ego’s selection and will result from a complementary selection on alter’s part.
Because of this double contingency, communication, which is the preoccupation
of cultural paerns, could not exist without both generalization from the partic-
ularity of the specic situations (which are never identical for ego and alter) and
stability of meaning which can only be assured by ‘conventions’ observed by
both parties.”
From Systems to Actor-Networks
126
from the realm of isolated, pre-communicatively constituted systems
of consciousness. The following considerations are concerned with the
problem of the emergence of the social as a realm in itself in contrast
to the realm of individual, isolated psychic systems. The introduction
of Parson’s theory of double contingency serves to synthesize three
theoretical traditions: Parson’s explanation of the emergence of society,
Husserl’s phenomenological analysis of intentional consciousness, and
systems theory’s requirement that a system must be founded on the
system/environment difference.
The first step in this synthesis is to extend the concept of contingency to
its “original modal theoretical version.”
Something is contingent insofar as it is neither necessary nor
impossible; it is just what it is (or was or will be), though it
could also be otherwise. The concept thus describes something
given (experienced, expected, remembered, fantasized) in the
light of its possibly being otherwise; it designates objects within
the horizon of possible variations. (Luhmann 1995:106).
The idea of contingency is to be understood in terms of Husserl’s notion
of intentional consciousness of an object against a horizon of possible
references or variations. If everything that appears, appears “as” this
or that (the hermeneutical “as”), then it can appear as something else
from a different perspective or in a different context. All meaning is
inherently contingent. According to Luhmann, double contingency has
two consequences:
It enables the differentiation of a particular world dimension for
socially distinct meaning perspectives (the social dimension),
and it enables the differentiation of particular action systems,
namely, social systems. (106)
The concept of contingency, therefore, connects Parson’s double-contin-
gency model with Husserl’s idea of intentionality as the consciousness
of an object against a perspectival horizon. These two together form
the background for explaining a unique system/environment difference
Meaning as a System 127
that constitutes social systems as specific forms of meaning systems. To
what extent is this strategy successful?
Let us base the following discussion on a simple example. Two children
are playing next to each other on a playground. Both have a ball in their
hands. Initially, they both, each separately, engage in an indeterminate
behavioral variation. The ball is pushed back and forth arbitrarily with
the hands, then with the feet; it is thrown into the air and dropped on
the ground or caught with the hands. There is dancing around the ball,
jumping over it, sitting on it, etc. During all this, the children do not
follow any behavior rules nor interact with each other. There is no ques-
tion of a social system yet. Then, one child throws the ball to the other.
The second child also selects this behavior; he throws the ball back to
the first child. This behavior is repeated, resulting in mutual behavioral
selection, which is then stabilized; it can be continued or stopped and
then taken up again. A structure is formed, distinguished from the arbi-
trary variations of behavior that occurred before. One could say that the
social system “ball game” has come into being. The structure called the
ball game, remains even when the children go home for lunch because
when they meet again at the playground afterward, they can resume
the ball game.
This model describes the situation of double contingency in that 1) the
behavior of each child is contingent, free, and undetermined, 2) both
children know this of each other, and 3) on the basis of interaction, both
become dependent or conditioned by each other’s behavior. They have
mutual expectations as to what action they should each perform. Each
child knows that the other knows that it knows what is expected. And
they honor these expectations and accept the “obligation” of acting
according to what is expected. Without mutual acceptance of the rules
of the game, the game would not exist. Without mutual acceptance of
rules governing social behavior, as Hobbes well knew, society could
not exist. Thus, it can be said that a social interaction or a social system
emerges autocatalytically via variation, selection, stabilization, and
structure formation under the conditions of double contingency. But
what does this have to do with meaning?
From Systems to Actor-Networks
128
It is important to note that double contingency presupposes the recog-
nition of the other as alter ego. The ball game, which is a simple model
of society, arises not as a result of solving the problem of double contin-
gency but on the basis of mutual recognition, which makes interaction
and thus double contingency possible. Why play ball with something
that cannot understand what I mean to do by throwing the ball to it?
Luhmann refers to this foundation of the social as “respect” (Achtung),
and he asserts quite plausibly that without recognition of the other as
alter ego, and vice versa, there can be no social interaction, no commu-
nication, and no society. Even when children play with their toys, they
speak to and relate to them as though they understand what is expected
of them and can respond appropriately. One can only have social
expectations with regard to others who are recognized to be capable of
communication. Thus Luhmann (1995:121-2):
Ego experiences alter as alter ego. But along with the nonidentity
of perspectives, ego also experiences the identity of this experience
on both sides. The situation is indeterminable, unstable, and
unacceptable for both the participants. In this experience the
perspectives converge, and that makes it possible to suppose an
interest in negating this negativity, an interest in determination.
Just as in Hobbes’ state of nature wherein all were at war with all, and
there was no common acceptance of any expectations or rules, all parties
realize that this chaotic and stressful situation must be negated and in
some way made determinate, that is, meaningful. Meaning, therefore, is
not only about mutually accepting rules of behavior or honoring expec-
tations. It is also not reducible to Husserl’s intentional thematization of
some content or information against a horizon of possible further refer-
ences. What prevents such a solipsistic consciousness from becoming
arbitrary and chaotic? Let us recall that the word “conscious” comes
from Latin and means “knowing together.” For Luhmann, meaning
must arise as a social system to the extent that information and mutual
recognition become a self-referential, operationally closed system. This
is a system of communication that excludes all that is not communica-
tion. This is society.
Meaning as a System 129
The moment an isolated psychic system recognizes another as like
itself, it is no longer an isolated psychic system, but is transformed into
something else. Thus, there is either a plurality of isolated conscious
systems, which then cannot be conceived as a social unity and there-
fore also cannot stand in a situation of double contingency, or there
is a multiplicity of communicative agents which already are the social
system; they belong to society and not merely to the environment of
society. Accordingly, when the children were playing side by side
before they had started their ball game, it must be assumed that this
behavior was already social. It took place on a playground conditioned
by many relations and references and is not conceivable as an indeter-
minate and unstructured behavior in the sense of a pre-social substrate
of society. The children are already “socialized” to play with balls
before they appear on the playground. The situation of double contin-
gency, therefore, does not model the emergence of communication.
Were the children not already communicative, we would be dealing
with psycho-social pathologies such as autism. Thus, the social does not
emerge from a non-social domain of some purely psychic meaning, just
as Husserl can’t escape solipsism once he begins with a lonely Cartesian
ego. Even Descartes had to call upon God to save him from this unten-
able situation. In conclusion, we must admit that the theory of double
contingency must presuppose that which it is supposed to explain.
If one nevertheless wants to try to use the model of double contingency
as a bridge between pre-socially conceived individuals and the social,
one must think of it behavioristically, i.e., as a model of mere behavioral
coordination or, according to Maturana, as “structural coupling” among
systems which are environment for each other. Indeed, like Maturana,
Luhmann seems to want to ground the social on the non-meaningful
communicatively conceived feedback loop of mutual observation,
which the behavioristically interpreted model of double contingency
depicts. The substructure presupposed in the theorem of double contin-
gency presupposes highly complex meaning-using systems that are
non-transparent and non-calculable to each other:
...they do not understand each other … They concentrate on
From Systems to Actor-Networks
130
what they can observe as input and output in the other as a
system in an environment and learn self-referentially in their
own observer perspective. They can try to influence what they
observe by their own action and can learn further from the feed-
back. In this way, an emergent order can arise that is conditioned
by the complexity of the systems that make it possible, but that
does not depend on this complexity’s being calculated or controlled.
We call this emergent order a social system. (Luhmann 1995:110)
As discussed above, the fact that “highly complex meaning-using
systems” appear to each other in the social dimension of ego and alter
ego means that the problem of the emergence of the social is already
solved in the presupposition. The behaviorist alternative, however,
tries to avoid ascription of subjective states by observing the other from
the outside. Therefore, it cannot explain how the alter becomes an alter
ego and, thus, how society could emerge from the interaction of ego
and alter ego. If two systems observe each other as system-in-an-envi-
ronment, for example, we observe the frog in the pond, and the frog is
watching us too, this does not mean that we are dealing with expecta-
tions of expectations, i.e., with meaning or with society. We could take
the frog as an alter ego, and then we would have a social situation as
we often do with pets and children with their toys. Mutual observa-
tion and behavioral coordination, which happens among organisms in
a niche, do not have anything to do with double contingency or with
the construction of commonly recognized expectational structures.
Behavioral coordination as adaptation occurs between any organism
and its environment continuously, i.e., as long as the autopoiesis of
the organism is maintained. For this purpose, meaning is not neces-
sary. The question arises: Can meaning systems observe each other as
system-in-an-environment so that communication and social interaction
come into being only based on such mutual observation? Or must not
communicatively constituted meaning be regarded as always already
conditioning any observation?
According to the behaviorist view of double contingency, the social is to
be built up only by the interaction of independent organisms with their
Meaning as a System 131
own areas of experience. Two “black boxes” react to each other until
behavioral coordination occurs, which Maturana considers as a struc-
tural coupling or the emergence of a “consensual domain.” This model
can be adopted from phenomenology in that it begins with a non-socially
conceived individual consciousness as a putative system of meaning. But
this only exchanges the problems of biological reductionism with the
problems of idealism. It leads to exactly the same problem that Husserl
was unable to solve, namely the problem of solipsism. How is it possible
to move from a purely private experience to the experience of the world
shared with others in language and communication?
Thus, if it does not fail because of idealism, the attempt to derive the
social as an independent system of meaning from the model of double
contingency leads back to biological reductionism. One has little reason
to speak of systems of meaning as a separate level of emergent order
beyond life. Only if the situation of double contingency is already
conceived as a meaning system and already as communicative can it
serve as a model to explain the emergence of expectations and coor-
dinated action in society, but then we do not need it. The explanation
is already in the presupposition. The recognition of the other and the
necessity of understanding is presupposed so that the two “black
boxes” could even come up with the idea of talking to each other.
Communication is already there as a condition of the possibility of
interaction. Communication cannot be deduced or explained from
isolated conscious agents.
We can now turn to Luhmann’s second strategy to get the social out
of isolated psychic systems. This is the idea of “interpenetration.”
Luhmann (1995:213ff.) speaks of “interpenetration” to designate the
relationship between human individuals and society, i.e., between
psychic systems and communication systems. The term is meant to
answer the question about the conditions of the possibility of double
contingency and, thus, society.
Interpenetration is not a general relation between system and
environment but an intersystem relation between systems
From Systems to Actor-Networks
132
that are environments for each other. … the concept of inter-
penetration indicates a very specific situation, which must be
distinguished above all from input/output relations. We speak
of “penetration” if a system makes its own complexity (and
with it indeterminacy, contingency, and the pressure to select)
available for constructing another system. Precisely in this sense
social systems presuppose “life.” Accordingly, interpenetration
exists when this occurs reciprocally, that is, when both systems
enable each other by introducing their own already-constituted
complexity into each other. (Luhmann 1995:213)
This remarkable statement poses several questions. How can a system
be related to something that is neither itself (self-reference) nor the
system’s environment? What does it mean to say there is a relation
between systems that are environment for each other but that this is
not a relation between a system and its environment? An autopoi-
etic, self-referential, operationally and informationally closed system
cannot communicate with the environment. It can only be perturbed
by changes in the environment and must construct information out of
these perturbations according to its organization. If it does this success-
fully, it can continue its autopoiesis and, for an observer, appears to
be adapted to the environment. How can such a system “introduce
its already constituted complexity (internal complexity) into another
system”? What could it possibly mean, based on general systems
theory, for a system to “introduce” internal systemic complexity into
another system? Furthermore, how can this happen between systems
reciprocally? In other words, how can a system accept, take in, and
process the internal complexity of another system that is environment
for it? And finally, if such a thing could happen, are we talking about
emergence, or are we talking about self-organization and internal
differentiation?
These questions lie at the heart of Luhmann’s theory of social systems
since the concept of interpenetration is called upon whenever it becomes
necessary to explain how information can be exchanged between closed
systems in society. We will see below that the social system differen-
Meaning as a System 133
tiates itself into many different kinds of subsystems, large functional
subsystems such as business, science, education, law, etc. Then, within
these functional subsystems, all sorts of organizations arise, which are
also systems. As general systems theory requires, all these subsystems
are modeled as autopoietic, self-referential, and informationally and
operationally closed systems. And all are merely environment for each
other. If there can be no communication between system and environ-
ment, how can all the different systems of which society is composed
share information and cooperate to build a unified society? The answer
that Luhmann gives is interpenetration. Therefore, we shall take the
opportunity that the problem of the relationship between individual
psychic systems and society offers to examine the concept of interpen-
etration more closely.
Let us take an example to illustrate the conceptual problems that the
concept of interpenetration brings with it. A frog lives in a pond that
is full of other organisms. The frog is adapted to this niche, or more
correctly, structurally coupled to various other organisms and natural
conditions in this niche. For example, the frog is structurally coupled
to the pond’s water and the insects it eats. However, it is not structur-
ally coupled to the concert being performed in the park nearby. In this
situation, is there any “intersystem relationship” that is not a relation
between system and environment and thus a relation of structural
coupling? The answer must be no. All systems in this niche, so long it
can be considered a balanced ecosystem, are adapted to each other and
structurally coupled. We can call this the “normal” situation. Therefore,
it is not surprising that Luhmann must emphasize that the concept of
interpenetration designates a “very specific situation,” a situation we
do not normally encounter.
What is “very specific” about the situation of interpenetration is that
we have to do with systems that somehow “introduce” their internal
complexity into each other, even though there can be no communica-
tion between the system and the environment. The internal complexity
of a system can be understood as the elements that are selected and
related to each other and thus constructed to perform the operations
From Systems to Actor-Networks
134
that the system needs to maintain its autopoiesis. If the system is an
autonomous, self-regulating learning system, it processes environ-
mental perturbations according to its internal complexity to produce
information that steers its operations. We must assume that the internal
complexity that is being made available to other systems in interpen-
etration is information since, otherwise, the systems would have to
eat each other to access their internal complexity. One could suppose
that the fly could make its internal complexity available to the frog for
its self-construction, but this means the frog eats the fly. Only if we
are talking about information can interpenetration mean that systems
can somehow “introduce” internal complexity into each other. The
exchange of information, however, is communication.
The situation of communication is indeed a “very specific” situation.
We do not observe this situation when describing the frog in the pond
or any physical or biological phenomena. We observe communication
only within communication and not as something that happens between
communication systems and their environment. If we suppose that
systems communicate with each other, they can do this only because they
are no longer autonomous systems but elements of an encompassing
communication system. The moment Husserl’s solipsistic intentional
consciousness opens its mouth and says something, it disappears and
becomes a construction of the social system, a necessary ascription of
the communicative act. Communication communicates, and to do this
efficiently, it constructs persons to whom communication is ascribed. In
addition, since society needs a certain communicative contingency so
that the system does not collapse into redundancy and static stability,
it constructs persons as sources of communicative contingency,
uncertainty, and restless construction of information. These persons,
however, are not the psychic systems that Luhmann begins with when
describing the situation of interpenetration. Individual psychic systems
disappear the moment they open their mouths and say something, that
is, try to introduce their internal complexity into each other. Interpene-
tration annihilates itself the moment it occurs. Interpenetrating systems
are instantly transformed into the social system. This leads to the idea,
which Luhmann explicitly suggests, that interpenetration describes
Meaning as a System 135
not the construction of a system using the complexity of another system
but the emergence of the social system from out of non-social psychic
systems.
When we are talking about the social system, we are talking about the
emergence of a system of a fundamentally different kind than anything
preceding it. We have argued above that one of Luhmann’s major
theoretical innovations is that he understands meaning to be a level of
emergent order beyond matter and life. If so, it would seem that inter-
penetration describes a situation in which a new type of system on a
new and different level of emergent order appears. Luhmann explicitly
compares the appearance of the social system from psychic systems
to the emergence of meaning from life. The question, therefore, arises:
Is the relation of the social system to psychic systems the same kind
of relation as that of meaning to life? Does it make sense to compare
the emergence of the social system from psychic systems, which,
according to Luhmann, are both meaning systems, to the emergence of
meaning from life? To say that life makes its complexity available for
the construction of a meaning system is, of course, in a very abstract
sense, plausible. Complexity, as we have seen in the discussion of the
self-organizing universe, seems to give rise to emergent phenomena.
Emergent phenomena are those that cannot be predicted and cannot be
reduced to or deduced from what precedes them. When discussing the
relation of psychic systems to social systems, we are not talking about
the emergence of meaning from non-meaning.
Let us take a moment to reflect on emergence. The emergence of life from
matter or meaning from life cannot be conceptualized as the communi-
cation of internal complexity from one system to another. Information
is not being transferred from one system to another. Instead, an entirely
new form of order arises (emerges) to reduce the complexity that the
existing codes can no longer efficiently reduce. General systems theory
allows us to think of matter and life as forms of systemic order. We
also know that systems are dynamic; they operate and have agency,
and therefore, talking about one kind of system “making its complexity
available” for constructing another kind of system is theoretically
From Systems to Actor-Networks
136
consistent in an inclusive and general sense. Furthermore, if we accept
the principle that every higher level of emergent order “integrates” the
lower levels into itself, that is, life takes matter into its processes just
as meaning takes life and matter into its operations, a certain mutu-
ality or exchange of information is conceivable. The idea of integra-
tion, however, is not interpenetration because it is life, once emerged,
that takes information from physical systems according to its specific
coding, just as meaning, once emerged, takes information from life
according to its semiotic coding. The emergence of semiotically coded
meaning from genetically coded life cannot be understood as a situa-
tion of interpenetration. The relation of psychic systems to society is
not a relation between systems that are environment for each other.
Individuals are not integrated into society; they are constructed by
society for the purpose of reducing the complexity of communication.
The idea of interpenetration implies that the two systems mutually
make their complexity available for constructing each other. Thus, it is a
matter of mutual enabling, a co-evolution, whereby the two systems do
not merge into a single comprehensive system. “The interpenetrating
systems remain environment for each other” (214). They function as a
source of “incomprehensible complexity” (214) for each other in a way
that Luhmann explains in reference to Heinz von Foerster’s “order from
noise” principle.
Thus, one could say that psychic systems supply social systems
with adequate disorder and vice versa … Social systems come
into being on the basis of the noise that psychic systems create
in their attempts to communicate. (Luhmann 1995:214)
Apart from the conceptual difficulties with the notion of any kind of
relation between autopoietic, self-referential systems, which is not
structural coupling, the recognition of the other as alter ego, which is
the basis of double contingency, is sufficient to open up the social level.
And it is difficult to imagine how or why one would recognize an alter
ego if alter didn’t communicate in some way, that is, if both ego and
alter ego were not already within a social horizon of meaning. Double
contingency explains why the interaction of ego and alter ego estab-
Meaning as a System 137
lishes mutual expectations, such as rules that social participants should
follow normatively. In fact, when Luhmann explains how psychic and
social systems can provide complexity for one another, it amounts to
the establishment of norms, that is, “binary schematisms” (229) such as
“conformity and deviance” (229) or right and wrong.
A system may schematize the use of another system’s complexity
as friendly/unfriendly, true/false, conforming/deviant, useful/
harmful, or whatever it wants. The schematism itself forces the
system to admit the contingency of its behavior and thus the
autonomy of the other system. … (230) For the interpenetration
of human beings and social systems, this implies that the social
meaning of an action is judged primarily by whether it corre-
sponds to the norm or not. (233)
Friendly or hostile, right or wrong, conforming or deviating are not
purely private states of consciousness but communicative acts. What
is “provided” here as the complexity of consciousness for the commu-
nication system are precisely not states of consciousness but rather the
actions of communicatively constructed “persons” in society. Schema-
tisms or criteria thus do not have the function of admitting something
merely conscious and non-communicative into the communication
system, but instead, they subsume communicative complexity under
certain expectations or criteria; they include and exclude; they consti-
tute the ways in which contingent and unforeseeable communicative
actions can be made meaningful. Schematisms are expectational and
evaluative structures of the communication system whose function is
to reduce communicative (rather than conscious) complexity and to
permit connecting operations in the communication system. Somebody
does something wrong and is blamed. Somebody does something right
and is praised.
Luhmann admits that one of the consequences of interpenetration is that
psychic systems now appear as subjects, “…binary schematisms are the
precondition for the emergence of the figure that in modern philosophy
has gone by the name of the subject” (233). The subject is characterized
From Systems to Actor-Networks
138
as “having true and false opinions” and “acting correctly and incorrectly
or morally right and wrong” (233). These are the socially constructed
persons we referred to above. It seems, then, that Luhmann is merely
reintroducing double contingency here under the title of interpenetra-
tion. The application of binary schematisms runs through interactions
and mutual observation and serves to specify and build up structures
of expectation. If interpenetration is to answer the question of the possi-
bility conditions of double contingency, then double contingency is
used to explain itself. However, as already shown above, this model
is only useful within the communication system – where commonly
recognized criteria/expectation structures are already present – for
the account of the concretization and differentiation of communication
and not for the explanation of the emergence of a social system from
pre-communicatively conceived systems of consciousness. Can there
be any sense in judging actions as right or wrong if the agent – as a
non-communicative consciousness – could not be told this, i.e., if they
could not have their actions constituted by intersubjective control? If
double contingency is to serve as a model of the specifically social, then
the social must be the precondition and the condition of the possibility
of double contingency. One can only get out of this circle if one admits
from the outset that meaning emerges based on semiotic coding and that
there is no fundamental difference between psychic and social systems.
This does not preclude that a meaning system such as modern Western
society chooses to differentiate itself into individuals and groups.
Finally, it would also have to be considered that the model of double
contingency cannot be justified by the modal logical notion of contin-
gency anyway. In the case of double contingency, it is not a matter of
merely being able to be different but of disappointable structures of
expectation. The difference between environment-1 and environment-2,
i.e., between indeterminate and determinate complexity, is ignored if
one interprets the phenomenological possibility horizon in a modal
logical way. Modal logical contingency is not the contingency that the
idea of double contingency presupposes. Modal logical contingency
does not exclude the indeterminate arbitrariness of environment-1,
whereas the model of double contingency presupposes communi-
Meaning as a System 139
catively applied symbolic generalizations, i.e., the already reduced
contingency of environment-2. The children on the playground are not
somewhere in chaos. They know quite well where they are and what
they can do, which is why there can be a right and wrong way to do
it. Ego expects a particular response to its communications from alter.
This expectation may be disappointed. As a reaction to this, ego may
correct his/her communications. As argued above, we are not dealing
with coordinated behavior or structural coupling but with mutual
understanding based on communication and language.
In our opinion, Luhmann’s orientation towards Husserl’s phenomeno-
logical analysis of meaning, on the one hand, and the Parsonian model
of double contingency, on the other hand, blocks the development of
an adequate theory of meaning based on systems theory. Contrary to
Luhmann’s view, the modal logical concept of contingency, by which
he wants to link phenomenology with sociology and thus avoid the
traditional problems of empirical versus transcendental cognition, can
neither overcome the solipsism of transcendental phenomenology on
the one hand, nor the biological reductionism of the double contingency
model on the other. Only communication can reduce contingency to
meaningful contingency. Thus, meaning exists only under the condi-
tions of communication. There is only one system of meaning, society,
consisting of persons and many other entities, depending on how the
meaning system differentiates itself internally.
2.8 Communication
Within the framework of a general systems theory, we may thus postu-
late that the elements of a system of meaning are not adequately grasped
merely as operations of purely psychically understood observation
or self-observation but only as semiotically coded “communication.”
This theoretical decision requires that the concept of communication
carry an enormous burden. It demands a theory of communication that
goes far beyond what the tradition has previously offered. Systems of
meaning are systems of communication. But what is communication?
From Systems to Actor-Networks
140
Following the various theories of communication (Shannon/Weaver
1949; Wiener 1948; Bateson 1972 and others), the concept of communi-
cation is often taken to be so broad that it can denote any interaction at
all. For example, Sebeok (1991:22-3), who advocates a “pansemiotism”:
All living things – whole organisms as well as their parts – are
interlinked in a highly ordered fashion. Such order, or organi-
zation, is maintained by communication. Therefore, commu-
nication is that criterial attribute of life which retards the
disorganizing effects of the Second Law of Thermodynamics,
that is, communication tends to decrease entropy locally. In the
broadest way, communication can be regarded as the transmis-
sion of any influence from one part of a living system to another
part, thus producing change. ... An implication of this way of
looking at communication is that the capacity for message
generation and message consumption, which are commonly
attributed only to humans, is here assumed to be present in the
humblest forms of existence, whether bacteria, plants, animals,
or fungi, and, moreover, in their component parts, such as
subcellular units (for example, mitochondria), cells, organelles,
organs, and so forth. The global genetic code, too, can (as it has
been) quite fruitfully be analyzed in communicational terms:
the message originates in a molecule, the master blueprint
called DNA, its end being marked by a protein. The intricate
interplay of nucleic acid and protein, the essence of life on earth,
provides a prototypical model for all forms of communication.
Because of the danger of reductionism as well as the blurring of the
essential difference between signal and sign associated with this view
of communication, as well as the problems with covering over discon-
tinuities in levels of emergent order Köck’s (1987:359) critical reserva-
tions are to be agreed with:
It is quite obviously unproductive to call any interaction
‘communication’ that may occur anywhere in the universe or
in or between inanimate and/or animate bodies. ... It seems to
Meaning as a System 141
me, therefore, reasonable and useful to understand the term
‘communication’ as a designation of a class of specific inten-
tional interactions between living beings, namely those interac-
tions which take place mediated, via media, i.e., with the help
of signs, more precisely: whose necessary condition is the exis-
tence of a code, by which a (possibly, as in ‘natural’ languages,
structured in itself) set of signs is linked with the corresponding
set of meanings, without saying anything about sign modalities,
stability of the code and the like. One could thus also speak of
‘semiotic’ interactions of all kinds.
Let it be emphasized again that it is not our job to deny the fruitfulness
and scientific validity of research approaches that consider inorganic
and organic processes as “communication.” We do not in any way deny
the legitimacy of “physiosemiotic,” “phytosemiotic,” “biosemiotics,”
and “zoosemiotic” approaches. It should be clear, however, that the
concept of sign thereby loses any unified meaning and disintegrates
into mere equivocation. Thus, to avoid the danger of a biologistic
reductionism in social systems theory, we recommend, in agreement
with Eco (1972:31ff), to postulate a “semiotic threshold.” According to
Eco, phytosemiotic and zoosemiotic research are
...to be identified as a kind of lower limit of semiotics, as the
point at which semiotics emerges from something that is not
semiotics, as the connecting link between the world of signal
and the world of meaning – as in physical anthropology between
the last primate and homo sapiens.
Plants, insects, and animals that coordinate their behavior with each
other via the exchange of genetically encoded signals are not elements
of a meaning system, that is, as long as the sciences that study these
phenomena are not the object of study. In terms of systems theory,
communication, in the restricted sense we propose here, functions as
a reduction of complexity by means of a boundary-drawing system/
environment difference that constitutes an autopoietic, self-referential,
operationally and informationally closed system at the specific level of
From Systems to Actor-Networks
142
emergent order called meaning. Thus, a semiotically organized meaning
system does not consist of plants or animals that coordinate their
behavior with each other by means of structural coupling – although,
of course, these are described by a meaning system; a meaning system
does not even consist of “rational animals,” even it prefers to differ-
entiate itself by this term, but it consists of meaning and information.
Man, as Peirce said, “is” a sign.
A second important consequence of the view that meaning systems
consist of meaning or information and that meaning and information
must be thought of as communication is that we must think of a systems
theory in the social sciences as a communication theory. The leading
question of a social systems theory is: How is communication to be
analyzed in terms of systems theory? It will become apparent in what
follows that a systems-theoretically oriented theory of communication
has to deal primarily with three problems: 1) the problem of the subject
of communication, 2) the problem of communication and action, and
3) the problem of the unity and differentiation of communication. We
turn first to the problem of the subject of communication, that is, who is
it that “speaks,” and related to this, we will discuss the problem of the
identity and/or difference of communication and action.
2.9 Communication, Subject, and Action
According to Luhmann, communication is the distinguishing feature
of social systems. Therefore, he must offer a theory of communication
that does not conceptualize communication as the action of individ-
uals or understand communication as one possible social activity
among others. If communication defines the social, it cannot be some-
thing that people could do but choose not to do as they wish. If the
meaning system is the social system and the social system is a system
of communication, it cannot be the case that individuals can decide to
communicate or not or do something else such as hunting, gardening,
building a house, etc. These are actions that Habermas, for example,
would call “instrumental” as opposed to “communicative.” Not only
Meaning as a System 143
does the action theory model of society undermine systems theory, but
it also undermines the primacy of meaning for the social. It undermines
systems theory because a society that consists of individual subjects is
not a system but a group, a collective whose order must be somehow
derived from the actions of the individuals and not by processes of
selection, relationing, and steering at the systemic level. This is what
Hobbes did by explaining how the Leviathan was created by the social
contract which integrated individuals into the state. From the begin-
ning of the discipline, how to integrate individuals into society has been
the fundamental question of sociology. Luhmann emphasizes that the
theory of social systems does not ask this question since society does
not consist of individuals. From the point of view of systems theory,
a society composed of individuals and their actions does not consist
of its own dynamic operations. Therefore, it cannot be conceived of as
a system. By insisting that society is a phenomenon that can best be
explained as an autopoietic, operationally and informationally closed
system on the level of emergent order of meaning, Luhmann must
propose a theory of communication that does not depend upon indi-
vidual subjects and their freely chosen communicative actions. This is
why Luhmann explicitly distances himself from the enthusiasm that
classical cybernetics shared for Shannon and Weaver’s transmission
model of communication. The transmission model sees communication
as a linear process initiated by a sender that sends information over a
channel to a receiver. Instead, Luhmann describes communication as
a self-referential, circular process, “a synthesis of three selections, as a
unity of information, communication, and understanding” (1995:147).114
The transmission model of telecommunication and computer science
also has the problem that it has nothing to do with semiotic coding.
Shannon and Weaver posed the question of the most efficient tech-
nological transmission of electronic signals. According to this model,
information is reified into a thing-like identity transmitted through a
channel from sender to receiver. Luhmann rejects this view of informa-
114 With this model Luhmann also wishes to distinguish systems theory from sub-
jectivist positions that understand communication as “action, speech, proclama-
tion, uerance” (147) or as some kind of action initiated by a subject.
From Systems to Actor-Networks
144
tion for systems of meaning since it makes communication a process
completed when the information is received. For systems of meaning,
on the contrary, communicated information is necessarily interpreted
and interpretable. It must somehow come back to the sender so that one
communication can connect to another and so on, thus guaranteeing the
autopoiesis of the system. The open world horizon of references Husserl
describes serves as the context of any selection of information. Whereas
signals trigger behavior, signs must be “understood” within a world of
meaning. It is remarkable that although Shannon and Weaver themselves
clearly denied that their model applied to semantic meaning, almost all
communication or information theory in all areas of physics, biology,
and sociology is based on this machine model. Meaningful information,
however, has no substantial identity that can be sent through a channel
to a receiver with minimum distortion but is continuously transformed
in the communication process, a process that is completed, as Luhmann
argues, only when it refers back to itself in a circular way such that ego
knows that alter ego has “understood” and responded which makes
further communication possible. Therefore, communication includes
a third selection beyond merely utterance and information, namely,
the selection of understanding. The selection of understanding, that is,
the circular self-reference of communication, closes the loop and leads
automatically from one communication to other communications and
thus to the ongoing autopoiesis of the communication system. This
means that semiotic coding requires a specific selection and relationing
process that cannot be modeled according to the encoding/decoding
model of machine signal transmission.115
Looking now more closely at the transmission model, if someone sends
something to someone, e.g., a man sends a lady flowers, the only rele-
vant question for the transmission model is how many flowers, how
quickly, are transmitted without damage. This is it. For social commu-
nication, on the contrary, the relevant question is how the lady receives
115 Nonetheless, one idea that Shannon and Weaver introduced into communication
theory has signicance beyond mathematical information theory, namely, the
idea that information has something to do with improbability and negentropy.
We will return to this idea below when discussion information and networks.
Meaning as a System 145
the flowers and what she interprets them to mean. The information that
the flowers are intended to communicate, and the expected response are
not the flowers themselves but what they are taken to mean and how
they are interpreted. The man no longer has the flowers but still has
the information. Luhmann rejects the entire “thing metaphoric” (139)
of the transmission model, which suggests that information is the same
for the sender and receiver. The lady may “understand” the flowers
in many different ways and respond unexpectedly. The recipient
selects an interpretation from among several possibilities. According to
Luhmann, this is where communication actually happens. And there
is no communication if it doesn’t happen. If the flowers are lost on the
way, no communication has taken place.
Luhmann’s model reverses the order of sender to receiver and under-
stands communication as a form of circular causality.116 Only when the
response comes back from the receiver to the sender has communi-
cation taken place. The sender selects, for example, certain flowers as
information which at a certain time are then “sent” in various possible
ways – let’s say red roses instead of white lilies, on their wedding anni-
versary instead of to a new acquaintance, and through the post instead
of bringing them himself. But there is no communication until the lady
somehow responds and says “thank you” or sends the flowers back or
ignores the communication, which is also a response. As Watzlawick
pointed out, one cannot not communicate. Social communication is
a “selective process” (140) and not a technical transmission problem.
The process includes three different but related selections. First, a piece
of information is selected, for example, the flowers; second, a form of
utterance – Luhmann also speaks of a “Mitteilungsverhalten” (196) –
such as sending the flowers by post or currier, or attaching a note, or
calling, etc., and third, the communication must be recognized as such.
The lady must receive the flowers and understand this as a communi-
cation and thus respond in some way. Recognizing a communication
means understanding it as a communication through an “acceptance
selection,” which could also be rejection that manifests itself in a subse-
116 “Communication is made possible so to speak, from behind contrary to the tem-
poral course of the process” (143).
From Systems to Actor-Networks
146
quent response. All these three selections of information, utterance, and
acceptance/understanding must happen before communication in the
complete and relevant sense of the word can take place.117 Commu-
nication “only comes about when ego establishes its own state on the
basis of a communicated information,” which implies that alter ego
is actually the sender, or at least that it is a circular relation. Double
contingency implies this circularity, which in turn results from ego
recognizing alter as alter ego and vice versa. As Luhmann argues, this is
based on distinguishing between utterance (the act of communicating)
and information (that which is being communicated).
If all three selections are not present, the attempted communication has
failed. The addressee did not hear or heard differently, was unaware
that anyone was trying to communicate, did not consider the speaker
capable of communication, etc. Communication can only occur in a situ-
ation of double contingency, that is, mutual social recognition, which is
a circular, self-referential situation. If there is no expected behavior on
both sides, such that the sender is also the receiver and vice versa, there
are no possibilities of connection to further communication. Except,
of course, to try again. It is important to note that “understanding” is
not described as an intentional act of a subject but as a moment in the
systemic processing of communication. If the flowers do not result in
a reaction, if one does not even get a thank you note, then the commu-
nication has either not been established or the interaction continues
due to other connecting or associated messages. One calls and asks,
for example, whether the flowers have arrived. Understanding as the
third selection thus has the function to enable connecting or associated
further communication possibilities and so on such that the autopoiesis
of the social system is maintained. For the autopoiesis of the commu-
nication system to be maintained, communications must connect to
117 Luhmann (1995:147) adds a fourth selection which he calls “acceptance or re-
jection” which is conceptually hardly distinguishable from understanding but
serves the purpose of enabling the development of certain forms of communica-
tion which Luhmann, following Parsons calls “symbolically generalized media”
that inuence participants in favor of acceptance. These media become the ba-
sis for the dierentiation of the semi-autonomous functional subsystems which
modern society, according to Luhmann, consists of.
Meaning as a System 147
further communications. Someone says something, and someone else
responds in some way, and so on in one unbroken chain of commu-
nication that makes up the social system. The meaning system, there-
fore, is not composed of thoughts, feelings, or actions but of a chain of
communications. One communication is networked with the following
one by the connecting behavior, which alone provides information
about whether one has been understood or not and whether one has
understood oneself correctly or not. I can no more “look inside” myself
than I can look inside another person to determine whether the words
have been understood. In communication, communications do not
operationally refer to people, inner mental states, or things or events in
the world but self-referentially to communications.
According to Luhmann, communication is thus the indissoluble
synthesis of the three selections: the selection of information, utter-
ance, and understanding. Insofar as the unity of all three selections
is only analytically resolvable, there can be no information in the
operations of the communication system that is not communicated
or could not be communicated; there can be no utterance that is not
understood or could not be understood. And there can be no under-
standing of anything other than communicated information. Every
selection of understanding is also automatically and necessarily a
further communication. The ground of possibility of this synthesis of
the three selections lies in “coding” (142), i.e., the specific organization
of communication into a semiotically coded meaning system.118 The
smallest possible unit of communication is when someone says some-
thing to someone. This is a semiotically coded synthesis of the three
selections and thus is constituted as a supra-individual interaction
context that cannot be thought of as the achievement of conscious-
ness, a subject, or an individual, or even as some collective action
of individuals who choose to communicate with each other or not.
Communication – and thus meaning in general! – is therefore never
only mine or yours, but always ours, always social. If communication
is the basic unit of society, and the selection of a subject is not part of
118 “The combination of information, uerance, and expectation of success in one
act of aention presupposes ‘coding’” (Luhmann 1995:142).
From Systems to Actor-Networks
148
the threefold synthesis, then the subject or the human being is not an
element of the social system, and the actions of individuals cannot
serve as the basis for creating society. When Hobbes’ free individuals
in the state of nature agreed to sign the social contract, they must have
been able to communicate and, through communication, understand
themselves as “free individuals” – which is not a natural state but a
social role! – who are authorized to enter into a social contract. Having
thus gotten rid of the subject, the question arises: who communicates,
who is it that speaks?
Interestingly, even though the communication synthesis as a whole is
the basic unit of the system of meaning, which cannot be resolved into
the actions of subjects, and which therefore does not imply subjects or
individual speakers, at least not human beings, it still remains impos-
sible for the communication system to dispense with the subject as
the originator of communication. This is evident in Luhmann’s sharp
distinction between actions and communications on the one hand and
his insistence that actions, not communications, constitute the elements
of the social system after all. On the one hand, it is asserted (164) “that
communication cannot be conceived as action, nor can the process of
communication be conceived as a chain of actions,” while on the other
hand, we hear (165) that “communication systems...must interpret
utterance as action” and that “a social system is constituted as an action
system.” If the word “action” here does not mean simply “operation” of
the system but something that someone does, how can we understand
this apparent contradiction?
According to Luhmann, the communication system itself must conceive
communication as non-communication, i.e., as action, because the
three selections of information, utterance, and understanding cannot
be observed directly. Communication can be “only inferred” (164)
through visible communicative behavior. We perceive not syntheses
of three selections but people talking to each other, exchanging things,
cooperating in various endeavors, and so on. In addition, there is the
problem that communication is a “symmetrical relation” of the three
selections, which leads to uncertainties and potential blockages that
Meaning as a System 149
must be “asymmetrized” (165). “Communication is symmetrical insofar
as each selection can lead the others and the leadership relations can be
continuously reversed.” Making an utterance, for example, can lead to
looking for some information that one wants to say, and this, in turn,
can lead to eliciting a response from someone, which can lead to saying
something because someone said something to which one wishes to
respond. The symmetry problem here is that one doesn’t know which
selection comes first, where to begin, and how to proceed. Is this situ-
ation not like the famous chicken or egg problem? Where should we
start? To solve this problem, it becomes necessary for society to ascribe
communication to a speaker. Every communication also has an ascrip-
tion to someone doing the talking.119
How should we order the communication process and keep it going
if we don’t know who said something first and who must respond or
react? The symmetrical arrangement of selections creates an unaccept-
able complexity because everyone could be talking and responding
about everything in different ways simultaneously. In the face of this
problem, the communicative process threatens to end. This problem-
atic complexity can only be reduced by directing communication along
certain lines. This is done through the attribution of actions:
Only by building the understanding of action into the commu-
nicative occurrence does communication become asymmetrical,
only thus can a person who utters information give directives to
its receiver, and this can be reversed only if the receiver begins
to utter something of his own, that is, begins to act. (Luhmann
1996:165)
Communication, therefore, seems necessarily to become communica-
119 The theory does not specify who the speaker (communicator) is. This depends
of the available forms of self-reference within the system. In principle anything
can communicate, not only human beings, but animals, ETs, spirits, God, indeed
anything. This marks the post-human position of Luhmann’s theory. Human
beings are not only banned from the social system, but they are also not neces-
sarily reconstructed within it as the only initiators of communication. The theory
requires an ascription but says nothing about what it is to which communication
is to be ascribed.
From Systems to Actor-Networks
150
tive action. According to Luhmann, however, communication in itself
has nothing to do with action because communication is a synthesis
of three selections, and action is something that can only be related to
one selection, the selection of utterance. Luhmann’s theory implies that
information has no agency of its own. Contrary to Bateson’s definition
of information as a difference that makes a difference, for Luhmann,
agency lies solely on the level of the system as a whole; only the commu-
nication system does something, and it attributes its processes to a
speaker. Acting human individuals are banned from the system. Still,
they are reintroduced into the communicative processing of meaning by
interpreting communication selections as communicative actions and
thus constituting them as being able to be connected to other commu-
nicative actions in an ongoing chain. Communications are attributed
to actors or, as Luhmann puts it, “persons” who are communicatively
constructed by the social system precisely for this purpose. Thus, it
is communication that constructs persons and not vice versa. Actions
of persons are the “attributions” of communications necessary for the
connectivity of processing meaning. Without the system being able
to attribute communications to “someone,” the system would remain
unable to conceive of communication as an operation of an operator at
all and thus to refer to itself on any other level than the metaphysical
level in which every communication would only be attributable to the
system as a whole. It is on this level that Luhmann can say that commu-
nication communicates.
In general, all autopoietic systems are necessarily self-referential.
Without basal self-reference, the system cannot refer its operations to its
own operations. Self-reference is a necessary prerequisite of all autopoi-
esis. As noted above, the specific self-reference of a system of meaning
occurs through reflection on system identity, i.e., the system maps its
system/environment difference onto itself. This creates what we have
called a metaphysical or world-inclusive self-reference. The autopoiesis
and operational closure of the system require that the system refers to
itself as doing something, as a dynamic entity. The world is, therefore,
always a “product,” even if we prefer not to speak of a creator God
but of a self-organizing evolutionary process. If we do not want to
Meaning as a System 151
attribute all agency to God, it becomes useful to differentiate further
self-references within the system, that is, to identify speakers. Without
being able to answer the question “Who communicates?” on the level
of everyday life, processing communications and linking one commu-
nication to another would be challenging. It would be as though God
would have to do everything; even the most minor decisions and events
would have to be authorized and carried out by God. Some believers
try to live in this way; they leave everything to God, but these people
usually can only survive in very small communities of the like-minded,
that is, of those willing to wait until God has cleared everything so that
life can go on.
Internal differentiation of self-reference requires that the system
becomes the subject for itself on levels below the totality. The system
must make “a description of itself … in order to control the progress
of processes, the reproduction of the system” (165). It does not matter
who is speaking so long as someone is, and an ascription can be made.
There could be humans, gods, spirits, ancestors, animals, plants, nature,
the Freudian subconscious, and beings from other planets, but also the
state, reason, conscience, law, justice, fate, the stars, invisible voices,
etc. who all can appear as “persons;” the only important thing is that a
system reference exists and that an attribution of action is possible.
The communication system is indifferent to holders of the subject posi-
tion; all that matters is that the position is filled and there is a – any –
subject. The requirements of self-referentiality and operational closure
of the communication system thus create the appearance – Luhmann
speaks of a “self-simplification” – of a system of action and constructs the
actors necessary for it, whoever they are: “For purposes of self-observa-
tion and self-description, the symmetry of communication is asymme-
trized...” (165). Acting subjects are thus constructed as elements of the
social system and simultaneously distinguished from communication
since they can choose to communicate or not. As well shall see, society
must take account of this by constructing many supporting structures to
ensure the motivation and cooperation of such actors; among the most
important of these are “symbolically generalized media,” which lead to
From Systems to Actor-Networks
152
the self-organization of autonomous functional subsystems, which we
will discuss below.
In summary, Luhmann’s communication theory is based on the idea
of autopoietic, self-referential, and operationally closed systems. The
system in question is the social system that appears on the level of
emergent order of meaning where it is meaningful information that is
communicated in circular, self-referential operations that are a synthesis
of the three selections of information, utterance, and understanding,
and which is confronted with the problem of reducing complexity by
means of internal differentiation into actors or subjects of various kinds.
Much of this theory is quite useful for understanding communication
as a systemic form of order. The question arises, however, of whether
Luhmann is not unnecessarily subscribing to a minimal notion of
agency. Because of the unnecessary and, in our opinion, even unten-
able distinction between psychic and social systems, Luhmann puts
himself, it would seem, into the situation of excluding individuals and
their actions from society while at the same time having to reintroduce
them into society through the back door as socially constructed persons
who alone can initiate communication. What is the relation, we may
ask, between psychic systems and those actors who, as persons, popu-
late the social system? How can the one be transformed into the other?
For Hobbes, this was a simple matter of signing the social contract. For
Luhmann, however, the system/environment difference must be over-
come since psychic systems are banned from the social system and find
themselves in its environment. Since there is, in principle, no commu-
nication between system and environment, how are psychic systems
transformed into elements of the social system?
The first answer to this question is to simply refer to the general prin-
ciple of systemic order that all systems construct their own elements.
According to its organizing principles, the table constructs a top and legs
out of whatever material is at hand or desired. An organism constructs
tissue, organs, etc., according to its genetic coding from the material
substrate. A meaning system also constructs its elements. These are
information as processes of communication. This is called self-organi-
Meaning as a System 153
zation. The environment offers systems the problem of complexity that
they must solve. Luhmann follows Heinz von Foerster’s notion of “order
from noise” to understand this process. Generally, when systems are
environment for each other, they are nothing other than environment
and can only perturb each other, that is, offer contingency, indetermi-
nacy, variety, and complexity so that the system can exorcise its self-or-
ganizing power and create, as von Foerster put it, order from noise.
Luhmann, as we saw above, calls this “interpenetration.” Colloquially,
one could say that individuals make life complicated and unpredictable
for society, and society constantly makes life hard for individuals. This
traditional sociological commonplace is reinterpreted within systems
theory as interpenetration. It is the systems theoretical version of the
modern Western paradox of free individuals vs. constraining society,
the one vs. the many, the individual vs. the group, or the classic puzzle
of how to integrate individuals into society while at the same time
allowing sufficient freedom to prevent society from collapsing into
sterile conformism.
In the systems theory version of sociology, the social system uses the
contingency or “noise” generated by psychic systems to construct order.
On the other hand, psychic systems use the constraints of society to
construct deviation, contingency, freedom, and indeterminacy, which
is the unique way in which they self-organize. The general principle
at work here is that systems self-organize in response to complexity.
Psychic and social systems “interpenetrate” in so far as they mutually
introduce contingency into each other and thus instigate self-organiza-
tion. Why the social system needs the contingency of psychic systems
instead of the contingency of the world, or in other words, the question
of whether the world itself is not sufficiently complex and thus able
to introduce all the indeterminacy any social system could ever want
completely independent of individuals is not addressed by Luhmann.
One could suppose that the sociological interpretation of the system of
meaning as society and not as world leads Luhmann into steering the
systems theory of meaning into a sociological corner where questions
of the self-organization of meaning become automatically questions
of how individuals relate to society. Perhaps an adequate theory of
From Systems to Actor-Networks
154
systemic order on the level of meaning must fundamentally revise the
assumptions of sociology and look for other foundational concepts than
“society.”120
The above discussion allows us to ask why agency in Luhmann’s theory
must be limited only to the utterance selection and have nothing to do
with the selection of information. Ascribing agency only to utterance
seems to be a consequence of the aim to banish subjects from society, and
indeed from the theory of society as a system of communications as well.
But apart from the desubjectification program Luhmann embarks upon
and the somewhat questionable reintroduction of subjects as speakers,
one can still ask why information is devoid of agency. If information
is defined as differences that “make” differences, then where does this
efficacy and agency to make differences come from? How does infor-
mation “make” a difference without any agency of its own? And what
kind of agency could information have? Is it impossible to suppose that
meaning is a dynamic process that arises out of information? Could
it not be that it is information that “selects” agents and speakers? Did
Luhmann himself not say that the subject is not primary since meaning
is the condition of its possibility? As argued below, this is an option that
the network paradigm, in contrast to the systems paradigm, explores.
We will ask whether systems theory can also accommodate the agency
of information or whether it is precisely why systems theory tends to
break down and must be extended by network models. But we must
not get ahead of our story.
As a response to the question of agency and information, one could
argue, as Luhmann does, that differences can only make differences
insofar as they are communicated and understood. And, as we saw,
communication requires actors who communicate. But if one says
the same thing over and over again, there is no information, and no
difference is made. Redundance has no informational value, regardless
of how much it may be communicated. Utterance does not make the
difference that information can make, for there can obviously be a lot of
120 This is precisely what we will suggest when discussing the network paradigm
below.
Meaning as a System 155
communication – we experience this daily – without much information.
But when information is communicated, it makes a difference because
it has its own power to change reality quite apart from the utterance. If
one understands, for example, that tobacco, alcohol, butter, and frozen
meat are dangerous to one’s health, then, as Luhmann (203) notes, one
has become “another.” It can no longer be ignored. In what does this
understanding show itself, if not in some connecting behavior that gives
others and me an understanding of my understanding? How is under-
standing itself to be understood if not as a state-transforming operation
that can somehow be attributed to information and not to a subject or
individual actor? Philosophical hermeneutics does just this by attrib-
uting understanding to the efficacy of a text, an artifact, or a historical
event. The efficacy of information goes beyond the author and inter-
preter. It arises from the text or artifact that itself has the power to open
up a horizon of meaning. After all, where do the many “references”
come from that Husserl already identified as meaning? Philosophical
hermeneutics emphasized against Husserl that understanding is a
historical process and not an act of an autonomous subject. Gadamer
(1960), for example, takes his cue from theological and legal hermeneu-
tics, where understanding the Word of God or the law manifests itself
directly in practically changing the situation of the interpreters and
thus realizing the historical process.121
Given these considerations, it would seem that agency could be
conceptualized as distributed throughout the system, attributable to all
system elements and operations. Luhmann chooses not to pursue this
path when conceptualizing the system of meaning. Instead, a speaker
is constructed by communication for attribution. These speakers are
conceived as socially constructed “persons” who replace the human
individuals of traditional sociology. Luhmann is forced to relate action
to only one of the three constitutive selections of communication, and
this is only a kind of necessary evil to make communications asymmet-
rical and ordered in time. This is a requirement of the temporalized
processing of communication and has nothing to do with traditional
121 For a discussion of philosophical hermeneutics in relation to systems theory and
network theory see Krieger/Bellger (2014).
From Systems to Actor-Networks
156
sociological theories of agency vs. structure.122 Nonetheless, it seems to
be a truncated view of agency since, in effect, it introduces human indi-
viduals into the social system through the back door. Theoretically, this
is not necessary. It is possible to argue that agency is implied as much
by the selection of information and the selection of understanding as
by utterance and that the subject, therefore, does not need to be reintro-
duced into systems theory as a privileged source of agency.123 Despite
these objections to Luhmann’s theory, we readily admit that replacing
the act of speaking with the systems theoretical notion of selection and
operation is a significant theoretical innovation in sociology that must
be kept in view. The system is the actor!124
2.10 TheDierentiationofCommunication:Functional
Subsystems
For Luhmann, a systems theory of meaning is a theory of the social
system as a system of communication. As a system of communication,
the social system can only become internally differentiated by means
of differentiating communication. Luhmann succinctly puts it: “The
differentiation of social systems can emerge only through the differ-
entiation of communication processes” (1995:152). It is not semantic
content of any kind that is decisive for meaning; it is how communi-
cation of any content is structured. It is not what is said but how it is
said that differentiates society. How does the social system differentiate
communication? As the answer to this question, Luhmann proposes his
definition of modernity. Modern society differentiates communication
into functional subsystems, from which he concludes that modern society
is a functionally differentiated society:
Functional differentiation means … that autonomous subsys-
tems of society are formed in the orientation to their own respec-
122 See Giddens (1984) for an inuential statement of this position.
123 Contemporary non-Cartesian cognitive science (see Rowlands 2010) and theories
of distributed agency have persuasively shown under the name of “aordances”
that things also have agency and contribute to cognition.
124 We will return to this slogan in the discussion of actor-networks below.
Meaning as a System 157
tive functions, which reproduce themselves self-referentially,
orient themselves recursively to the respective self-produced
communications, and thus realize the characteristics of struc-
turally determined autopoietic systems. (Luhmann 1990b:479)
The social subsystems Luhmann has in mind are the economic system,
the legal system, the political system, education, art, religion, science,
the media system, and healthcare, to name only the most important
ones. These subsystems serve specific social functions and can “repro-
duce themselves self-referentially” and “orient themselves recursively
to the respective self-produced communications.” They can do this
because they operate autopoietically, self-referentially, and are opera-
tionally and informationally closed based on their unique binary coding.
Such binary-coded functional subsystems process all communication
in society. After all, if one is not engaged in business, education, and
training, political activities of some sort, healthcare, art, religion,
science, media, etc., what is one doing? Perhaps there is room for small
talk, but this is also systemically organized as “interaction systems.” As
sociology (see Gofman) has long known, socially constructed persons
in the modern functionally differentiated society are like actors playing
roles in different “frames,” settings, or contexts according to varying
scripts for various purposes. According to Luhmann, the codes that
structure this communication in one way or another emerge in society
from the differentiation of symbolically generalized media.
Media, in general, are “evolutionary achievements that start at ... ...
breaking points of communication” and serve to “transform the improb-
able into the probable” (Luhmann 1984:220), whereby the improbable
here consists first of all in the fact that communication occurs at all.
For Luhmann, each of the three selections constitutive of commu-
nication mentioned above (information, utterance, understanding)
is improbable. That one has something to say (information), that the
message reaches someone (utterance), and that one is understood and
responded to are all improbable. Assuming these three improbabilities,
three media function to make the improbable probable: Language,
dissemination media (such as the press and radio), and finally, what
From Systems to Actor-Networks
158
Luhmann, following Parsons, calls “symbolically generalized media.”
On the concept of symbolically generalized media Luhmann writes
(1995:161):
We would like to call “symbolically generalized” the media that
use generalizations to symbolize the nexus between selection
and motivation, that is, represent it as a unity. Important exam-
ples are: truth, love, property/money, power/law; and also in
rudimentary form religious belief, art, and, today, standardized
“basic values.”
Symbolically generalized media, as opposed to other media such as
language and dissemination media such as writing, radio, the press,
etc., play a major role in differentiating social subsystems.
In all these cases … it is a matter of conditioning the selection
of communication so that it also works as a means of moti-
vation, that is, so that it can adequately secure acceptance of
the proposed selection. The most successful and most relevant
communication in contemporary society is played out through
these media of communication, and accordingly, the chances
for the forming of social systems are directed toward the corre-
sponding functions. (161)
Luhmann ultimately attributes the internal differentiation or build-up
of internal complexity in the communication system via such symboli-
cally generalized media to the contingency of the environment, namely
to the assumption, already discussed above as problematic, of a multi-
plicity of pre-communicative consciousness systems that think and act
“obstinately” (1990b:40), and which would tend not to participate in
communication at all if special motivational structures (namely: the
symbolically generalized media) did not facilitate and steer their inte-
gration into the social system. What for Hobbes was accomplished by
the need to secure one’s life in the war of all against all is, for Luhmann,
accomplished by symbolically generalized media that “motivate”
people to communicate and cooperate.
Meaning as a System 159
According to Luhmann’s functional differentiation model, interactions
or communications are absorbed into a particular symbolically gener-
alized medium where a functional subsystem of society codes them to
make it much easier to “understand” what the communication is about.
The codes of the functional subsystems are two-valued or binary codes.
The most important functional systems structure their commu-
nication by means of a binary, two-valued code that claims
universal validity from the point of view of the specific function
in each case and excludes third possibilities. (Luhmann 1990:75)
Money, for example, is seen as a symbolically generalized medium
in which a certain communication is made possible by attributing a
monetary value to something. After something has its “price,” all that
remains relevant communication is whether to buy or sell. The money
code or the binary alternative of buying or selling filters the world into
only two values: buy or sell. All other communications, such as those
about religious belief, educational certification, legal judgment, political
office, artistic value, etc., are excluded from the economic system. The
binary coding of a functional subsystem is both inclusive and exclusive
but, paradoxically, for this very reason, universal. There is nothing that
doesn’t have its price; everything comes under the binary coding of the
economic system. But only with regard to buying and selling. Every-
thing else is banned into the environment of the economic system and
is not registered by the code. One may discuss the aesthetic value of a
work of art as long as one wishes, but at a gallery, one has to decide
whether to buy the painting. Since a price can be given to anything,
including love, the economic system becomes an autonomous, autopoi-
etic, operationally and informationally closed system.
The same happens for the other functional subsystems. Due to the medi-
um’s claim to universality and the binary code, in which all interactions
are structured by either the positive or the negative value, connecting
up communications to one another becomes much more straightfor-
ward than in a situation where one must first clarify what is being said.
Knowing, for example, that a particular communication is an offer to
From Systems to Actor-Networks
160
sell something and not the profession of a belief, an educational instruc-
tion, a legal argument, etc., makes it much easier to “understand” what
is going on and communicate reciprocally. All communications relating
to buying and selling are connected, and all other communications are
excluded. The code enables the construction of a system/environment
difference and, thus, the emergence of the economic system as a social
subsystem organized by its unique code. Luhmann calls the sum of
communications coded according to the binary scheme of money, that
is, paying or not paying, the economic system. Due to the all-inclusive
binary code, the system differentiates itself from the overall social
communication as an autonomous, autopoietic, operationally and
informationally closed subsystem, whose function is to facilitate the
material reproduction of society, that is, the production and distribu-
tion of goods and services.
Corresponding to this model of the economic system, other function-
ally differentiated subsystems emerge within society out of their own
symbolically generalized media. The political system is coded by the
binary alternatives of either power or no power: disposition over offices
and positions in the medium of power. The legal system is coded in
terms of legal/illegal in the medium of legal decisions. The science
system is coded in the binary alternatives of true/false in the medium
of knowledge. The educational system is coded in the alternatives of
certification or non-certification (pass or fail) in the medium of career
expectations. And the religious system, according to Luhmann, is coded
in immanence and transcendence in the medium of faith. One could
add the art system, which is coded by what is considered to be original
and dysfunctional as opposed to what is functional and already known
in the medium of contingency. Finally, the healthcare system is coded
by the alternatives of being sick or healthy.
The functional differentiation of social subsystems occurs as a reduction
of increasing social complexity. Society has developed in complexity
from segmented (archaic) to stratified (high cultural) to modern
functionally differentiated societies. Functional differentiation results
“autocatalytically” (1995:190) from the autopoietic reproduction of the
Meaning as a System 161
meaning system. The general “mechanism” of evolutionary differen-
tiation specified in relation to communication is called, according to
Luhmann (1990b:49,81), “surplus generation-and-selection.” What at
the level of genetic organization provides for surplus and variety of
living beings as random mutations is attributed at the level of semiotic
organization to the arbitrariness of observation as it happens in the
conscious environment of the communication system. “Distinctions are
always arbitrarily introduced and used” (99). The uncontrolled obser-
vations of psychic systems introduce distinctions arbitrarily, creating
a surplus of possibilities. Based on this variation, selections are made
first by the symbolically generalized media and then increasingly
by the various social subsystems. It becomes generally expected that
everything be given a price and that markets are established, that every-
thing can be either legal or illegal and that courts are established, that
specialized competencies can be learned and certified by schools, and
that true knowledge can be discovered by scientific methods leading
to research becoming institutionalized. Furthermore, it becomes insti-
tutionalized that religion has to do with faith, not politics, which is
organized in political parties and programs, and that healthcare has
to do with doctors, hospitals, and health insurance and not with reli-
gious rituals. What comes out of this differentiation process is modern
Western society.
According to the concept of autopoiesis, communications reproduce
themselves in such a way that they simultaneously reproduce their
own reproducibility, i.e., produce their own conditions of possibility.
Since everything depends on the chain of communications continuing
without breaks and gaps, the conditions of the possibility of communica-
tion can be found in whatever successfully ensures that communication
is linked up to further communications. The symbolically generalized
media and the functional subsystems ensure that the autopoiesis of the
communication system can be maintained. They reduce the complexity
of communication and make it manageable. The idea of differentiation
means that systems reduce complexity by building systems within
themselves.
From Systems to Actor-Networks
162
Confronted with the imperative that the autopoiesis of communication
must continue, the way to ensure communication becomes the construc-
tion of specialized subsystems within the social system.
system differentiation repeats the formation of the overall
system within itself. The overall system is reconstructed as the
internal difference between a subsystem and the subsystem’s
environment, and this reconstruction is different for each
subsystem. .... The overall system is contained within itself
many times over. (1995:191)
Importantly, the internal differentiation process is not purposefully
directed by the system as a whole. The subsystems that emerge within
the social system have no special relation or informational exchange
with the encompassing social system, which is for them nothing other
than environment.
internal system formation occurs autocatalytically, that is, by
self-selection. Internal system formation presupposes neither
“activity” by the overall system nor a capacity for dealing with
that system, not to mention any overall plan. Nor does the
overall system subdivide or break down into subsystems. The
overall system merely enables the self-selection of subsystems
through its own order. (1995:190)
Every symbolically generalized medium and the functional subsystems
that emerge from them tend to absorb all communications into them-
selves according to their unique codes. In this way, the complexity of
social interactions becomes reduced to those communications coded
by the subsystem. Each subsystem must determine for itself what is
system and what environment. The consequence of functional differen-
tiation is that after the differentiation of such autonomous, autopoietic
subsystems, no overall social communication remains. The various
subsystems compete to represent the system as a whole. The economic
system, for example, encodes everything exclusively in the alternatives
of payment or non-payment, just as the legal system perceives every-
thing as legal or illegal. For science, there is nothing that cannot become
Meaning as a System 163
the object of research. For the religious system, the world is divided
into sacred and profane. Everything else that is neither legal nor illegal,
neither sacred nor profane, which has no price and cannot be bought or
sold, which does not lead to an educational certification of some kind,
etc., is banned into the environment.
In what larger communication contexts, then, are the respective
subsystems contained? What code mediates communication between
the functionally differentiated subsystems? How can overall social
communication still be possible if the social system is fragmented into
autonomous, informationally and operationally closed subsystems? On
the other hand, it seems necessary to have overall social communication
because if the semiotic code were not internally repeated and specified
in the subsystems, we would not be dealing with a subsystem but with
a completely different society. Each autonomous subsystem would be a
different communication system, and there would be no encompassing
social system that somehow contains the subsystems within it. Subsys-
tems are subsystems only because they are in some way integrated into
society as a whole. After all, they fulfill functions for society. This raises
the question of the unity of society against the multiplicity of subsys-
tems. Society becomes the environment for the functional subsystems,
and each subsystem is environment for each other. How is the unity
of the system/environment difference to be theorized for the system
of meaning when society itself fragments into various autonomous
subsystems? This is the problem of “ecological communication.”
2.11 Ecological Communication or Can Society be Modeled
as a System?
Conceptualizing the system of meaning as society and then going on
to understand society as a system of communication that is differenti-
ated into autonomous subsystems, each of which is only environment
for the others, raises the question of the unity of society as a whole.
Wherein lies the unity of a functionally differentiated society? In
pre-modern societies, the unity of the “world” was religiously repre-
From Systems to Actor-Networks
164
sented. It makes little difference that there were and still are different
religions, worldviews, ideologies, cultures, etc. Religious and cultural
pluralism does not amount to systemic pluralism. The pluralistic world
is still one world; different convictions and views about the world
give society more to talk about. In the modern functionally differen-
tiated society, however, the “world” is no longer available since each
social subsystem sees the world differently and blocks informational
exchange between system and environment. When business people
are involved in politics, science, education, or religion, they are not
doing business. Modern functionally differentiated society requires
that socially constructed persons are theatrical actors playing different
roles in different dramaturgical settings for different audiences, etc.
Goffman’s dramaturgical role theory of social interaction describes
this situation in detail. And, just as there is no unity for society as
a whole, there is no unity for the actor. For a long time, modernity
dealt with this problem by substituting “reason” or “rationality” in
the place of God to represent the unity of the system of meaning. Not
only has postmodern critique and deconstruction eliminated this form
of metaphysical self-reference and raised the question of the extent to
which any self-representation of society as a whole is still possible, but
functional differentiation has replaced society with specialized forms
of communication that are autonomous, operationally and informa-
tionally closed systems in their own right.
Is society anything other than the mere sum of the differences of the
various subsystems to each other? What kind of communication thema-
tizes the unity of the differences between the subsystems and their
environment? This kind of communication Luhmann calls “ecological
communication.” It is misleading to assume that ecological communi-
cation talks about environmental problems such as depletion of natural
resources, pollution, climate change, etc. Of course, this is what everyone
means by ecology, and everyone expects communication about ecology
to deal with these important issues. This is precisely what happens,
but it occurs within the particular horizon of meaning of the various
functional subsystems. The “natural” environment is a topic of laws
and regulations, of business models, of scientific research, or religious
Meaning as a System 165
beliefs and ethical admonitions, educational programs, art, political
campaign promises, etc. But this is the problem and not the solution.
We must emphasize that we are not talking about ecological commu-
nication as it is commonly understood. We are not concerned with the
“natural environment.” We are concerned with the social environment,
for even nature exists only insofar as it plays a role in communication
and is constructed by meaning within society. We are concerned with
the unity of the difference between system and environment for a system of
meaning. Specifically, we are concerned with the problem of the unity
of society as a whole under the conditions of functional differentiation,
that is, under the conditions of a society that has differentiated itself
into autopoietic, operationally and informationally closed subsystems
such that each subsystem is related to the others and society – whatever
that might still be – as to an environment. This is why the concept of
ecology is of central importance to a theory of social systems. There
may be no such thing as the social system, but only social systems that
are more or less structurally coupled.
The attempt to conceptualize the social system under the conditions of
functional differentiation brings the idea of ecology onto center stage. As
we have already had occasion to note, ecology poses a serious problem
for systems theory. On the one hand, every system is constituted by the
exclusion of the environment, and on the other hand, ecology focuses
precisely on the unity of the system and environment. As we have
argued above, It would seem impossible to speak of an ecosystem, for
this system would have to exclude an environment, which would need
to be included again and so on infinitely or at least until one reached the
entire universe. Ecology, therefore, stands for the problem of concep-
tualizing the unity of the difference between system and environment.
This problem becomes acute when history arrives at functional differ-
entiation as the primary form of social order. Luhmann (1989:5) speaks
even of a “fundamental paradox” in “the theoretical structure of the
ecological question.” The unity of the difference between system and
environment cannot itself be a system. What cannot be a system cannot
become a topic for systems theory. Could it be that the unity of the
From Systems to Actor-Networks
166
difference between system and environment itself does not fall within
the scope of systems theory? Systems theory presupposes it but begins
only from the difference, not the identity. Could it be that the theory has
a blind spot and cannot see what it cannot see? And if this might be the
case, does this imply that society as a whole cannot be conceptualized
as a system? If so, then meaning cannot be adequately modeled as a
system.
In principle, there can be no communication, i.e., informational
exchange between an autopoietic, self-referential, operationally and
informationally closed system and its environment. Information is
constructed within the system and not imported from the environment.
Regarding information, there is no “instruction” but only construction,
which occurs only within the system. The environment can only perturb
the system but not inform it. Were this not the case, the “system” would
not be autonomous, autopoietic, and informationally and operationally
closed. Instead, it would be an element of a greater system, subject to the
information constructed by the greater system of which it is an element.
The tabletop, whatever material it might be made of or form it might
have, is an element of the table system that constructs it as a tabletop.
The organizing principle or structure of the table does not represent
environmental complexity that pressures some piece of wood or metal
to self-organize into a tabletop. This is also true for living systems.
Meaning systems, however, have a unique relation to the environment.
Meaning systems have the unique property of constructing their envi-
ronment as meaningful within their own boundaries. For this reason, it
could seem plausible that functional subsystems within society can be
“instructed” by the information already existing in the environment,
which, since it is itself meaningful, must also be in some sense of the
term “information” that can be communicated. The question of the
unity of the difference between system and environment, the ecolog-
ical question, would therefore be answered alone by the fact that for
meaning systems, there is no non-meaningful environment. On the
level of general systems theory, however, this solution leads only to
the further problem of how a system constitutive difference between
Meaning as a System 167
functional subsystems and their environment is conceivable. And if
there can be no constitutive difference between system and environ-
ment, it makes no sense to speak of functional subsystems as systems in
their own right. They must either be conceived of as system elements –
which only pushes the problem of a constitutive distinction back to the
social system as a whole – or, if they are autonomous systems, how can
their unity with society be conceptualized? We will argue below that
ecological communication is possible but not based on systems theory.
Instead, we will say that meaning can only be adequately theorized as a
network, not a system. In contrast, the functional subsystems within the
social network can be modeled according to systems theory.125
Given these open questions, we must ask what the concept of the
“social” means from a general systems theory perspective.126 Despite all
of Luhmann’s claims to the contrary and the entire program of a theory
of social systems, society as a whole seems unable to be modeled as a
system. Various reasons support this claim. These are:
1. A distinction between system and environment does not constitute
meaning and society.
2. Meaning is not subject to evolution. Even if one conceptualizes
meaning as a system, the system cannot be understood as an au-
topoietic, self-referential, operationally and informationally closed
125 The theory is not saved by simply thinking of meaning as a medium, as Luh-
mann suggests, since a medium is only environment for a system and not a sys-
tem itself.
126 The anomaly, that no communication can address society as a whole was recog-
nized already in the 90s; see Fuchs (1992). Instead of being allowed to challenge
the systems theory paradigm, however, the problem was dealt with in Ptole-
maic fashion by generated ever more theoretical epicycles. Luhmann (2012:40)
aempts to address this issue by distinguishing between three levels of analysis:
“(1) general systems theory and, within it, the general theory of autopoietic sys-
tems; (2) the theory of social systems; (3) the theory of the societal system as a
special instance of social system.” This merely pushes the question back to the
social system since society as a whole is a “special instance” of the social system.
Luhmann aempts to resolve these issues by describing the operational closure
of communication. The social, regardless of level of generality, is nothing other
than communication. The concept of communication, however, depends on a
theory of meaning which leads to a tension in Luhmann’s theory between mean-
ing and society.
From Systems to Actor-Networks
168
system. This means that the relationship between the system and
the environment cannot be understood as a reduction of complex-
ity under the pressures of evolutionary selection.
3. Although social systems can emerge within meaning, there can be
no social system, that is, no system of the whole of society, world,
or meaning.
Let us look more closely at these three reasons for doubting the systems
model of society.
1. Meaning is not Constituted by a Distinction Between System and Environ-
ment.
If meaning is to be modeled as a system, it must be constituted by a
difference between system and environment. Processes of selection,
relationing, and steering must reduce the complexity of the envi-
ronment so that order is established against a relatively disordered,
random, unstructured, or, at any rate, more complex environment. Of
course, no system arises in a completely unstructured environment in
the real world. If there were no regularity in the universe, no struc-
ture, no natural laws governing matter, energy, etc., then no kind of
order could establish and maintain itself. For this reason, Maturana and
Varela speak of “structural coupling” instead of adaptation. Organisms
and environments are adapted based on structures that somehow “fit”
together. Evolution is, therefore, a “structural drift” of organism and
environment, each influencing and conditioning the other. This does
not preclude, however, that the environment is always more complex
than the system, and it is this environmental overcomplexity, that is,
contingency or the fact that the environment can change in unforesee-
able ways, which poses the problem that systemic order solves.
In the case of meaning systems, the system/environment difference, as
we saw earlier, must fall within the system, for the environment must
also have some meaning. It is, therefore, difficult to see how society as
a whole can fulfill the requirements of self-reference and operational
and informational closure. Self-reference requires other-reference.
Meaning as a System 169
Operations of the meaning system make meaning, and since meaning
extends to include the environment, the system cannot refer its opera-
tions to itself, for there is no other from which it can distinguish itself.
As we remarked earlier, the world can be the subject of a metaphysical
self-reference, but of course, this is paradoxical in that no “other” can
be distinguished from the self. If there can be no self-reference without
reference to an “other,” then we are faced with the metaphysical ques-
tion of Being. Nothingness “is” in some sense, too.
Furthermore, since the environment must also be meaningful, infor-
mation is constructed within the system and the environment. We
must conclude that for meaning, self-reference is paradoxical. Where
does meaning stop and non-meaning begin? Where is the boundary
between meaning and non-meaning? How can meaning be a system
characterized by operational and informational closure? If not, there
can be no constitutive system/environment difference. That from which
the meaning system must differentiate itself is always already within
it. What is outside society? As long as society was understood to be a
kind of thing, a substance, with visible territorial boundaries, it could
be observed from without, and there could be something “outside,”
for example, outside the city walls or the borders, whether cultural
or physical. One could stand, for instance, on one side of the border
between Germany and France and observe French society. Or one could
stand within the city walls and look beyond into a world of uncivilized
nature. As long as society was a territory, there was something outside
of society. The entire discipline of cultural anthropology and ethnology
is built on this assumption, and as the methodological discussion in
these disciplines finally realized, the “others” are us.127 The moment
society becomes a system of meaning, it has no boundaries. Commu-
nication does not stop at any borders, no matter how heavily guarded
they may be. It is a general principle of systems theory that systems
are constituted by inclusion and exclusion. What does communication
exclude? Of course, within society, exclusions of all kinds are the rule
and not the exception. But as a communication system, one cannot even
127 See Krieger (1991) for discussion of the methodological debate in ethnology.
From Systems to Actor-Networks
170
argue that society excludes non-communication. As Watzlawick noted,
one cannot not communicate.128 It could be argued that since society is
a meaning system, it excludes everything meaningless. However, the
binary distinction of meaning/non-meaning is paradoxical since even
non-meaning must have some meaning to be used by the meaning
system as a boundary. Even the distinction between information and
noise must be information and not noise since it is a difference that
makes a difference.
Social systems, understood as meaning systems, represent a very
unusual system. They are all-encompassing, thus violating the prin-
ciple of selection, which demands exclusion. They are not operation-
ally closed since self-reference depends on distinction from something
other, and there is no other. They are not informationally closed since
no undifferentiated perturbations are coming from outside that could
serve as the stuff out of which information is constructed. Could it
be that meaning is not a systemic phenomenon but must be modeled
by other principles of order than selection, relationing, and steering?
Could it be that although the functional subsystems can be modeled as
systems, the society they are subsystems of is not itself a system?
2. There is no Evolution of Meaning
The dominance of biological models in systems theory has led to the
widespread use of the concept of evolution and its application to
systems of all kinds, not just living systems. Physicists speak of the
adaptive behavior of complex dynamic systems and the evolution of
the universe, and sociologists of the evolution of culture and society.
Indeed, it would seem that everything is in some way evolving. But
what exactly is meant by evolution? The principles of evolution are
variation and selection. We wish to argue here that these principles
128 Walawick (1967). Luhmann on the contrary holds on to “The deniteness of the
external boundary (= the distinguishability of communication and noncommu-
nication)” which he claims “enables the operational closure of the world society
system” (2012, 87). However, this is dicult to reconcile with the statement that
“World society is the occurrence of world in communication” (87), since it is
impossible to imagine where the world stops and non-world begins, and it is
equally impossible to say it.
Meaning as a System 171
are to be found exclusively on the biological level of emergent order
and not on the physical level or the level of meaning. We argue that
only living systems are subject to the forces of variation and selection.
What does evolution mean on the level of life? Living systems are
proposed solutions (variations) to the problem of autopoiesis within the
constraints of an ever-changing environment (selection). If the problem
is not adequately solved or an existing solution becomes inadequate,
the organism is “selected” against and it disappears. This is a funda-
mentally different concept of selection than self-organization, which
selects elements, relates them, and steers system operations. As long
as a living system can continue its autopoiesis, it is considered to be
“adapted” to the environment and thus “viable.” As long as the envi-
ronment is considered “nature” one speaks of “natural selection.” It is
important to note that no living system selects itself in the evolutionary
sense of the term. There can be no self-selection in biology or evolution.
Systems self-organize and operate to maintain autopoiesis but do not
themselves select for viability. Self-organization is not selection in the
evolutionary sense. Since the emergence of meaning, nature as well as
life have become “integrated” into meaning; it is becoming increasingly
“culture” rather than “nature” that selects for or against living systems.
Indeed, the much-discussed Anthropocene expresses precisely this fact,
implying that contemporary debates about climate change are debates
about society and not nature.
What does evolution mean for society? Luhmann argues that such
things as physical reality, chemical and organic processes, and even
neurophysical and cognitive phenomena constitute the environment of
society.129 We have had occasion to cite Luhmann’s remarkable exclusion
of human beings from society more than once. But this claim disregards
the fact that all these apparently non-communicative phenomena are
constructed in communication and endowed with meaning. Otherwise,
129 “The thesis of self-production by communication postulates clear boundaries
between system and environment. The reproduction of communications from
communications takes place as society. All further physical, chemical, organic,
neurophysiological, and mental conditions are environmental conditions. Soci-
ety can substitute for them within the limits of its own operational capabilities”
(Luhmann 2012; xiii).
From Systems to Actor-Networks
172
we couldn’t discuss them or describe them in any theory. We couldn’t
take measures to influence them. And we couldn’t manipulate them
through technology so they do not select against our projects. The fact
that supposedly extra-communicative phenomena are constructed
by communication means they could be constructed otherwise or not
at all. An organism cannot do this. If an organism does not construct
information that is “viable,” it vanishes. It must adapt to the constraints
of the environment or perish. The social system has no environment
in the sense that organisms do or even in the same sense in which the
functional subsystems of society do. Luhmann is well aware of this
problem and attempts to solve it by introducing the concept of “self-se-
lection.” He uses this concept to explain the emergence of functional
subsystems within society and to explain cultural and social evolution.130
If the concept of “self-selection” is not merely another word for self-or-
ganization and autopoiesis, then it is an oxymoron, a contradiction,
because self-selection is no selection at all in the sense of evolutionary
and environmental selection. As Wittgenstein pointed out in his famous
argument against the possibility of a private language, if one claims that
whatever I say is right is right – this would be self-selection – then there
is no longer any right or wrong because the very nature of selective
constraints is that they are not arbitrary and are external to the system.
How can there be evolution in any sense of the term if no external
constraints are selecting which forms of system organization are viable
and which are not?
In distinction to the social system, which excludes nothing, the codes
of the functional subsystems exclude themselves. The distinctions
between profit/non-profit, legal/illegal, certified/not-certified, true/
false, and so on cannot code themselves, and they cannot code each
other. These distinctions are not themselves either profitable or legal
130 See Luhmann (1995:190) with regard to subsystems, and (2002:407.) for a dis-
cussion of social, cultural and political evolution. Luhmann explicitly gives up
applying Darwinian evolution to society and claims that natural selection is re-
placed by building up internal structures reinforced by the systems own feed-
back. But building up internal structures is nothing other than self-organization
and stabilizing autopoiesis. This makes it impossible for the theory of social sys-
tems to use the term evolution.
Meaning as a System 173
or true or certified; they make it possible to decide what is profitable or
not, legal or not, certified or not, within the domain of the system from
which these decisions are made. The codes of the functional subsystems
construct the borders and limits of the subsystems, and because these
are system boundaries, they do not fall within these subsystems. Also,
the subsystems effectively ban all other forms of communication than
those their binary distinctions can code. The functional subsystems
do not allow information to come from outside system boundaries.
If the legal system bans certain forms of environmental pollution or
racial discrimination in employment, this is not information for the
economic system but perturbations to which the economic system must
react within the abilities of its own information construction. It reacts
by pushing costs onto the consumer or developing business models
for pollution control. The functional subsystems can function precisely
because they are operationally and informationally closed systems.
They are confronted by perturbations from an environment consisting
of communications that are coded differently, that do not make “sense”
within their constitutive borders, and to which they must adapt by
means of internal information construction and system conform oper-
ations. To speak of a social system beyond the functional subsystems
implies that these are somehow integrated into a social whole, a super-
system. How is this possible?
It would seem that the biological and evolutionary idea of environment
and selection by the environment does not apply on the level of emer-
gent order of meaning. Therefore, it appears that functional differenti-
ation, which makes society as a whole the environment for the various
functional subsystems, at the same time makes “ecological communi-
cation” impossible. Once society has become functionally differenti-
ated, no unified and encompassing society can contain its functional
subsystems as a kind of supersystem. Luhmann attempts to solve this
problem by appealing to interpenetration. But, as we have argued
above, interpenetration is merely another word for communication.
Appealing to interpenetration simply assumes there can be communi-
cation where there cannot. Other ways of dealing with this problem
are equally unsuccessful. For example, if one compares society to an
From Systems to Actor-Networks
174
organism consisting of functional organs, such as heart, liver, stomach,
lungs, etc., it is highly questionable if the metaphor is applicable. The
body’s organs are not autopoietic, self-referential, and operationally
closed systems, for whom the rest of the body is merely environment.
The cells of a multicellular organism are not informationally closed to
each other.131 What kind of integration is needed to solve the theoretical
problem of systems that are, on the one hand, constitutively and radi-
cally distinguished from the environment and systems that exist within
other systems as functional elements? In the case of meaning systems,
it can be asked if meaning can be conceived as a system or instead as
some kind of order in which certain types of systems can emerge and be
maintained. This leads to the question of the unity of this encompassing
form of meaning. The answer to this question in modern thought has
usually called forth theories of rationality. But as Luhmann (1995:477)
admits, “Modern societies’ principle of differentiation makes the ques-
tion of rationality more urgent – and at the same time insoluble.”132
Another attempt to solve the problem of the unity of the system of
meaning under conditions of functional differentiation could refer to
the status of semiotic coding. Is semiotic coding, which organizes all
communication and meaning, fundamentally systemic or not? Is the
semiotic code a system code? Or is it a code for some other kind of
order? Luhmann is very clear that language alone is insufficient to
order social communication. Language is structured by the equally
probable options of acceptance or rejection of communication, of yes
or no. Successful communication, however, depends upon acceptance
and further communication. This is why the emergence of symbolically
131 Levin (2022) has shown that biological electical networks inform indivdual cells
how they are to cooperate in order to build organs or morphological structures.
Cells to not do what they want, but what they are told to do.
132 Luhmann solves the problem by reinterpreting rationality as biological adapt-
ability. “Translated into the language of causality, this idea [rationality, AB/DK]
decrees that a system must control its eects on the environment by checking
their repercussions upon itself if it wants to behave rationally. A system that
controls its environment in the end controls itself” (1995:475). This is also the
basis of Levin’s (2022) concept of intelligence and cognition which he nds in all
living systems. It should be noted that this is a biological denition of rationality
that has nothing to do with meaning.
Meaning as a System 175
generalized media and functional subsystems becomes necessary; they
motivate acceptance and guarantee, as far as possible, the continuing
autopoiesis of the communication system. Interestingly, Luhmann often
seems not to understand meaning as a system but rather as a medium
in which systems emerge, specifically, the psychic and social systems.
Both are meaning systems and depend on language and semiotic
coding. We have already noted severe problems with this view since
it supposes that psychic systems use language but do not communi-
cate. We argued above that this makes no sense and that the distinction
between psychic and social systems is not well-founded. But beyond
the problem of distinguishing between psychic and social systems,
there is the problem that meaning is a level of emergent order and not
a system in itself. Could the unity of the social system be achieved by
thinking of society as a whole as synonymous with a level of emergent
order in which the social subsystems arise?
Following this line of thought, one could argue that meaning is like
life. Is there a system of life or only living systems? Life is what all
living systems somehow participate in. Therefore, life could be thought
of as a kind of medium in which organisms emerge, a medium with
its own general principles, such as autopoiesis and evolution. But is
life a medium? Luhmann understands a medium as a state of loosely
coupled elements. Systems arise as “form” within a medium; they arise
as closely coupled elements. From a systems theoretical perspective,
form means that self-organizing processes of selection, relationing, and
steering construct a system/environment difference that constitutes an
autopoietic, self-referential, operationally and informationally closed
system. There can be no communication between system and environ-
ment, and thus, no communication between form and medium. How
could meaning as a medium still allow communication between the
social subsystems?
Can one think of psychic and social systems as different kinds of systems
in the medium of meaning? The problem here, as with theorizing life as
well, lies in the constraints of systems theory itself. Systems theory has
no basis for the concept of medium other than environment. Systems
From Systems to Actor-Networks
176
emerge in distinction to an environment. Neither life nor meaning can
be equated with the environment in which either living systems or
systems of meaning emerge. Life and meaning are characteristics of the
system and not the environment. Just as it is possible and perhaps neces-
sary to consider all organisms as expressions of life, all linguistic and
communicative operations are expressions of meaning. The possibility
remains to theorize life and meaning as levels of emergent order, as we
have suggested. The difference between life and meaning lies in the
specific characteristics of these levels of emergent order. For meaning,
the particular characteristic, as opposed to life, is that meaning contains
its environment within itself. Meaning is not subject to any natural
selection by the environment. Meaning is not exclusively directed
toward adaptation. Meaning is not structurally coupled to anything
other than itself. We will discuss other differences between life and
meaning as different levels of emergent order below. For the moment, it
is important to note that unlike living systems, of which there are many,
there can be only one meaning system. There are many different organ-
isms, but only one system of meaning exists. Indeed, Luhmann argues
that there is only one all-encompassing global social system since all
communication can be connected to all other communication without
any territorial or other boundaries. The social system, for Luhmann, is
world society. It is, therefore, questionable if one can speak of social
systems in the plural or only of the one, all-encompassing world system.
Just as meaning is world, so is society. These issues and questions will
accompany us throughout this book. They point to problems within the
systems paradigm and toward solutions that may lie in the network
paradigm. But let us not get ahead of our story.
According to Luhmann, each functional subsystem of society can only
be an environment for the other subsystems. The problem of overall
social communication can, therefore, only be solved according to the
model of double contingency, i.e., of structural coupling and coordi-
nated behavior. Understanding “subsystem rationalities” does not, in
principle, go beyond the model of structural coupling and consequently
does not add anything new to Luhmann’s model of society. One
subsystem, for example, the legal system, does not communicate with
Meaning as a System 177
the other subsystems, e.g., the economic system, but generates “pertur-
bations” in the environment of the economic system in such a way that
the economic system reacts to them according to its own constitutive
coding. This is true of the relations between all subsystems. That this
model cannot replicate the emergence of meaning – unless, of course,
one presupposes commonly accepted structures of expectation, which
is explicitly denied here – seems to be confirmed by Luhmann in that it
follows from his theory that communication addressed to the whole of
society is impossible in a “polycontextural” society:133
Fundamentally, any attempt to make the unity of the system
the object of an operation of the system runs into a paradox; for
this operation must exclude and include itself in the process. As
long as society was differentiated according to center/periphery
or according to a hierarchy, at least positions could be fixed in
which it is possible to ‘represent’ the unity of the system without
competition, namely in the center or at the top of the hierarchy.
The transition to functional differentiation destroys this possi-
bility by leaving it to many functional systems to represent the
unity of society through their respective subsystem/environ-
mental differences, exposing them to competition among them-
selves for which there is no superior standpoint of super-repre-
sentation. ... ...the unity of society is then nothing other than this
difference of functional systems; it is nothing other than their
mutual autonomy and non-substitutability. (1990:216)
Luhmann’s diagnosis of communicative disintegration in modern
society is valuable as a problem analysis. Undoubtedly, the diversity of
societal communication is so complex that it results in the formation of
subsystems within society as a whole, quite apart from cultural, ethnic,
political, ideological, and moral pluralism, or what generally has been
called the postmodern condition. It stands to reason that symbolically
generalized media can significantly reduce communicative complexity
133 Luhmann orients himself on Gohard Günther’s theory of polycontexturality
hps://de.wikipedia.org/wiki/Polykontexturalit%C3%A4tstheorie which at-
tempts to formalize relations between radically dierent perspectives.
From Systems to Actor-Networks
178
in this situation. Nevertheless, the notion of a subsystem itself has no
meaning if the overall system is replaced by it, i.e., if “the unity of
society...is nothing but this difference of functional systems.” The only
“system” left are the many functional subsystems, and society becomes
nothing but the sum of their differences. In this case, systems theory
offers the same vision of fragmentation and disunity as postmodern
deconstruction. The internal differentiation of society documents only
the internal disintegration of society and allows complexity to grow
to such an extent that it can no longer be managed. The problem of
over-complexity, it would seem, is “solved” at the cost of generating an
even more severe problem. Differentiation leads to a situation in which
there is no longer any social communication as such, i.e., no longer
social unity. This is not only a problem for (post)modern society but a
problem for the theory, which seems to have lost sight of its constitutive
object, the social system.
3. Society as a Whole does not have the Unity of a System
The idea that there are subsystems within systems is problematic for
systems theory since it must explain how the relation between system
and environment that is constitutively necessary to all systems can
be conceptualized if systems are somehow contained within other
systems. We saw that systems construct their own elements and relate
them in certain ways to reduce complexity and achieve some goal. In
the most general sense, it can be said the goal of all systems is to reduce
complexity by establishing a certain form of order and then operating
such as to preserve this organization and maintain the structures and
operations that have emerged. All living systems operate so that they
can continue to operate, that is, continue their autopoiesis. One could
say that all autopoietic systems have only one goal: to maintain their
autopoiesis by operating to attain their specific set points. The elements
of the system are necessarily functional. The legs and top of a table,
whatever they may look like or whatever material they are made of,
function as legs and top, or they are not elements of the table system.
However, if there are systems within systems, these cannot be merely
functional elements like the top and the legs of a table. As autonomous
Meaning as a System 179
systems, they must be allowed to pursue their general systemic func-
tion without regard to the system they are somehow “within,” which
attempts to ascribe them to a specific function. Indeed, what does the
word “within” mean when it is said that internal differentiation repeats
the system/environment difference within a system? The autopoiesis of a
subsystem must be somehow aligned with the autopoiesis of the larger
system whose internal differentiation the subsystem represents. How
is this alignment to be understood? One explanation is to claim that
subsystems set their goals to align with the goals of the larger system
in which they operate. The problem with this solution is that one must
explain how the subsystem knows the goals of the larger system when
the larger system is environment for the subsystem. There can be no
communication between system and environment. If the subsystem
cannot know to what it must align its goals, it must rely on variation
and selection, that is, structural coupling, to achieve alignment. In this
case, the larger system, that is, society as a whole, is merely a source
of constraints, indeed, an environment, selecting for or against the
operations of any subsystem. Subsystems would have to be modeled as
structurally coupled organisms in an environment or “ecosystem.” We
argued above that there can be no “ecosystem” because there can be no
system of everything. But if the unity of society cannot be understood
as the closure and unity of a system, what kind of unity is it?
Let us ask again what it means to say that a system is within another
system. To begin with, let us note that in a certain sense, it cannot be said
that a system is “in” an environment since the environment is all that
is excluded when the system/environment difference is constructed.
From the point of view of the system, outside, there is only environ-
ment and no other systems. From the point of view of an observer, the
system, for example, the frog, can be understood to be “in” the pond
and interact with other organisms. The observer, however, is not the
system; the observation does not constitute the operational and infor-
mational closure as well as the internal differentiation of the system.
How, then, is the unity of systems internally differentiated into subsys-
tems to be conceived? Only an observer can see the unity of the differ-
ence of system and environment? But when the unity of the system and
From Systems to Actor-Networks
180
the environment is in question, how can this be conceived if not as an
ecosystem? One could suppose that society constructs the subsystems
of which it is composed as “organic” functions, much like an organism
constructs a cardiovascular system, a central nervous system, or a
digestive system, or much like the cells of a multicellular organism form
tissue, etc. On the biological level, the body’s various organs depend
upon the organizing principles, the DNA, and morphogenetic patterns
of the living system as a whole. For this reason, it can be said that the
whole is more than the sum of the parts and, indeed, something entirely
different. For this reason, the organs and the individual cells do not go
off on their own and become pathogenic. For this reason, it is difficult
to conceptualize individual cells or organs as autopoietic, operationally
and informationally closed systems in their own right and thus their
relation to the organism and each other as structural coupling.134 The
relations between elements of a living system are not similar to the rela-
tions of closed systems to an environment. Integrating cells and organs
in a unitary living system is not adequately conceptualized as structural
coupling. Can the integration of functional subsystems into society be
conceptualized in this way? And if structural coupling cannot explain
the unity of an organism, how can it explain the unity of society?
Concerning the social system and the functional subsystems that
compose it, it must be asked what is the DNA, or the morphogenetic
pattern, the “code” of society that determines what functional elements
it should construct. For society, there is no analogous code. The semi-
otic code has nothing against contradictions and absurdities. There is
nothing that cannot be said in some way. Consequently, no analogous
environmental constraints could select or “force” the construction
of any particular function for a system of meaning. On the contrary,
an organism must take up energy by eating and acquire oxygen by
breathing, and some circulatory system or metabolic processes must see
to it that nutrients and oxygen are distributed throughout the body. It
134 Not even Levin (2022) can maintain this kind of autonomy for individual cells in
a multicellular organism since these cells must be informationally open to other
cells in order to cooperate and build organs or tissue etc. and follow the morpho-
logical paern of the organism.
Meaning as a System 181
must be able to recognize danger and react accordingly. But what does
a meaning system have to do other than create meaning? And who or
what is to say how this will be done or not done? Of course, everyone
knows what makes sense and what doesn’t, and everyone quickly
points out that what others say and do is meaningless. But this use of
the difference between sense and nonsense is part of how we make
meaning and not an actual boundary of meaning. Beyond this, what
kind of economic system, forms of education, government, science, art,
or religion are needed for society to be “viable?” Is a hunter-gatherer
society less viable than a modern industrial society? Is one religion or
worldview more viable than another? How could we judge? What time
scale is required? What criteria apply?
Of course, meaning does not exist independently of the physical and
biological constraints from which it emerged. There are fundamental
constraints. But what they are and how they may influence the order of
society is not given, and as mentioned above, such constraints have no
selective function. The force of gravity, it might be objected, constrains
physical and biological systems and meaning systems as well. But does
it? Has society not found different ways to “defy” gravity? What tech-
nologies of the future may bring, we can only speculate. The theory of
emergent order implies that higher levels of emergent order integrate
lower levels and contextualize and condition them, not vice versa. As
history demonstrates, society is infinitely variable, and the future is
open. Where are the principles of variation and selection that would
make evolution for meaning systems possible? The biological analogy
seems unconvincing. There can be no social or cultural “evolution” in
the technical sense. Although such talk is common and, for the most
part, unquestioned, it would seem that it has neither theoretical basis
nor empirical justification.
As opposed to organisms, meaning systems and society do not achieve
closure through exclusion but through inclusion. Any distinction that
could be made between the social system and its environment neces-
sarily falls within the system, that is, within society. Business is not
politics, and education is not art, religion, or healthcare. But society
From Systems to Actor-Networks
182
is society. Where or what is the environment? Luhmann, as we know,
says it is individual psychic systems. But we have attempted above
to show that this idea is not tenable. For business, one could say that
politics is the environment. Political decisions perturbate business,
and businesses must construct information out of these disturbances
to continue operating in the new regulatory environment. But what
is the environment of society as a whole? Nature, both physical and
biological, human individuals, and technologies and artifacts of all
kinds Luhmann wishes to locate outside society, but all these, as well
as persons, are social constructs. They are not outside of society. They
exist “as” (hermeneutical “as”!) what they are in meaning. Ecosystems
exist only in meaning since they are constructs of an observer. If society
has nothing outside, no environment, it is impossible to describe the
operations of the social system as a whole in terms of function, that
is, in terms of adaptation to an environment. What complexity, other
than its own, does meaning reduce? An organism with an appropriate
sensory-motor system, central nervous system, etc., can be judged
viable if it constructs information out of environmental perturbations
to continue its autopoiesis. If lions hunt zebras instead of butterflies,
this is because their principle of organization, their code, allows them
to construct appropriate information out of everything that is moving
in the environment. If this code functions appropriately, the lions
will find zebras, eat them, and survive. Society is not like a lion that
lives in a particular environment to which it must adapt or perish.
Society is only confronted with moral constraints on what is suitable
for communication and what is not. Despite moral concerns, one can
talk about anything, and in fact, one does. The autopoiesis of society
is the communicative construction of meaning. If this stops, then no
one is there to tell the story. If it continues, there is no way of judging
whether this or that way of communicating is functionally better or
worse than any other. Even if every religion, worldview, and ideology
claims they know the answer to the best meaning, there is no way to
know or decide whether a particular society is viable except after the
fact. Civilizations, as history teaches, rise and fall. Religions and philos-
ophies are founded, flourish, and then disappear. Apart from always
premature judgments about history, who can say which culture, world-
Meaning as a System 183
view, language, religion, nation, or people are more meaningful, more
communicable, and therefore more viable than another? A pessimist or
perhaps a realist would say that history decides. This is indeed so, but
until history makes its judgment, assuming those coming later are in a
position to appreciate it, what environmental constraints, conditions,
limitations, or resistance does communication have to conform to in
order to be viable?
Of course, the entire ecological discussion, with its supposition of unity
based on structural coupling, can be interpreted as a question of the
viability of a particular kind of society within a particular interpre-
tation of the natural world. Nevertheless, we are not dealing directly
with nature but with different interpretations of nature. If nature didn’t
have its advocates to “speak” for it, there would be no ecological issue.
Ecology and what it stands for is a conflict within the domain of meaning
and not a conflict between society and an otherwise mute nature. If the
outside is inside for systems of meaning, we must conclude that perhaps
questions of some kind of environmental adaptation of society to nature
are misplaced. These questions are nonetheless important because they
show that society cannot be modeled as a system. If society cannot be
modeled as a system, then even if society consists of communications,
these communications cannot be described as operations of an autopoi-
etic, operationally and informationally closed system. Communications
cannot be understood as functional elements or even functional subsys-
tems because communication’s only function is to continue to commu-
nicate. All it needs to do is to refer one communication to another, one
meaning to another, and so on. In this process, communication and
meaning do not run into any boundaries other than its own making.
History shows us how well society does at blocking communication
and how these blockages have always been overcome. We do not live in
a global world today because communication has successfully blocked
communication. The fact that we do live in a global world indicates that
if is a unified whole, this cannot be the unity of a system, for there can
be no system of everything.
Part II
The Network Paradigm
Chapter 3
Network Science
3.1 Network Science
Networks have a long history.135 Recently, what has become known as
“network science” or even the “network paradigm” is often considered
a branch of general systems theory.136 Although we disagree with this
view, it seems quite plausible at first glance. Let us recall that one of
the fundamental principles of systemic order is relationing that is,
constructing relations between the elements of a system. If one inter-
prets network science as the study of relations, much as cybernetics
can be understood to focus on the principle of steering or control,
then network science appears to be a part of general systems theory.
Network science can be seen as the study of the relations among the
elements in complex systems. The complex system that is being studied
is resolved into nodes and links, which represent the connections or
relations between elements, whatever they may be, whether proteins or
people.137 Also, similar to general systems theory, there are two major
methodological approaches to modeling networks. Network analysis
can model the nodes and links quantitatively or qualitatively; of course,
there are mixtures of the two. The quantitative and, therefore, math-
ematical modeling of networks is dominant in the natural sciences,
whereas qualitative network analysis is typical in the social sciences.138
135 For a detailed timeline going back to 1736 when Euler solved the problem of the
bridges of Königsberg see Lewis (2009:3-4). It is interesting to note that although
network science often speaks of systems, it shares practically no common history
with general systems theory and cybernetics.
136 See (Lewis 2009:7) for a typical statement of this view. “…network science is
essentially the science of systems … networks often model complex systems…”
137 For an overview of what kinds of topics can be investigated from the network
perspective see the collection of publications as hps://barabasi.com/publica-
tions. A collection of resources for network science can be found at hp://www.
network-science.org/.
138 See Hidalgo (2016) for an overview.
From Systems to Actor-Networks
186
As Cerquite and Shi (2022:1) remark in a recent collection of essays on
network science:
In general, networks are mathematical models that describe
unified frameworks composed of disaggregated compo-
nents—the nodes—by including components’ interconnections.
Network models are then a step beyond the simple case of
systems collecting different unrelated entities; indeed, the links
between the nodes provide unique information on the investi-
gated phenomena.
The qualitative form of network science is usually represented by
Social Network Analysis (SNA). SNA studies the communicative rela-
tions among persons in an organization or defined group.139 Generally,
network science, like systems science, is an interdisciplinary field of
study, including the social sciences and physics, biology, computer
science, and many related disciplines. Lewis (2009:6) summarizes this
situation succinctly:
each subfield using network science had a different working
definition. Communication engineers think of networks as
systems of routers and switches; sociologists think of networks
as influence diagrams representing the social interactions
among humans; marketing business people think of networks
as populations of buyers; and the physicist thinks of networks
as models of phase transition, magnetism, and so on. Biologists
use the network metaphor to understand epidemics, genetics,
and metabolic systems within cells, and power engineers think
of electrical power grids. Network science appears to be in
the eye of the beholder with different nomenclature, different
vocabulary, and different methods of analysis in each field.
Despite claims to universality and interdisciplinary scope, it can be said
that network science, like systems theory, is characterized by a division
between the natural and the social sciences, that is, sciences of phys-
139 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Social_network_
analysis.
Network Science 187
ical and biological phenomena and sciences of meaning. Furthermore,
because of the different disciplinary and methodological perspectives, a
specific terminology has arisen, which is found in network science and
does not depend on concepts derived from systems theory. Nonethe-
less, some shared concepts exist between general systems theory and
network science. Among the most common are “complexity,” “emer-
gence,” and “evolution.” Despite using these terms from systems theory
and occasionally referring to networks as systems, network science can
do quite well without them as they seem to have no well-defined and
well-founded usage in network theory. With the exception of actor-net-
work theory, which we will discuss below, network science does not
usually refer to networks as agents. Although self-organization is some-
times applied to networks, for example, in self-organized criticality
theory,140 self-organization does not imply agency. In social science
network theory, the nodes of the network are indeed often people who
are agents, but the network is not an agent. Network science, except for
actor-network theory, as we shall see, does not inherit pan-agentialism,
that is, the view that everything is an agent of some kind, typical of
general systems theory. Pan-agentialism is an important shared view
between general systems theory and actor-network theory.
When speaking of networks, one often speaks also of graphs. A graph is
usually a visual and mathematical representation of a network. The two
terms are, for the most part, interchangeable.141 Instead of elements, as
in systems theory, networks are made up of nodes (also called vortices
or actors). The nodes are connected by links (also called edges, ties,
relations, or associations). The nodes (vortices, actors) are connected in
various ways and by various amounts (the number of links connected to
a node is called a degree) by the links (edges, relations, ties, associations)
that are described by network modeling. Like systems, networks are
140 See Wikipedia hps://en.wikipedia.org/wiki/Self-organized_criticality.
141 For this reason, Steven Wolfram’s (hps://writings.stephenwolfram.
com/2020/04/nally-we-may-have-a-path-to-the-fundamental-theory-of-phys-
ics-and-its-beautiful/) multi-hypergraph model of reality can be considered a
network model.
From Systems to Actor-Networks
188
not static structures. They are dynamic and changing.142 The dynamics
of networks can be described on two levels: on the level of “micro rules”
governing linking mechanisms and on the “macro” level in which
emergent properties and behaviors of the network become apparent.143
Nodes can appear in various topographical configurations, sometimes
called clusters or communities. There are also “scale-free” networks which
contain hubs. A hub is a node often said to arise due to “preferential
attachment,” which leads to one particular node becoming connected
to more nodes than other nodes share. Hubs are network nodes with
many more links than other nodes.144 In the social sciences, such hubs
are often known as “intermediaries,” nodes that have great influence
since actors must go through them to link up with other actors in the
network. Actor-network theory speaks of all actors as potential “medi-
ators” and hubs as “obligatory points of passage.”145
It is important to note that network theory is not based as is systems
theory on a constitutive difference between system and environment
and, therefore, not on operational and informational closure. Networks
are not closed systems. Instead, networks are constituted by relations,
connections, and openness to changes in configuration and extensions
of relations. The difference between systems and networks becomes
immediately apparent in the illustration below.
142 This aspect of networks is usually where talk of “evolution” comes into network
science, but the term should not be taken as a reference to variation and selec-
tion as in systems theory but simply a word for the fact that networks grow and
change. As mentioned above, change does not imply agency.
143 As Lewis (2009:6) puts it: “behavior of nodes and links is dened by a set of mi-
crorules governing the behavior of nodes and links. These rules are given at the
microlevel, to distinguish them from macrolevel behaviors of networks. Specif-
ically, microlevel rules dictate the behavior of links and nodes, and macrolevel
rules dictate the emergence of global properties of a network.”
144 See Barabasi (2014) for the concept of “preferential aachment” as well as an
introduction to network science and its terminology.
145 See Callon (1986). We will discuss the terminology of actor-network theory in
relation to network science in detail below.
Network Science 189
System Environment
Network146
It should be evident from this illustration that we are dealing with two
entirely different things when speaking of systems and networks. None-
theless, a network could be found within a system, as with cell protein
replication networks. In this case, the network is a functional element
– not a subsystem – of a system. On the other hand, there can be systems
within networks, as different companies within a supply chain network
are each individual closed organizations or systems linked together
by open and changing relations. We will argue that Luhmann’s func-
tional social subsystems should be understood this way. Concerning
society as a whole, Luhmann’s social subsystems are indeed systems
but not within a larger social system. Society is an open network, but
the functional subsystems can still be modeled as closed systems. The
economic system, for example, functions within the open network of
society as a whole. We will return to the differences between network
theory and systems theory below. For the moment, we are concerned
with a general description of network theory, which can be the basis for
comparing the two paradigms.
As noted above, there are two basic approaches to describing networks:
quantitative and qualitative. Among the two approaches, the quantita-
tive approach abstracts from what is connected, the nodes, and focuses
on the numerical aspects of the relations between the nodes. It is the
intensity, distribution, topology, and regularities of the links that are
the object of study, regardless of what the nodes may be. Typical of the
146 Image from arronparecki cc by 2.0 hps://openverse.org/image/ca55d60a-f05f-
4a41-8e71-582d405d3e8d?q=network
From Systems to Actor-Networks
190
natural sciences, the goal is objectifiable, general regularities and not
individual characteristics of the things related. Quantitative network
theory abstracts from individual characteristics and focuses on quan-
tifiable, generalizable relations. The method is, therefore, usually
mathematical modeling and topological, that is, graphic representation.
Typical examples are the graph of Internet traffic or the spread of a
disease. On the other hand, the qualitative approach is also concerned
with the relations among nodes in networks but derives them from the
nature of the connected nodes. The question is not primarily how links
are distributed but what the network is made of. The focus is on whether
one is studying people, emails, organizations, political parties, etc. Each
type of node makes a difference. Depending on what the nodes are, the
relations have different meanings. It is not simply how everything in
the network is connected; it is a question of what is being connected
and why.
Granted these two perspectives, it can nonetheless be said that just as
systems theory sees systems everywhere, network theory also finds
networks everywhere. After all, what is not a composite of some kind
with components connected in some way? As Barabasi (2014:5) notes:
Networks are also at the heart of some of the most revolutionary
technologies of the 21st Century, empowering everything from
Google to Facebook, CISCO, and Twitter. At the end, networks
permeate science, technology, business, and nature to a much
higher degree than it may be evident upon a casual inspection.
Consequently, we will never understand complex systems
unless we develop a deep understanding of the networks
behind them.
The graph, or the representation of a network, can be used to model
almost everything from physical processes to protein networks to social
networks. Just as systems theory, network theory also claims univer-
sality; it claims that network structures are the basis of all phenomena.147
147 See Lewis (2009:8) “…advocates of network science began to demonstrate the
universality of network science as they applied it to diverse elds that seemingly
had no relations to one another.”
Network Science 191
Examples of networks are electrical power grids, online social networks,
neural networks, both biological and electrical, telecommunication
networks, transportation networks, conglomerates, Internet traffic
and the network effects of monopolies, markets, economic sectors
and regions, protein interaction networks, food networks, epidemics,
industry and technology networks, small world networks, etc. When
networks are visualized, what is seen is a pattern that is usually a highly
complex branching structure but not random. When the methodology
is quantitative, what is seen is usually networks that follow a power
law distribution explained by “preferential attachment.”148 It is also
seen that networks in various areas, whether it be physics, biology, or
sociology, exhibit specific common characteristics. This justifies talking
about a “network paradigm,” which claims that network order is to
be found in almost all phenomena no matter what level of emergent
order.149
Since our concern in this book is the universality claim of the systems
paradigm and the network paradigm regarding self-explanation, we
will focus on the question of meaning. Can meaning be successfully
modeled as a network? Can network science explain itself? Questions
such as these imply that the purely quantitative approach will not be
the center of our interest. We will not primarily be interested in the
distribution or topography of network links but in how meaning condi-
tions both nodes and links. As a theory of meaning, network theory
cannot abstract from what is being related, from what it means to be an
actor in a network. Our guiding questions will be: What makes nodes
and links in a network meaningful? Why does meaning arise from
associating nodes with each other via links? The question of meaning is
not primarily quantitative unless the amount of links affects what and
how the network constructs meaning. We are interested in the meaning,
not how many connections a word, idea, artifact, etc., may have with
148 Preferential aachment means that nodes which have more links tend to acquire
more links and become hubs. Barabasi (2014) has inuentially proposed this
model and demonstrated it in many typs of networks.
149 As an example, see hps://www.sciencedirect.com/topics/computer-science/net-
work-paradigm.
From Systems to Actor-Networks
192
other associated meaningful elements. Nevertheless, in some cases, as
we shall see when discussing actor-network theory, the number of links
can condition meaning. But what is primarily important for a theory
of meaning is what a “node” is and means and not solely the number
of connections it may have to other nodes.150 Our question is to what
extent network science offers a coherent theory of meaning.
3.2 Actor-Network Theory
The network science literature in the natural and social sciences offers
no theory of meaning. It is assumed that one knows what one is talking
about, mathematically modeling and describing. The meaning of what
constitutes a network from quantitative and qualitative perspectives is
simply taken for granted and not considered a question that network
science must address. However, this is not the case for actor-network
theory (ANT). We admit that the canonical literature of network
science usually does not contain references to actor-network theory.
Actor-network theory stands outside the usual parameters of what is
considered network science. This could be attributed to the fact that
although ANT shares basic concepts with network science, it has very
different methods and fundamental assumptions about the nature
of what is being studied. ANT is concerned with networks from an
entirely different perspective than is “normal” network science. ANT
was developed in the sixties and seventies, primarily by Bruno Latour.
Other significant contributors are John Law, Michel Callon, Steve
Woolgar, Trever Pinch, Wiebke Bijker, and others within the broad area
of the sociology of knowledge and what has come to be called “science
and technology studies” (STS).151 The unique perspective taken by STS
and ANT developed out of questioning such fundamental distinctions
150 For an example of the quantitative approach to literature see the network anal-
ysis of Shakespeare’s dramas. hps://www.martingrandjean.ch/network-visual-
ization-shakespeare/.
151 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Science_and_tech-
nology_studies. For an overview of ANT see Wikipedia hps://en.wikipedia.
org/wiki/Actor%E2%80%93network_theory. And for an introduction from La-
tour himself see Latour (2005).
Network Science 193
as those between subject and object and society and nature. In addition,
ANT used the descriptive methods of ethnology to show that science
was deeply embedded in society and that technological artifacts were
not passive instruments but more like social partners in the construction
of knowledge. The origins of ANT in the ethnology of science explain
why all networks are social for ANT. In opposition to systems theory
and network science, ANT is not a theory of physical or biological
networks but a theory of how physics, biology, and all sciences and
their various objects are fundamentally social phenomena. ANT, there-
fore, is a theory of order on the emergent level of meaning and views
all forms of order, including physical and biological order as forms of
meaning, that is, as social phenomena.
For systems theory, the world consists of systems. Whenever order
appears in the world, it is systemic. Similarly, for network science,
order on all levels, physical, biological, and semiotic, exhibits network
characteristics. In short, reality consists of networks. This universal-
istic approach is also true for ANT. Everything is an “actor-network.”
Actor-networks, however, are distinguished not only from systems but
also from the networks that network science studies by ANT’s founda-
tional assumptions and methodology.
The first of these assumptions and methodological principles is what
could be termed relational ontology. Networks for ANT consist of actors
and their relations or “associations.” In a similar way to how the actors in
a novel reveal what roles they play and who they are only in the course
of the unfolding story, an actor-network is fundamentally a narrative
form of order in which both humans and nonhumans can become actors
who play specific roles for certain purposes.152 Just as in ethnological
fieldwork, one looks at how actors are related and then derives what the
actors are from their relations. Relational ontology, however, goes even
further by assuming that actors are themselves networks. There is no
“individual” in the literal sense of the word as something that cannot be
152 ANT speaks of “actants” which is a term derived from semiotics referring to any
role played by an actor in a narrative whether it be human or nonhuman. In a
medieval tale an actant can be a sword, the knights horse, a chalice, a castle, or
whatever can change the course of events in the story.
From Systems to Actor-Networks
194
further divided, as a “substance,” or something that could exist without
relations. ANT presumes there are no substances, atoms, individuals, or
last elements into which everything could be resolved. In this respect,
and in its own way, ANT shares systems theory’s insistence that the
elements of a system are always constructed by the system within a
system and do not exist independently as things in the environment.
Reality consists of networks nested within networks. This implies a
kind of fractal structure of reality that is fundamentally at odds with the
view of normal network science in which networks are first composed
of individual nodes that are then linked in various ways. We will have
occasion to return to the idea of relational ontology below.
A second important principle of ANT is what could be called agnos-
ticism. ANT makes no presumptions about the nature of the actors
since it is their associations that determine what they are. Agnosticism
is a methodological principle of ANT that says one does not assume
anything about what is associated in a network. As is often claimed, one
must “follow the actors” without presuppositions (Latour 1987; 2005) to
discover the network and who or what the actors are. The actors reveal
themselves to the observer by tracing the activities that construct the
network. No traditional boundaries or distinctions should be assumed
as given. Actors move freely between social domains and even between
nature and society.
A third principle of ANT is that all actors should be treated alike;
that is, there is “symmetry” regarding whether actors are humans or
nonhumans. ANT refers to this principle as “generalized symmetry.”153
Generalized symmetry requires that there are no a priori distinctions
between actors. Generalized symmetry has several important corollary
principles that characterize ANT. The fourth principle is one of these.
It can be called heterogeneity. Heterogeneity means that networks are
usually composed of many different kinds of actors. An actor-network
is composed not only of humans but of things, technologies, institu-
tions, organizations, legal procedures, documents, economic interests,
markets, territories, religious traditions, languages, etc. Networks
153 For explanations of generalized symmetry, see Latour (1996; 2005).
Network Science 195
do not respect ontological domains. They associate entities across all
sectors. For this reason, it can be said that actor-networks are always
heterogeneous and hybrid.
Fifth, and perhaps the most important implication of generalized
symmetry is nonhuman agency. This principle means that there can be
nonhumans in a network and that these material things possess agency
and can change the course of a network. Generalized symmetry and
heterogeneity imply distributed agency. Technologies, for example, are
not assumed to be mere tools, instruments, or passive objects. Instead,
they influence the construction of actor-networks, possess their own
forms of agency, often called “affordances,” and must be considered
social partners. Already, general systems theory introduced the idea
that order is processual and dynamic. ANT carries this notion a step
further by distributing agency across all boundaries and not limiting it
to functionalism. Agency, for ANT, is not equivalent to function or the
operations of a system. It is a much broader concept that includes any
form of construction of associations, that is, networking.
This leads to the sixth principle characteristic of ANT, namely, the
priority of process over structure. Networks, like systems, are dynamic.
Networks are not static structures but dynamic processes of networking.
Indeed, one should probably not speak of networks at all but of networks
networking. The word “network” is primarily a verb and not a noun.
The processual nature of networking, the consequent displacement of
structure, and the accompanying goals of maintaining structure typical
of systems make networks flexible, contingent, continually changing,
and unstable. Not only do networks change constantly, but when they
change, the actors also change. ANT’s relational ontology prevents
actors from being understood as substances with fixed identities. When
relations change, human and nonhuman actors change roles and iden-
tities.
Finally, ANT fundamentally differs from systems theory by assuming
only one kind of network exists. General systems theory assumes the
existence of physical, biological, and social forms of systemic order. On
From Systems to Actor-Networks
196
the contrary, ANT assumes that all forms of order are actor-networks
constructed with “society.” However, for ANT, society is not what it
is usually understood to be. Society is not one ontological domain in
distinction to the domain of nature. The distinction between society
and nature is a construction of typically “modern” actor-networks and,
thus, a historical artifact that can change. ANT is not a theory of physics
or biology but a theory of society that insists that physics and biology
must be understood as social phenomena, as actor-networks. ANT
maintains a universality claim by arguing that there is nothing outside of
society, that is, outside of actor-networks. This claim is understandable
if society is understood as meaning and not merely the particular ways
in which certain intelligent apes organize their cooperative activities.
Meaning is all-inclusive. There is nothing outside of meaning. From the
ANT perspective, this implies that there are only actor-networks.
One reason why ANT is not usually included in network science lies in
the fact that ANT has developed its own terminology to describe what
networking means and what an actor-network is. It was mentioned
above that network science generally refers to various mechanisms of
network construction. Scale-free networks, for example, form through
preferential attachment. Other networks exhibit clustering mechanisms
or various topological formation principles that influence network
structure. One speaks of “micro rules” conditioning how nodes are
linked together and macro structures that emerge as network proper-
ties. On the level of micro rules, ANT uses the concepts of “translation”
and “enrollment.” On the level of network structures, ANT speaks of
“trajectories” or “programs of action.” We will explain what these terms
mean in the discussion below.
In contrast to general systems theory, ANT and network science do not
share the mythical opposition of chaos and order underlying systems
theory. Networks do not arise as solutions to the problem of complexity
or as islands of order swimming in an ocean of chaotic complexity.
For networks, complexity is not a problem but a solution. Networks
consequently do not have an environment from which they are consti-
tutively distinguished. There is no network/environment difference.
Network Science 197
For network science, the world consists of nodes and links; for ANT, the
world is constructed out of actor-networks. ANT, therefore, does not
project the mechanisms of biological adaptation and evolution onto the
level of meaning. There are no biological metaphors in ANT. In contrast
to systems theory, for ANT, meaning cannot be modeled as the emer-
gence of systemic order. Instead, meaning emerges as an actor-network.
Network order replaces systemic order. An actor-network is not an auto-
poietic, operationally and informationally closed system that could be
classified as physical, biological, or social. Every actor-network is social.
Things are social. Technologies are social. Meaning emerges from the
mutual and symmetrical agency of humans and nonhumans. Meaning
emerges not as a system on the level of human consciousness, cogni-
tion, or social communication but as a network that necessarily includes
both humans and nonhumans, both things and ideas, the material and
ideal, both subject and object, and nature and society. Indeed, as ANT
proposes, actor-networks are always hybrid and heterogeneous.
In our discussion of ANT, the concept of meaning plays a central role. We
must openly admit that ANT does not see itself as a theory of meaning.154
Surprisingly, the concepts of meaning, information, and communica-
tion do not play significant roles in ANT. Nonetheless, we will attempt
to interpret ANT as a theory of meaning, and we will make special use
thereby of the concept of information. This will allow us to compare
Latour and Luhmann as proponents of theories that describe order
on the emergent level of meaning. In contrast to Luhmann, however,
Latour does not speak of self-organization or emergence and does not
consider meaning a particular level of emergent order. Where Latour
and Luhmann would undoubtedly agree is that what we are talking
about is society. But as we shall see, there are significant differences in
154 Latour (2013:266) does discuss the theory of meaning in terms of the mode of
existence named PRE. “For every mode, there is a distinct theory of meaning, a
particular semiotics. If there is an even slightly general metalanguage, it has to
be entrusted to the mode of prepositions [pre]. To dene the sense of an existent
is to identify what is lacking, what must be added in order to translate it, to take
it up again, to grasp it anew, to interpret it. consequently, in this inquiry, trajec-
tory, being, and direction, sense, or meaning, are synonyms.” We will return to
these ideas below.
From Systems to Actor-Networks
198
what each considers society to be. Returning to the central concept of
meaning, neither Luhmann nor Latour is alone in focusing on meaning.
Heidegger and Wittgenstein, for example, have also made meaning,
language, and communication central concepts of 21st-century thought.
Although Latour does not use the concept of meaning, we shall argue
that it is meaning that he is talking about. Latour’s aversion to the
concept of meaning comes from his critique of deconstructive post-
modernism as a theory of “discourse” based on signs and structuralist
semiotics. According to postmodernism, structuralism, post-structur-
alism, and hermeneutics, at least as Latour reads them, all the world is
a text. There are only signs. In structuralism, signs are ideal, virtual, and
immaterial. Latour rejects this view and asks, where are the things, the
artifacts, and the technology of which the world is made? We agree with
Latour’s criticism of structuralist, post-structuralist, and postmodern
deconstructivism and explicitly reject notions of meaning that do not
include material reality. We are not forced to accept the postmodern
version of semiotics or hermeneutics. Our understanding of meaning
is not primarily linguistic or semiotic in the usual sense of these terms
but explicitly includes things and their role in constructing meaning.155
In ANT, the inclusion of things of all kinds and, primarily, artifacts and
technology in society as co-constructors of meaning is expressed in
terms such as “hybrids,” “quasi-objects,” and “quasi-subjects.” Central
ideas of the semiotic and hermeneutic traditions, such as interpretation,
text, language, and signs, are not simply ignored in ANT. They are
taken up and transformed in terms of “articulation.”156 We will have the
occasion to explicate these and other special terms in ANT below.
From the point of view of actor-network theory, there are indeed
155 ANT is well-known as a major proponent of what has come to called “material
culture” or “material semiotics” which emphasize that things have meaning and
not merely signs or language. On material semiotics see Law (1919)
156 “Articulation does not mean ‘to be articulate’ in the sense of being able to speak
uently, but rather ‘to join together’ in the sense of constructing a composite en-
tity out of heterogeneous parts. This is why I have chosen this term to dene not
only the action of joining together but also the result of this action. Each of the
composite entities made up of heterogeneous parts will, from now on, be called
articulations.” (Latour 1999:306).
Network Science 199
systems, but these are to be understood as a certain kind of network.
The autopoietic, operationally and informationally closed systems
Luhmann describes as functional subsystems of society can be under-
stood from ANT’s point of view as particular forms of network order.
They are functional networks that have become solidified into “black
boxes.”157 Black boxes are networks in which the complex relations
between the actors have been covered over and reduced or subsumed
to fixed functions that result in standard input/output operations.
The many internal constitutive links, relations, and associations have
become invisible. We perceive only fixed processes of input and output,
which seem almost mechanical or automatic. For Latour (2005:37ff) this
means that the actors in the network have become “intermediaries”
instead of “mediators.”
An intermediary, in my vocabulary, is what transports meaning
or force without transformation: defining its inputs is enough
to define its outputs. For all practical purposes, an intermediary
can be taken not only as a black box, but also as a black box
counting for one, even if it is internally made of many parts.
Mediators, on the other hand, cannot be counted as just one;
they might count for one, for nothing, for several, or for infinity.
Their input is never a good predictor of their output; their spec-
ificity has to be taken into account every time. Mediators trans-
form, translate, distort, and modify the meaning of the elements
they are supposed to carry. (Latour 2005:39)158
From the perspective of ANT, Luhmann’s theory of social systems
does not describe fundamental forms of order on the level of meaning,
whether psychic or social, but rather black boxes within networks. The
challenge of ANT to systems theory is the claim that the concept of
“system” is not equated to the idea of “order.” Order can also appear
as networks; indeed, networks are the primary and original form of
157 “The word black box is used by cyberneticians whenever a piece of machinery or
a set of commands is too complex. In its place they draw a lile box about which
they need to know nothing but its input and output.” (Latour 1987:2-3)
158 We will return to the all-important concept of “mediation” below when discus-
sion networking.
From Systems to Actor-Networks
200
order, at least on the level of meaning. On the level of meaning, where
ANT locates all the sciences and their physical or biological objects of
study, one finds actor-networks. Systems are a particular kind of an
actor-network, which is why science and technology are embedded
within society, or as Latour often says, technology is society.159 ANT
does not tell the age-old and almost ubiquitous story of order emerging
from chaos. There are not systems on the one side and chaos on the
other, even if it might seem that systems “perceive” the complexity of
networks as chaotic and a problem that must be solved. The starting
point of both praxis and theory in ANT is more or less chaotic networks.
Within these networks, small areas or delimited islands have become
rigid, standardized, and functionalized. It is in systems that mediators
are turned into intermediaries. It could be said that what is seen as
systems are forms of order in which information has been reduced to
a high level of redundancy, and the ongoing process of the construc-
tion of meaning has become channeled and slowed down, giving the
appearance of durability, fixed order, and structure over against vola-
tile agency and its unforeseeable vicissitudes.
In the legal system, for example, redundancy is enforced by state
power and the application of the binary code of legal/illegal. In science,
redundancy results from applying the binary code true/false by means
of established and institutionalized research methods and programs.
In each functional subsystem of society, regularities are enforced by
the operational and informational closure of the system. These regu-
larities appear to sociologists as structures, institutions, organizations,
etc.160 Durkheim famously referred to these structures as social facts,
the specific objects of social science. Luhmann speaks of social systems.
From the point of view of actor-network theory, social structures, and
social systems are located on a level beneath the originary process of
networking which constructs meaning. At the basis of the social in all its
159 See for example Latour (1991).
160 Latour does not speak of social systems to describe the black boxes which law,
religion, business, politics, etc. appear as, rather, he considers these to be “modes
of existence.” See Latour (2012) for the theory of modes of existence. We will
discuss this text below.
Network Science 201
forms – which includes, from Latour’s point of view, not only humans
but nonhumans as well – is networking and not system building.
Theoretically, this is expressed in the fact that for Latour and ANT,
the basic principles of order are not processes of selecting, relationing,
and steering but processes of “translation” and “enrollment.” Transla-
tion and enrollment designate the micro-rules that construct meaning
by creating associations among actors. In and through associations,
actors become an actor-network that has meaning. Meaning emerges
not as perception and consciousness or even communication but as
the construction of actors in networks of associations. What Luhmann
describes as the cognitive and communicative operations of systems of
meaning are for ANT operations of translating and enrolling that are
performed by both humans and nonhumans. From the point of view
of network order, actors can be anything, both human and nonhuman.
Therefore, contrary to what is everywhere taken for granted, cogni-
tion and communication are not the prerogative of human agents or
persons. Let us take a closer look at what these fundamental processes
of meaning construction are.
3.3 Translation and Enrollment
The actor-networks that ANT first described were things like daily
activities in a scientific laboratory, disputes concerning fishing rights
in a small town in France, the development of an electrical railway in
France, in general, the social context of scientific discovery and the
complex processes by which technologies were adapted and integrated
into society. All these empirical studies were done according to the
methods of ethnology, and the shared focus of research was always the
interaction of humans and nonhumans, especially in science and tech-
nology. The classic example of what is meant by nonhumans is technol-
ogies, artifacts, but also animals, bacteria, fish, social institutions, laws,
and everything else that makes up society. An actor-network is always
made up of heterogeneous and hybrid actors. Generalized symmetry
and rejection of presuppositions as to who and what can become an
actor in a network allows ANT to describe associations in ways inacces-
From Systems to Actor-Networks
202
sible to traditional sociology and philosophy of science. For Luhmann,
technology, artifacts of all kinds, and every material entity must be
banished into the social system’s environment, which consists only
of communications. With this exclusion strategy, Luhmann reflects a
long history of opposing society and technology. The typically modern
division between society and nature and the material and culture has
been expressed in either technological determinism (technology deter-
mines society) or in the view that society determines technology (social
determinism).161 Latour rejects both views since, for ANT, technology
is society. Meaning is always and necessarily material, and the trans-
formation of matter into meaning does not simply add a “virtual” or
“ideal” layer onto an otherwise meaningless material world, a level of
culture onto nature, but transforms the material world and nature into
society. Society, Latour prefers the term “collective,” as with Luhmann,
is not made up of human beings, but against Luhmann, the “collec-
tive” is made up of networks of humans and nonhumans involved in
constructing associations which make meaning.
We will attempt to illustrate the theory of meaning that could be
grounded in ANT and Latour’s work by closely examining a concrete
example of an actor-network. Our example will not be any of the actual
networks that ANT has described, but an “original,” perhaps even the
very first actor-network that can be described. Let us go far back in
history, even before the appearance of human beings.162 According to
the current state of the archeological record, Homo sapiens appeared
161 For technological and social determinism see the respective Wikipedia articles.
hps://en.wikipedia.org/wiki/Technological_determinism; hps://en.wikipedia.
org/wiki/Social_determinism.
162 This allows us to answer the challenge of Meillassoux (2008) who speaks of the
“ancestral” in order to designate a realm beyond the “correlationist” presuppo-
sition of modern Western subjectivism, especially since Kant, and as exempli-
ed in Husserl and Phenomenology which inuenced Luhmann. The idea of
the ancestral questions the assumption that Being – and therefore meaning – is
necessarily correlated to human consciousness, to intentionality. Meillassoux
asks, how can Being be correlated to human consciousness when there was obvi-
ously an “ancestral” world long before human beings arose. Correlationalism is
a modern idea arising from the strict distinction between subject and object and
the necessity of then somehow puing them together again, since the one cannot
exist without the other.
Network Science 203
about 300,000 years ago.163 The first use of tools, the appearance of tech-
nology, and all that is implied by it goes back about 3.3 million years. In
this period, known as the Paleolithic Age, hominins, our earliest ances-
tors, used stone, wood, and bone tools for hunting, scavenging, and
other purposes. Stone tools can be dated to ca. 3 million years before
Homo sapiens appeared. These are the so-called “Mode 1” tools from
what is known as the “Oldowan Industry,” that is, sites found at the
Olduvai Gorge in Tanzania. These tools go back to 2.6 million years.
They were made by hammering oval-shaped stones found in riverbeds
into forms with sharp edges on one end by breaking away flakes. One
end was sharpened, whereas the other end was round and could thus
be easily held in the hand. This primitive stone ax, also called the “hand
ax,” is not the oldest example of stone tools. 3,3 million years ago, stone
tools were discovered at Lomekwi 3 in West Turkana, Kenya.164 This
discovery shows that the stone ax is undoubtedly one of the first and
even the longest-used tools in protohuman and human history. In
many ways, it prefigured later technologies and can be considered a
simple model of what it means to use tools or what technology is. When
we consider that tool use is often said to be that which distinguishes
humans from other animals and thus an essential aspect of culture, the
question of this technology could lead to a deeper understanding of
human existence than can be attained by beginning with full-blown
linguistically mediated conscious experience as does modern philos-
ophy as well as Luhmann’s systems theory.165
Since the ability to make and use tools and to transmit these skills over
many generations is considered a hallmark of culture and civilization,
what conclusions can be drawn from the fact that hominins were making
and using tools several million years before Homo sapiens appeared?
163 For an overview of the archeological record see Wikipedia hps://en.wikipedia.
org/wiki/Early_modern_human
164 See Wikipedia hps://en.wikipedia.org/wiki/Stone_tool.
165 It should be noted that with regard to scientic plausibility Latour’s story is
much more credible than Parsons’ and Luhmann’s mythology of the originary
encounter of double contingency. See for example the Material Engagement The-
ory of Malafouris (2013), the Mirror System Hypothesis of Arbib (2012), and Cor-
ballis (2003; 2017) on the origins of language.
From Systems to Actor-Networks
204
It implies that human beings, with their big brains and linguistic skills,
were not a necessary precondition for the emergence of meaning, or at
least, that meaning emerged long before the appearance of specifically
human intelligence. Furthermore, it implies that meaning is intimately
associated with nature, material reality, and nonhuman forms of life.
Let us, therefore, ask: What does it mean to use a stone ax? How did
this technology emerge? What is the difference between how hominins
made and used tools and how behavior that might appear similar in
apes at first glance never developed into culture and meaning? Despite
all the evidence of tool use among apes and other species, there seems
to be some fundamental difference between what animals do and what
could be called the non-episodic use of tools on the level of meaning.166
Let us attempt a phenomenology of the stone ax. Let us suppose we
have some privileged access to the dawn of history and can go back in
time and carefully describe, as an ethnologist might do, what happened
when the first tool was made, when technology was born, and when
meaning emerged more than 3 million years ago in Africa. Let us adopt
the methodology of actor-network theory and attempt to discover the
original actor-network, which helpfully illustrates all the essential
elements of an actor-network and how it can be understood. Of course,
this is fiction.167 But is it not the case that the emergence of systemic
order or the emergence of life is also shrouded in mystery? Are not all
origin stories, including the event of the Big Bang, in some way or other
attempts to illustrate and define first principles with the help of imag-
ination? Be this as it may, what does our phenomenology of the stone
ax reveal?
We emphasize at the outset that we are not describing the adaptation
166 Simian tool use is characterized by its “episodic” nature. See Donald (1991:149),
who describes the “culture” of apes as “episodic.” “In fact, the word that seems
best to epitomize the cognitive culture of apes […] is the term episodic. Their
lives are lived entirely in the present, as a series of concrete episodes and the
highest element in their system of memory representation seems to be at the
level of event representation.”
167 But ction is nothing new in science. See for example Luhmann’s original situ-
ation of double contingency which is just as much ction as is Hobbes’ state of
nature.
Network Science 205
of an organism to its environment. We are also not offering a phenom-
enological description of human conscious experience since we find
ourselves at a point in time 3 million years before Homo sapiens
appears. We see a process by which relations or “associations” are
established between particular things and animals so that a unique
behavior appears that must be ascribed to all of the participants. Let us
call these participants actors and note that the behavior resulting from
their association cannot be reduced to any of the actors alone. We see
the outcome of a unique and unprecedented kind of co-operation between
hominin and thing in which agency and what could be called “cognitive
function” is distributed among all actors.
We can never understand and infer the nature of the “cognitive
function” responsible for the creation and use of a tool without
first recognizing that the various processes responsible for the
transformation of raw material to tool, as well as the tool itself,
actively and reciprocally participate in the co-construction of
what counts as “cognitive function.” (Malafouris 2013:163).
In the case of the stone ax, it is a “cooperation” between a hominin and
a certain type of stone. What the hominin does cannot be adequately
understood as a “reaction” to perturbations coming from an envi-
ronment. There is no operational or informational closure. There is
no purely internal construction of information that may or may not
be viable. Therefore, we are not talking about an organism’s interac-
tion with its environment. What takes place is a unique cooperation
between hominin and stone. Concepts derived from systems theory
do not adequately describe what happened when a hominin 3 million
years ago, for the first time, picks up a particular stone from a riverbed,
holds it in the hand in a certain way, wields it with the arm in specific
ways, strikes things, etc. In this world, there is neither a system nor
an environment. Concepts such as “system” and “environment” carry
the weight of a long history of theoretical development and abstraction
from experience. Our ethnologist has gone back in time to witness and
describe this event, which is not a single event but a historical process
lasting thousands of years. Interestingly, the ethnologist does not see
From Systems to Actor-Networks
206
systems distinguished from environments. What do they see?
Our ethnologist sees a hominin picking up a particular stone in a
riverbed and holding it in various ways in the hand. The stone has a
specific shape, weight, size, and consistency. All these characteristics
of the stone suggest, offer, nudge, propose, and influence the hominin
so that the stone be held in the hand in a certain way and swung by the
arm in a particular manner such that an animal could be struck and
killed, or an enemy frightened off, a piece of wood split and formed,
etc. It would not be untrue to say that the stone “instructs” the hand
how to hold and wield it. Other kinds of stones with different char-
acteristics don’t do this. Other ways of holding the stone in the hand,
swinging the arm, and striking something don’t have the effects that
this particular stone and particular movements with respect to these
particular objects have. It could be said that this special stone, but not
only the stone, also the particular anatomy of the hand and arm of the
hominin, as well as the characteristics of the animals, enemies, wood,
etc., had certain “affordances” that somehow could be associated with
each other such that over thousands of years of acting upon one another
created something that did not exist before, namely, a “hunter,” or a
“warrior” who wielded a stone “ax.” Although Latour does not use the
term “affordance,” it is helpful when attempting to describe the agency
of nonhuman actors.168 Gibson (1979:129) writes:
[A]n affordance is neither an objective property nor a subjective
property; or it is both if you like. An affordance cuts across the
dichotomy of subjective objective and helps us to understand
its inadequacy. It is equally a fact of the environment and a fact
of behavior. It is both physical and psychical, yet neither. An
affordance points both ways, to the environment and to the
observer.”
168 Latour does not use the term aordances for the contribution of humans or
nonhumans to an actor-network. Instead, he speaks of “programs of action” or
“propositions” (Latour 1999, 309). A proposition is what “an actor oers to other
actors” (309). Also, it should be noted that an aordance is not merely a “con-
straint,” but an active contribution to the construction of what is going on, of the
“meaning” of the event.
Network Science 207
Not only did this unique and exceptional cooperation between certain
stones, hominins, and other things that were involved change the being,
the identity, and the behavior of the hominins, who became hunters
or warriors, it also changed the behavior and identity of the stones as
well, which were no longer mere stones, laying about on the ground,
but “axes.” What is seen is best described as a network of actors, all of
whom are transformed by being associated in the way they were and
thus becoming a network. The stone became a “tool,” an artifact, and
the hominin became a “tool-maker.” The ax is not a mere stone, and a
hunter wielding an ax is not a mere hominin. The ax and the hunter
mutually enable and condition each other. There would be no hunter
without the ax and no ax without the hunter. Together, they, and of
course the animals and enemies that experienced what the ax does
to them, build a network in which they subsequently exist “as” what
they are. To say, as Heidegger does, everything that appears, appears
“as” this or that is to say that hominin, stone, etc., appear as actors in a
network of certain associations.
An actor-network is a mutual conditioning of human and nonhuman
that makes all “actors” into something that none of them was before. It
could, therefore, be said that the associations, the relations, determine
what and who the actors are. This is a relational ontology because the
actors do not exist prior to the associations out of which they emerge
transformed. This mutual conditioning, that is, the construction of asso-
ciations and relations, is where we propose to locate the construction of
meaning. Meaning does not emerge in psychic systems of perception and
cognition or social systems of communication, as Luhmann proposes.
Our hypothesis is that meaning emerges as associations constructing
actor-networks in which human and nonhuman actors become what they
are. Actor-networks are neither subjective nor objective, neither natural
and social, neither material nor ideal. They undercut all the traditional
distinctions of Western ontology and systems theory. The process by
which actor-networks and thus meaning are constructed Latour calls
“technical mediation.” What happens in technical mediation?
Technical mediation (Latour 1994) is a process of translating and
From Systems to Actor-Networks
208
enrolling both human and nonhuman actors into actor-networks, that is,
cooperative programs of action. Latour is well aware that the concept of
the “technical” is ambiguous; therefore, he is careful to point out that he
is using the word in a special sense.
Technical … designates a very specific type of delegation, of
movement, of shifting, that crosses over with entities that have
different timing, different properties, different ontologies, and
that are made to share the same destiny, thus creating a new
actant. (Latour 1994:44).169
The word “technical” refers to the fact that human action is always
associated and interwoven with nonhumans, material and biological
entities, and their affordances. It implies, against Aristotle’s famous and
influential distinction between techne (constructing artifacts) and episteme
(knowing first principles), that knowing is doing, making, constructing,
and constructing is always the construction of meaning. The taken-
for-granted distinction between subjective human beings acting upon
objective, passive things in the world does not adequately describe what
is happening. The phenomenology of the stone ax illustrates this very
well. Our hominin did not sit down and think about how to make an
ax. Such “things” or “ideas” did not exist 3,3 million years before Plato.
We can assume there was neither language nor thinking in the form we
know today. Nonetheless, and this is the crucial point to this entire story,
there was meaning. The meaning of the ax was nothing other than the
ax itself and not some mental state, a cognitive act, a sign, or an idea. It
would take millions of years and the construction of a myriad of complex
actor-networks before Plato could come up with the notion that ideas
were located in some immaterial realm distinct from things and used
as blueprints for their construction, whether by divine or human hand.
This traditional understanding of “construction” obscures the original
experience of technical mediation. Latour (2013:157) writes:
169 Further: “Technical skill is not uniquely possessed by humans and reluctantly
granted to nonhumans.
Skills emerge in the zone of transaction, they are properties of the assembly that cir-
culate or are redistributed among human and nonhuman technicians, enabling
and authorizing them to act.” (Latour 1994, 45)
Network Science 209
To say that something—a scientific fact, a house, a play, an idol,
a group—is “constructed” is to say at least three different things
that we must manage to get across simultaneously.
First, the activity of constructing is strangely “doubled.” Since there is
always more than one actor involved in a network, it is unclear from
where the activity comes, from only one, from others, and who was
first. Was it the stone that first suggested its use as an ax to the hominin?
Or was it the hominin who suddenly realized that this particular stone
could have a use? Or was it the enemy who forced one to do something
to defend oneself, the wood that needed to be split, or the animal that
had to be killed and cut up? Agency cannot be isolated and ascribed to
only one of the actors. All actors are involved “symmetrically” in the act
of construction.
Every use of the word “construction” thus opens up an enigma
as to the author of the construction: when someone acts, others
get moving, pass into action. (Latour 2013:158)
Second, the act of constructing does not know in advance what is going
to come out at the end. As Latour puts it, the “vector” of the action is
uncertain. There is an “oscillation” between various outcomes or direc-
tions.
the arrow can go in either direction: from the constructor to the
constructed or vice versa, from the product to the producer,
from the creation to the creator. (158)
Latour cites the example of a puppeteer who is guided by the puppet’s
personality to pull the strings in specific ways. The hominid did not
know it would become a hunter, a warrior, or a builder the moment
the stone ax was constructed. Presumably, the stone did not plan this
outcome either, and the potential enemy was undoubtedly surprised
when being attacked by a warrior wielding an ax. With regard to the
construction of scientific facts, it is decisive that the experiment is so
constructed that it is not merely the ideas, intentions, and desires of
the scientist that appear as results, but instead, what the entities under
From Systems to Actor-Networks
210
investigation contribute to the outcome. Technical mediation constructs
by means of translation and enrollment of the actors in the roles they
play within the network, and none of the actors knows what they will
become before the act of construction. The hominin did not set out that
day 3 million years ago to become a hunter. The stone had no intention
of becoming a status symbol or a cult object that mediated not dead
animals or split wood but contact with transcendent forces or helped
some hominids find a partner. Construction, as with all networking, is
open-ended and contingent.
Finally, the third component of the meaning of construction is what
Latour refers to as a “value judgment.” Whenever something is
constructed, it can be judged as either well-constructed or poorly
constructed.
To say of a thing that it is constructed is to introduce a value
judgment, not only on the origin of the action…but on the
quality of the construction… it is not enough for the experi-
menter to construct facts through artifices; the facts still have
to make him a good experimenter, well situated, at the right
moment, and so on. Constructed, yes, of course, but is it well
constructed? (159)
The hominin and their children probably tried out countless stones,
ways of wielding them, and uses that they might be put to, always
judging whether it was well done or ill-done. We will argue below
that this meaning of technical mediation as construction directly links
action to design. The hominid was a designer because the construction
of meaning is a question of good or bad design, just as Heidegger’s
carpenter reaches for his hammer to make a good chair or table, and
others judge the result of his work accordingly. Ethics, therefore, is built
into reality and not added to it as an “ought” that is opposed to what is.170
ANT reaches back into a time before the idea of ideas as somehow
distinct from things arose, before the distinction between matter and
170 For a discussion of the ethical implications of ANT see Belliger/Krieger (2021).
Network Science 211
mind had become a foundation of Western thought, and before the
idea that what is constructed is somehow “artificial” and not “real.”
For this reason, ANT is often associated with what is called “material
culture” or “material semiotics.”171 When discussing systems theory
above, we named the principle of order of meaning systems “semiotic
coding.” From ANT’s point of view, semiotic coding is not primarily
about syntax, semantics, or signs somehow referring to things but
about technical mediation. What technical mediation does is transform
the world into meaning, or in other words, it is the process by which
meaning emerges as a level of order on its own. Just as autopoiesis is the
process by which life emerges from matter, so is technical mediation,
translation, and enrollment, the process by which meaning emerges.
Before the actor-network of the stone ax was constructed by technical
mediation, there was only the potential of meaning. Just as there was a
point in time when there were only chemical reactions on Earth and life
had not emerged, the potential was there. When that hominin picked
up that stone for the first time, meaning had not yet emerged. What our
ethnologist witnessed over 3 million years ago was the emergence of
meaning. Meaning emerged as both material and semiotic, both subjec-
tive – the experience of the hominin – and objective – the transformation
of the stone into an ax. If we consider all the things that are in any way
involved in actor-networks as “technology,” then this is why Latour
can say that technology is not merely a tool, a passive instrument, but
a partner, a social actor. Indeed, technology is society, and society is
always technologically mediated.172
As a theory of meaning, the concept of technical mediation does not merely
claim that human action is mediated in the sense of being influenced by
tools, artifacts, or material entities of some kind, or even visual or acoustic
signs, as in language. Technical mediation does not mean that people use
tools to do things they might otherwise not be able to do. It does not mean
171 See Wikipedia hps://en.wikipedia.org/wiki/Material_culture; on material semi-
otics see Law (2019).
172 In his later work Latour (2013) understands technology as a particular “mode of
existence” among others such as politics, religion, science, etc. We use the term
technology in the sense of technical mediation which, we maintain, applies to all
modes of existence as networks.
From Systems to Actor-Networks
212
that tools and technologies extend or augment natural abilities. This is
undoubtedly true. The ax did influence the hominin, just as industrial
production influences modern society, human self-understanding, and
everything else humans do and are. But technological mediation means
more than mere influence, augmentation, or extension of what humans
can do based on the natural bodies. In our example of the stone ax, tech-
nology constructed the very being, the identity, and the existence of the
hominin, who eventually – most probably because of technical mediation
– became a human being. Once technical mediation emerged, that is, once
meaning emerged, the large brain and linguistic abilities that later evolved
had a purpose, something that gave certain hominins an evolutionary
advantage.173 The agency of nonhumans is expressed in the concept of
“affordance.” What affordances and technical mediation suggest must not
be misunderstood as technological determinism.
On the contrary, if mediation is not a mere influence upon an actor but the
ontological constitution of the actor, then it is the network of associations
among humans and nonhumans that constructs the actor. This implies
that the actor is not only the product instead of the producer but also
that the producer is technical mediation or meaning. Just as Luhmann
can say that communication communicates – and it is communication
that produces communicators –ANT says that meaning creates meaning
and actors are entities that become meaningful through technical medi-
ation. But here again, one must resist the temptation to misunderstand
the term technical and fall into technological determinism. On the level
of networking, there is nothing outside of technology to be determined
by it. Society is technology. As Latour never tires of saying, the network
is the actor. The hominin without the ax is not a hunter but a mere ape,
and without the hunter, the stone ax is merely a stone.
The central point of technical mediation is that the idea of mediation
refers to the associations among the actors that constitute the actors and
not to any influencing done by a subject towards an object or vice versa.
The associations first make the actors into what they are and give them
173 For the inuence of toolmaking on the evolution of cognitive and linguistic abil-
ities see Orban/Caruana (2014).
Network Science 213
the specific roles they play within a network. As counterintuitive as it
may be, it is the relations, the associations, that are primary, whereas the
identities and roles of the actors emerge from the relations. The theory of
meaning we propose claims, therefore, that being is relational, and onto-
logically, relations are fundamental, not the relata. ANT is not a substance
ontology but a relational ontology.174 To a certain extent, systems theory
can also be considered a relational ontology since systems come into
being by virtue of processes of selection and relationing and thereby
construct their own elements. Indeed, self-organization is defined as the
emergence of order out of interactions between elements. Systems, just
as networks, are not substances. But even if we assume that for systems
theory, relations are real and not merely added onto reality, systems,
as opposed to actor-networks, do not primarily consist of relations. For
systems theory, the relationing of elements depends upon their selection
and the specific purpose or goal of system operations, and all of this
depends on the system/environment difference. This is not the case for
networks with no constitutive border or purpose.
Networks, as opposed to systems, are nothing but relations, even if
these relations are differentiated into what Latour (2013) calls “modes
of existence.” It is difficult to claim that what our ethnologist saw when
observing the hominin becoming involved with the stone ax was pure
relations and nothing else. After all, the stone was there in the river bed.
It had been there for perhaps millions of years. How can mere relations
change all this? How can one see relations where before that had only
been stones? One can see stones and hominins, but where are their
relations? We do not stumble over relations as we might stumble over
stones. Relations are not things. What are they? We propose the term
“information” to designate the associations from which actor-networks
arise. For our purposes, which are primarily directed toward a theory of
meaning, relations are information. What is “seen” when we see relations
174 Latour (2013:162) expresses this idea by distinguishing between “Being-as-Being”
and “Being-as-Other.” As opposed to being-as-being which rests upon a foun-
dation of substance or sameness, being-as-other must establish “subsistence” or
continuity. It “does have to pass through a leap, a hiatus, to obtain its continuity,”
and he goes on to cite Tarde (163), “dierence proceeds by diering.”
From Systems to Actor-Networks
214
is information. Relations appear as information.175
Just as systemic order requires that the elements out of which a system
is constructed are related in a non-random order, network order also
requires that information be ordered in specific ways.176 Technical media-
tion constructs the associations or relations among the actors, which allow
the actors to appear as actors. What does it mean for an actor to appear
“as” an actor? It implies that actors play roles in a narratively constructed
“story.” Information, therefore, is primarily ordered as a narrative.177
Actors do things in a temporally unfolding trajectory of events, which
assign roles to actors and order events in a certain way. One thing happens
after another. One actor does something, and other events follow from
this. What describes these roles and coordinates them into a narrative can
be called a “program of action.”178 A program of action describes what
may be called the “trajectory” of the network and applies to humans
and nonhumans symmetrically. As Latour (1992:233) puts it, “Parts of
a program of action may be delegated to a human, or to a nonhuman.”
Agency is always distributed and always a narrative construction.
When systems are said to pursue goals or act to maintain set points, this
is a narrative describing who does what, when, and why. Although for
both systems and actor-networks, agency can be ascribed to individual
actors – for example, Luhmann is compelled to ascribe communication
to persons – it must not be forgotten that it is the social system or the
network that is the actor. Communication communicates. So also for
ANT, the network networks. Agency for ANT, as for systems theory,
175 This denition of information goes back to Bateson’s denition of information as
a dierence that makes a dierence, but goes beyond Bateson by claiming that
things are relations and nothing else, or in other words, information is reality on
the level of meaning, and after meaning has emerged there is no other reality. Or,
as Heidegger puts it, whatever appears appears “as” this or that particular thing.
The hermeneutical “as” signies the emergence of information.
176 Latour (2013) refers to these various ways of ordering information as “modes of
existence.”
177 For a discussion of narrative in establishing social order see Belliger/Krieger
(2016).
178 See Callon (1991:136). Programs of action, also called “trajectories,” are usually
formulated as narratives. See Belliger/Krieger (2016) for a discussion of the nar-
rative construction of social order.
Network Science 215
is never individual. For systems theory, it is the steering function of
systemic order that accounts for goal-directed behavior. For ANT, “Action
is a property of associated entities” (Latour 1994:35). In other words,
networks network. Networking is an ongoing process that constructs
information ordered in a narrative. For ANT, technical mediation is a
form of practice, a problem-solving activity, but one must never forget
that the actors doing this practice are not exclusively humans.
To explain how processes of translation and enrollment construct infor-
mation in the unique sense of the term we are proposing, let us briefly
return to our hominin and the stone ax. What is unique and special about
this particular hominin and this particular stone? To formulate the ques-
tion somewhat differently: What distinguishes technology from animal
uses of “tools?” When animals use tools, they pick up a rock or stick
to break open a coconut or forage for food, then drop it and move on.
Animal tool use is characterized as “episodic.”179 The stone ax certainly
was also left behind occasionally, but something else was “holding” on
to both our hominin and a particular stone that made a difference. Para-
doxically, it can be said that the moment a hunter wields a stone ax, quite
literally, “things have gotten out of hand.”180 In other words, the stone or
the stick is no longer a tool for the moment, only for as long as it is held in
hand. Even when it is not being used, even when the hand is doing other
things, the tool somehow stays with the tool user, or as Heidegger might
say, it exists as “ready to hand” even when not in the hand or not even
in sight. The tool does not disappear as a tool and become a mere stone
again the moment it is no longer being used, as can be assumed to be
the case for animal use of tools. Technical mediation changes everything.
When the hominin is not holding the stone, the “ax” and what it means
to be a “hunter” is holding on to the hominin and the stone. Neither
the hominin nor the stone return to what they were before becoming
179 Simian tool use is characterized by its “episodic” nature. See Donald (1991, 149),
who describes the “culture” of apes as “episodic.” “In fact, the word that seems
best to epitomize the cognitive culture of apes […] is the term episodic. Their
lives are lived entirely in the present, as a series of concrete episodes and the
highest element in their system of memory representation seems to be at the
level of event representation.”
180 See Belliger/Krieger (2016:29.).
From Systems to Actor-Networks
216
involved with each other in the actor-network that is the stone ax. Once
it becomes a “hunter” or a “warrior” it remains what it has become even
when not holding the ax and not hunting or fighting off enemies. On the
contrary, the animal drops the stone when its purpose has been achieved.
Nothing has changed. Neither the animal nor the rock has become some-
thing other than what they were before. Stone and animal do not become
an actor-network. They are not transformed into something they were
not already. The actor-network fundamentally differs from any single
episodic use of a stone or a stick as animals do this. With the emergence
of the actor-network, the stone has become associated in meaning with
the hunter or warrior as the hominin is with it. Technical mediation
constructs the relations or associations between hominin and stone such
that something emerges that holds on to all actors and integrates them
into a network even when the stone is dropped, and the hominin is no
longer using it. To say that “things get out of hand” means that they
remain and are “held” in the network and do not simply revert to their
former mode of being. They take on their own life and remain what they
have been transformed into. That which is holding onto the hominin and
the stone beyond any episode of actual use is information.181
We have already mentioned Heidegger regarding the primary or originary
mode of existence of tools or things in general. For Heidegger, humans’
primary or original relation to things is that of practical use or practical
engagement with the world. The terms “primary” and “original” refer to
how Dasein relates to the world before reflection abstracts from imme-
diate experience, that is before one begins to think of the world in terms
of appearance and reality, substance and accidents, subjects and objects,
etc. Everything appears “as” something specific, as this or that particular
thing bound up in some praxis. The originary way things appear is not as
passive objects of disinterested knowing. Dasein is not primarily a disin-
terested observer of objects but an engaged and concerned user of tools.182
181 As Latour (2013:223) puts it, “’technology’ does not designate an object but rath-
er a dierence.”
182 What Heidegger calls “care” (Sorge) which characterizes Dasein’s existence as
practical Being-in-the-World seems to reappear in Latour’s return to “experi-
ence,” but also in his notion of “maers of concern.”
Network Science 217
Only when the tools break or do not work as expected does Dasein step
back from engagement and ponder the situation. But here again, practical
engagement with the world, the question of fixing the broken tool, guides
attention. The attitude of scientific or objective observation and the tradi-
tional assumptions about substance and properties, subject and object,
knowledge and reality, nature and society, arose very late in the history of
meaning. Things first appear not as passive objects with measurable prop-
erties or as substances with so-called “primary” and “secondary” qualities
but as tools involved in practices or, as Latour would say, programs of
action. The hammer, to cite once again Heidegger’s famous example from
Being and Time – an example that is not too far away from our stone ax – is
primarily given in its many practical uses, both what it may currently have,
for instance, building a table, and also what it could have, and not as a
mere object of disinterested knowing. Recalling Luhmann’s definition of
meaning from Husserl, it is these many uses, practices, and ways of doing
things that make up the horizon of references that constitutes meaning.
Luhmann interprets these references to refer to what they are used for. The
hammer, for example, “refers” to wood, nails, the workshop, etc., but for
Luhmann, this happens only as either perception or communication. For
ANT, on the contrary, it really happens in the world of things. Meaning lies
in doing, participating in an actor-network, and building tables, not merely
perceiving or talking about them. It seems that ANT stays much closer to
the practical, primary, and original nature of meaning than Luhmann.
Technical mediation is not “technical” in the usual sense of the word,
which describes the workings of a machine or its engineering. Latour’s
notion of technical mediation is a metaphysical concept that describes
how meaning as such is constructed. As Heidegger put it, the essence
of technology is not itself technical.183 When ANT speaks of technical
mediation, what is being spoken about is not any artifact or machine. It
is also not exclusively a matter of tools unless one wants to consider all
things as “tools” in some sense. It is what Luhmann would call the oper-
183 It should be noted that Latour’s technological mediation is much dierent than
Heidegger’s Gestell, which is a systems concept and not a network concept.
Heidegger was describing modern Western society much as Luhmann later de-
scribed it under the dictates of universal functionalism, while Latour is already
beyond modernity and concerned with a post-modern world.
From Systems to Actor-Networks
218
ation of a meaning system. For Luhmann, a meaning system operates
to construct meaning via perception and communication and nothing
else. This is the autopoiesis of meaning. Unlike Luhmann, in ANT, we
no longer talk about “observation” or the functionally directed internal
information construction of a closed system. As stated above, network
order results from what may be called “networking.”184
Networking is what happens when things seem to “get out of hand.”
They are taken up into a network of associations that transforms all
the actors involved by allowing them to appear on the level of emer-
gent order of meaning. Meaning, therefore, is not a cognitive act that
somehow adds semantic significance, signs, or mental representations
to things. Meaning does not reside in signs that somehow stand for
things. Signs, significance, language, and much more come much later
in history than the emergence of meaning, which we are describing
with the example of the stone ax. There is nothing that “has” meaning
as opposed to things that do not have meaning. Once meaning has
emerged, matter and life; indeed everything that exists, appears “as”
something meaningful. As soon as one stone became an ax, all the
other stones became non-axes or potential axes.185 Meaning is not some-
thing other than the things themselves. Meaning is not something that
things “have” or do not have, something with which a cognitive agent
somehow endows them. Meaning, or as we prefer to say, information,
is what things “are.”186 And it is only because Being is meaning that
cognitive agents can arise and that there is an evolutionary advantage
to having big brains and special linguistic abilities. As Latour (2013:230)
puts it: “technologies have preceded and generated humans” and not
the other way around.
With these remarks, we hope to have set the stage for a more detailed
discussion of how technical mediation constructs meaning. Technical
184 See Belliger/Krieger (2016) for a discussion of networking.
185 As Wigenstein noted, it is impossible to say just one word alone. To speak one
word is to speak a language.
186 This is the idea behind what has come to be known as material semiotics, or ma-
terial culture.
Network Science 219
mediation consists of processes of “translation” and “enrollment,” just
as systems consist of selection, relationing, and steering processes. For
any tool to be used as a tool in a specific activity, a series of relations
has to be established between the actor using the tool and the tool itself,
as well as between the tool and what it is supposed to work upon. The
stone ax, for example, works, let us say, upon an animal or a piece of
wood. These elements, the hand, the stone, and the animal or wood,
have to be related to each other in specific ways. Latour speaks of a
“detour” that must be taken by someone who wants to work on some-
thing using a tool. In the case of the stone ax, the hominin wished to kill
an animal or split a piece of wood. Since they couldn’t do this very well
with their bare hands, they had to take a detour via the stone ax. Latour
calls this detour “translation.”
Translation does not mean a shift from one vocabulary to another,
from one French word to one English word, for instance, as if
the two languages existed independently. Like Michel Serres,
I use translation to mean displacement, drift, invention, medi-
ation, the creation of a link that did not exist before and that to
some degree modifies two elements or agents. (Latour 1994:32)
Translation is not simply the establishment of a relation between two
previously unrelated things. It is a relation that transforms what it
relates. It is not merely a matter of placing two existing things into a
relationship with each other, whereby the relationship is something
added to the things and may, therefore, also be subtracted. A relation
that is added onto already existing relata does not change the things it
relates. On the contrary, translation is a unique way of creating associ-
ations that transform what is being associated. It transforms them into
something new and unexpected. The hominin, who before had only
their bare hands to work with, is now a being that can kill animals or
split wood by wielding an ax. The stone is no longer something lying in
a riverbed but has become an ax, a weapon or tool held in the hand of a
“hunter” or a “warrior.” It is important not to forget that the animal that
is killed or the wood that is split also contribute to this activity because
only some animals can become game and only some types of wood
From Systems to Actor-Networks
220
can be cut with a stone ax, and this only in specific ways which they
decide as much as the ax or the wielder of the ax. When the animal and
the wood don’t “cooperate,” nothing happens. What is essential in this
account is that agency is distributed. All actors are equally, or “symmet-
rically,” involved in constructing the actor-network that is hunting or
chopping. It is not the hominin alone who is the hunter hunting or
the builder chopping wood. It is the actor-network constituted by the
unique associations between hominin, stone, and animal or wood that
the translation process has constructed.187
Translation brings enrollment with it. Symmetrical agency means that all
actors in a network have influenced each other such that they become
enrolled into an actor-network that has a particular trajectory or program
of action.188 When the hominin sets out to defend the camp against
enemies, this is a different trajectory and program of action than setting
out to cut wood or hunt animals. Everything is different: the hominin,
the ax, the way of wielding the ax, the addresses of the action, etc. It
is translation that makes the associations among actors and enrollment
that determines the purpose, trajectory, and program of action the
network follows. According to the network model, meaning is not what
Husserl described as the conscious thematization of an intentional object
against the horizon of possible references. Meaning emerges through
the translation and enrollment of actors into programs of action that are
really carried out and not just thought about. It is only on the level of
the emergent order of meaning that actors, both human and nonhuman,
are constructed with specific roles to play. Heidegger showed how the
intentionality that Husserl described is actually Being-in-the-World,
a practical engagement with things guided by “caring” (Sorge) or
“concern” (as Latour will say). The hammer makes the carpenter in
Heidegger’s workshop as much as the carpenter makes the hammer.
From the ANT perspective, the emergence of meaning constructs the
actors: the hunter and the ax, the carpenter and the hammer. It is the
187 The reader is reminded of the threefold meaning of “construction” (Latour 2013:151.)
including distributed agency, uncertainty of outcome, and value judgement.
188 Latour will use this idea as the basis for distinguishing “modes of existence,” or
dierent kinds of networks.
Network Science 221
actor-network constructed by translation and enrollment that creates
what it means to be a “hunter,” a “warrior,” or a “builder,” and what an
“ax” is. The hunter has different goals than the warrior or the builder.
They hold different things in their hands and use them differently. The
network and the possibilities that it creates did not exist before the work
of translation and enrollment, which enabled a program of action to
come into being. When an ape picks up a stone and uses it to crack open
a coconut, he drops it again, eats the coconut, and moves on. He has not
become a hunter or anything other than he was before. The stone used
and dropped has not changed and become something it was not already.
Nothing is there to hold on to the ape or the rock after the episodic use,
or at least not enough to form an actor-network.189 When an actor-net-
work appears, however, there is something new. Something remains.
Something is there to hold on to all actors in the network even when they
move on to do other things. Something has come into being that was not
there before. What is this something that has come into being that did
not exist before? What is holding on to the actors and transforming them
into something new that does not disappear when the activity ceases?
We propose calling this special and unique something that exists only on
the level of emergent order of meaning “information.”
3.4 Information
Unlike systems theory, at least in Luhmann’s proposal, network science
has not explicitly developed a theory of meaning. Network science
takes the nature of the nodes and links that are the object of study for
granted and is concerned mainly with how many and in which kind of
order the nodes are linked. The meaning of what is in the network is
never at issue, or at least, it is not explicitly problematized. Regularities,
distribution, topography, etc., constitute network science’s object. For
ANT, the point of departure for network description is much different.
189 The use of tools by animals is well-documented and exhibits extraordinary sim-
ilarities to human tool-use. This seems to imply that we are not dealing with a
radical discontinuity, but much rather with a case of emergence. In the same way
life emerges from maer, there is always continuity despite jumps.
From Systems to Actor-Networks
222
The associations among the actors are also the object of study, but the
associations ANT talks about are an exceptional kind of relation. They
are relations that do not merely link actors but constitute actors. For
network science, the nodes exist whether they are connected or not.
The nodes are first things that then secondarily become related through
links in some way or another. What the nodes are, whether computers,
people, proteins, emails, etc., is already given, and the links among them
are being studied. After all, everyone knows what a person is, what an
email is, what constitutes an infection in epidemiology, what users are
on a social media platform, etc. Network science is not concerned with
metaphysical or epistemological questions about how all these things
considered nodes in a network came to be what they are. For systems
theory, the situation is different. In systems theory, Luhmann must
address this issue because he departs from the usual assumptions of
systems theory by proposing meaning as a level of emergent order in
its own right. Luhmann, and he is to be credited for this, introduces into
systems theory a theory of meaning. As we saw, he relies on Husserl
and the notion of the thematic intention of an object against a horizon
of possible references. For Husserl and Luhmann, this is what meaning
is. However, Luhmann goes on to distinguish between how psychic
systems process these references as perception and social systems as
communication. He proposes to model meaning as an autopoietic,
operationally and informationally closed system.
In contrast to network science, but in agreement with systems theory,
ANT is also concerned with the constitution of meaning; perhaps
“construction” would be a better word. ANT proposes that associations
construct actor-networks in which things first appear “as” (recalling
Heidegger’s hermeneutical “as”) what they are. It is the associations
among the actors that “translate” and “enroll” the actors in the roles
they play in the network. It could seem that this idea has similarities
to systems theory. One could argue that in systems theory, the system
constructs its own elements. Elements do not exist before being inte-
grated into a system. No tabletops or table-legs are lying around in
the environment waiting to be taken up into a table system. The table
system constructs tops and legs out of materials of many different kinds
Network Science 223
to make a table. No lungs, hearts, or skin lie around in the environment,
waiting for an organism to integrate into itself. The organism constructs
its own elements that did not exist before being constructed by the
living system. Also, in ANT, an actor-network constructs the actors it
is made of through translation and enrollment. A hominin becomes a
“hunter” or “warrior,” and a stone becomes an “ax.” But as we will see
later, an actor-network is not a system. It is not operationally and infor-
mationally closed. It has no environment. It is not functionally defined.
It does not “adapt” to anything. We will discuss the differences between
systems and networks in detail below. For the moment, it is important
to ask: If the associations between hominin and stone construct the
actor-network, what do we see when we describe an actor-network?
ANT describes the actors and what they do. But what makes actors
into actors with a specific program of action are the associations
without which there would be no hunter, warrior, or ax. When the
stone becomes an ax, it is transformed into something entirely different
from a stone. It is taken up into a network of associations determining
what it is. However, associations and relations do not seem to have the
same nature as things we can pick up or see and describe. We cannot
hold a relation in the hand and use it as an ax. Or can we? What do we
see when we see an actor-network? What do we do when we are an
actor-network? To answer these questions, we propose introducing the
term “information.” Information can and does mean many things. We
wish to propose a new understanding of what information is well aware
of all the risks involved in attempting to change the meaning of a word
that is already overdetermined in many ways. Nonetheless, we propose
to define information as what we “see” and what we “are” when we
appear in an actor-network, when an actor-network comes into being.
When Husserl says that meaning is the intentional thematizing of
something against a horizon of further references, and when Heidegger
pointed out that whatever appears or comes into being appears “as”
this or that particular entity embedded in a world of associations, ANT
says that networking, processes of translating and enrolling construct
From Systems to Actor-Networks
224
actor-networks. In short, networking constructs information.190 Actors,
both human and nonhuman, indeed, all beings are information. If we
take the information away, we don’t see anything and can’t do anything
either; the “world” disappears, and we rejoin our simian relatives. Let
us not forget that our ethnologist, who went back in time to describe
the first actor-network is also an actor in a much more complex and
extended network of modern science and ethnology that arose much
later in history due to the hard work of constructing networks over
centuries. Nonetheless, being an ethnologist is no different from being a
hunter or a warrior, for which there are still many examples. We propose
to think of all of this, from the first hominin who used the first stone ax
to ethnologists, physicists, and biologists of today and everything they
are concerned with as information. But what is information?
We have argued that processes of translation and enrollment laid the
foundation for the evolutionary advantages of big brains and language.
These processes arose long before Homo sapiens with big brains and
linguistic abilities appeared on the evolutionary scene. Meaning emerges
with the emergence of actor-networks and not with the appearance of
Homo sapiens. In contrast to Luhmann’s theory, we assert that there
were actor-networks and meaning long before the psychic or social
systems Luhmann describes. In contrast to Husserl and all of modern
philosophy and cognitive science, meaning does not depend on the
intentional consciousness of empirical or transcendental egos. Systems
theory assumes self-organization based on selection, relationing, and
steering processes. As we have argued, this model does quite well
for physical and biological systems but cannot adequately describe
meaning since systems require a constitutive difference from an envi-
ronment. The moment this difference becomes meaning, it falls within
the system and can no longer function as a boundary of the system.
For ANT, however, processes of translating and enrolling distribute
agency among all actors, both (proto)human and nonhuman. A network
190 Luhmann would of course agree that a meaning system constructs meaning, but
Luhmann and systems theory do not use the term information in this sense. Al-
though he introduces the idea of a meaning system it is perception and commu-
nication that are constructed by the system. Information plays a secondary role.
Network Science 225
does not need to distinguish itself from anything; it has no constitutive
boundaries and can scale up indefinitely. This corresponds more closely
to Husserl’s original idea of an infinite horizon of possible references
and Heidegger’s notion of Being-in-the-World. Translation and enroll-
ment, therefore, are not uniquely human activities; they are not mental
processes, they are not forms of subjectivity; they do not take place in the
brain, whether it be a big brain or a small one; they are neither psychic
nor social nor natural since they are basis for all these distinctions. They
are processes of meaning construction in which Homo sapiens came
to participate in a unique way millions of years after they had already
emerged. The implication is that meaning can no longer exclusively be
attributed to human individuals or social systems. Meaning, therefore,
is not reducible to what we humans may know or communicate but is
the condition for the possibility of being human and knowing anything.
Because the world is meaningful, we can understand it and technologi-
cally transform it, and after millions of years of evolution, even come to
talk about it, and not the other way around.191
Because the world is meaningful, having big brains and a command of
language can prove to be an evolutionary advantage. But once meaning
has emerged, evolution no longer plays a role in how meaning develops.
We humans do not give meaning to the world. On the contrary, with the
emergence of meaning, a “world” appears for the first time, a world in
which humans, for the time being at last, have gained a special place.
In words that remind one of Heidegger, we could say that it is not we
humans who “use” language, but language which “uses” us. Of course,
big brains are important. Big brains and linguistic abilities enable
networks to be scaled up almost indefinitely, allowing quick movement
through uncountable links. Big brains and language make it possible to
191 This marks the departure of Latour’s theory from modern subjectivism in the
line from Descartes and Kant to Husserl to which Luhmann despite disclaimers
is commied, as well as the solution to the problem of “correlationalism” posed
by Meillassoux (2008) which he denes as “the idea according to which we only
ever have access to the correlation between thinking and being, and never to
either term considered apart from the other.” This is not a problem unless think-
ing is the exclusive property of Homo sapiens or being is considered an object
known by a subject.
From Systems to Actor-Networks
226
integrate many more links and relations into a network than possible with
more limited cognitive abilities or none. Nevertheless, human cognition
alone does not account for meaning, nor is it the source of meaning.
As the phenomenology of the stone ax shows, meaning arose long before
Homo sapiens. It is, therefore, not observation (Luhmann) but mediation
(Latour) that constructs meaning. It is not the logical operator of nega-
tion, which Luhmann supposes is the foundation of meaning – following
Spencer Brown, “draw a distinction” – but the construction of relations
and associations, in short, networking, that makes meaning. For Luhmann,
it is systems that construct information; for ANT, it is information that
constructs networks. Again, we could say with Heidegger that after the
emergence of meaning, it is not dasein alone but every being that exists as
an issue for itself. This is not because all beings are somehow endowed
with intentional consciousness or the ability to make observations and
draw distinctions but because all beings are endowed with agency in the
sense of being able to enter into processes of translation and enrollment. All
beings, as beings, are potential mediators capable of becoming involved in
constructing relations. This is what the emergence of meaning means.
When we speak of information in this context, let us state clearly that we
are not speaking of some mental state. Information, as we wish to define
it, is not a cognitive act. Information is not, as Luhmann proposes, obser-
vation, that is, binary distinctions founded upon negation. Furthermore,
according to Luhmann’s theory of communication, information is not that
which is talked about in communication, one of three selections along with
utterance and understanding. Information in the view we are proposing
is a form of being sui generis. It consists of relations and nothing else. On
this level of being, that is, on the level of emergent order of meaning, there
are only relations, and what exists does so because it is constructed of asso-
ciations through the mediating activities of both human and nonhuman
actors, who themselves can only perform such activities because they are
taken up into processes of translating and enrolling.
Emergence means that phenomena appear that cannot be derived
from anything other than themselves. Information can only come from
Network Science 227
information, just as for Luhmann, communication can only come from
communication. Again, information, as we define it, is not a special
kind of thing. Information is not a substance with attributes. Information
is a process, the process of networking. We cannot hold information in our
hands, on paper, or in a database. We do not have information; we do
it, or rather, it does something with us, and it is through information
that we become what we are. Information is not a new name for a
particular kind of substance. Information must not be interpreted on
the basis of Western metaphysics or on the computational information
theory that dominates discussions of the topic today. In the view that
we are proposing, it is through the ongoing construction of information
by information that humans and nonhumans come to be what they are
and do what they do in a meaningful “world.” It cannot be emphasized
enough that information emerges not as mental states or linguistic acts,
consciousness or communication, or the transfer of bits and bytes but
as actor-networks. Information is at once material and mental, at once
human and nonhuman, at once signified and signifier, at once given
and constructed, at once subject and object, at once nature and society,
at once word and thing, at once map and territory. Indeed, all these
traditional distinctions are not helpful and even misleading when
defining information. This situation is not new.
Since the concept came into focus at the beginning of cybernetics, the
ontological status of information is notoriously uncertain. Wiener’s
(1949) famous declaration that information is information and not matter
or energy says more about what information is not than about what it is.
In addition, Shannon and Weaver’s mathematical theory of information
has come to dominate computational models and has uncritically been
taken over by media theory, communication theory, and other social
sciences. The result is that information has become a term that is not
very informative. It is, therefore, not surprising that Latour does not use
this concept and has not explicitly developed a theory of information.192
We wish to intervene in this unsatisfactory situation by suggesting an
192 Floridi (2011) is one of the few who explicitly aempts a systematic philosophy
of information. For a critique of Floridi’s position which overlooks the relational
nature of information see Belliger/Krieger (2021)
From Systems to Actor-Networks
228
interpretation of Latour’s principle of “irreduction” as a definition of
information and as the foundation of a theory of information. The prin-
ciple of irreduction was enunciated by Latour in an early philosophical
work of the same title. In this short essay, which was published together
with his study of Pasteur, Latour says that “nothing is, by itself, either
reducible or irreducible to anything else” (Latour 1993:158). What
does this mean? If something can only exist insofar as it is neither the
same, that is reducible to, something, nor totally different from (that is,
irreducible to) something, then it can only exist as a relation. In short,
nothing can exist “by itself.” This implies that Being is relational.
We spoke earlier of ANT’s relational ontology. According to a relational
ontology, identity, sameness, continuity, and persistence are not given
but constructed from relations. What do relations do? Relations mediate.
Mediation implies that whatever exists, exists because it is a mediating
mediator. Based on this principle, it becomes understandable why
technical mediation constructs information. It does so because it creates
associations, relations, and processes based upon constantly overcoming
reductions to identity.193 If we interpret the principle of irreduction as
a definition of information, what it states is that information emerges
as the mode of being sui generis, which integrates actors into networks
and transforms the actors, elevating them, as it were, onto a new level
of emergent order, the level of meaning. As with all levels of emergent
order, meaning integrates the levels below it: the physical and biolog-
ical. Once meaning has emerged and “world” appears, nothing exists
outside or beyond the world. Physical and biological science and all
they assert to know exist not outside of meaning but within the world
of meaning. They are parts of what Luhmann would call “society.”
Latour does not speak of society but of the “collective” to emphasize
that nonhumans are also involved in constructing meaning. From this
perspective, it is understandable why ANT arose out of Science and
Technology Studies and the sociology of knowledge. What happens
in the scientific laboratories is through and through social, that is, the
construction of actor-networks. The fundamental distinction between
193 See Latour (2013) for a detailed discussion of this aspect of networking.
Network Science 229
subject and object, society and nature, facts and opinions, the ideal and
material, and many other such dualities that Western modernity has
created are placed by this view into question. Indeed, Latour (1993) can
even claim that if one looks closely at what is really happening in the
world, one sees that “we have never been modern.”
Perhaps we must accept that hardly any traditional theories used to
define information adequately describe what information is on the level
of emergent order of meaning. Again, information is not something
we have and could perhaps do without; it is what we are.194 Following
Latour’s principle of irreduction, we can claim that what exists is
information and nothing else. Where Luhmann proclaims, “There are
systems,” Latour declares, “There are networks.” Both are universal
claims. Both talk about meaning and information, but they do so in
entirely different ways. Systems are not networks. It makes a difference
if information is produced by the autopoietic operations of an opera-
tionally and informationally closed system or whether the actor-net-
works in which we exist are constructed by information. For systems
theory, we, that is, the system, construct information, whereas for ANT,
it is information that constructs us.195
194 Floridi (2014) speaks of humans as “inforgs” instead of cyborgs. Inforgs are be-
ings that exist as information in a world that is also information, an “infosphere.”
Floridi’s concept of information, however, is dierent from what we are here
proposing. We are equating information with meaning. It is only very much lat-
er in the history of information when under certain circumstances it becomes
useful to distinguish information from meaning, for example, when studying
animal communication, or developing a mathematical theory of information.
195 In this context it is interesting to note that there is a tension in Levin’s (2022) the-
ory of universal agency and collective intelligence which is founded on systems
theory. For Levin, individual cells are at once autonomous, operationally and
informationally closed systems and still they necessarily obey the instructions
of morphological bioelectrical networks. If cells can only cooperate and form
collective entities such as organs or limbs by being instructed by a bioelectrical
network, then they cannot be autonomous. They do not do what they want, but
what they are told to do. They are not informationally closed but informationally
open. If all intelligence, as Levin claims, is collective, and collectivity demands
informational openness, then intelligence must be a network and not a system.
From Systems to Actor-Networks
230
3.5 Society: System or Network?
We have asked repeatedly throughout this book whether society could
not be better modeled as a network than as a system. Coupled with this
question was whether or not the universality claims of systems theory or
ANT were justified, that is, the claim of the theory to be able to explain
everything, including itself. Starting from Luhmann’s introduction of
a third level of emergent order, the level of meaning, and then quickly
focusing on “society” as the form in which meaning is ordered, we
noted that Luhmann was very clear in distinguishing his approach to
sociology from traditional social theory. Luhmann excludes everything
from the social system that is not communication or the way in which
meaning is expressed beyond individual psychic experience. From the
point of view of systems theory, at least four traditional assumptions
about the nature of society must be rejected:
1) that society consists of actual people and relations between people
2) that society is constituted or at least integrated by consensus
among human beings, by concordant opinion and complemen-
tary purpose
3) that societies are regional, territorially dened entities, so that
Brazil as a society diers from Thailand, and the United States
from Russia, as does Uruguay from Paraguay
4) that societies, like groups of people and like territories, can be
observed from outside. (Luhmann 2012:6)
Luhmann did away with the first assumption that society consists of
human beings and their relations by declaring human individuals to
belong to the social system’s environment, which is a system of commu-
nications and not of people. All communications are on the same onto-
logical level, the level of meaning. People, as such, have no meaning
other than what is said about them and attributed to them in commu-
nication. Although Luhmann distinguishes between society as a whole,
large functional subsystems, and small interactions among “persons,”
this does not carry the traditional distinction between individual and
Network Science 231
society, agency and structure, and micro and macro levels that charac-
terize traditional sociology. For Luhmann and Latour, social space is
“flat.”196 Systems theory does acknowledge social differentiation in the
form of systems within systems; indeed, this is what “internal differen-
tiation” in general systems theory means. Nonetheless, if subsystems
are indeed systems, then they relate to the systems they are “within”
as to an external environment. This means that persons and their face-
to-face interactions are only in the social system to the extent that inter-
action systems construct them, that is, only as communication. There
is consequently no problem in sociology of “integrating” individuals
into society. There is no fundamental opposition between agency and
structure and no micro or macro levels of society that are understood as
separate ontological domains.
Social systems theory eliminates the second assumption that society is
based on consensus because it does not assume that communication
depends on agreement. Indeed, the opposite is the case.197 The more
disagreement and controversy, the more communication, and the more
agreement, the less communication, for there is less to talk about. If
everyone were to agree upon what counts as true, socially acceptable,
and as honest and sincere, as Habermas (1987) claimed was necessary
for communication to be possible, then there would be no controversy,
no conflict, no diversity of interests, no real politics, indeed, no society
as we know it. All the real problems would already be solved. The fact
that the forces of social constraint and conformism are so great, a fact
that for Luhmann is illustrated by the emergence of symbolically gener-
alized media and functional subsystems, testifies to the constant danger
of communication failure and the threat that this poses to the ongoing
autopoiesis of the social system. Society is based not on consensus but
on the need to reduce communicative complexity. Systems are always
a complexity gradient, a balance of internal and external complexity.
196 This means that there are no ontological domains and no micro and macro levels
in society. For Latour’s version of this idea see (2005:165).
197 Habermas’s Theory of Communicative Action (1987) argued that communication
depends upon a consensus about fundamental criteria of truth, correctness, and
sincerity.
From Systems to Actor-Networks
232
Systems self-organize and come into being as attempts to reduce
complexity, which they necessarily presuppose, and not based on
sameness, conformity, unanimity, and consensus.
The third assumption that societies are territorially bounded entities
confuses society, at least in the modern period, with nation-states.
Nation-states, ethnic regions, cultures, etc., should not be equated with
society. At least in the modern period, where globalization removes
any territorial constraints on communication, it must be assumed that
nation-states exist within society, the one global society, and are not
society itself. If the social system is defined as the system of communi-
cations, where are the boundaries of communication? Does communi-
cation stop at the border of a nation-state? The theory of social systems
as a theory of communication is a theory of one global society, a “world
society”198 consisting of obviously not territorially bounded commu-
nication. All the functional subsystems of society are global. Business,
science, law, religion, art, education, etc., once differentiated into
functional subsystems with exclusive/inclusive binary coding, are not
territorially bounded. The only exception is the political system, which
has the function of making collectively binding decisions whereby
the “collective” is defined as a territorially limited nation-state. The
collective of the nation-state, however, is not the collective of global
society. There is no political system on the global level, which implies
that society currently lacks the political function. How can the territo-
rially bounded nation-state “steer” society without powers beyond its
borders? And when nation-states attempt to steer global society, there
ensues a Hobbesian war of all against all on the international level.
Finally, assumption four about the possibility of external observation of
society is rejected because it would place the observer outside society,
which is impossible. After all, the social system is a meaning system,
and one cannot go outside of meaning. Meaning alone can observe
meaning. Meaning, for Luhmann, exists as self-observation or self-dif-
ferentiation in perception and communication. The self-reference of the
198 See Luhmann “Globalization or World Society: How to Conceive of Modern So-
ciety?” published in International Review of Sociology (1997).
Network Science 233
meaning system is a major topic in Luhmann’s theory and takes up
much of the early discussion of second-order cybernetics. If society is a
system of meaning, then sociology must become a theory of meaning,
which, as we have argued, implies that it must be able to explain itself
and not merely social “objects” that are observed in a domain located
somehow outside of the observer.199
Having quickly reviewed Luhmann’s four objections to traditional
sociology, let us turn to ANT. Latour would agree with Luhmann in
rejecting these typically modern assumptions about society and the
nature of the social. For ANT, just as for Luhmann’s theory of social
systems, the social and society cannot be reduced to human beings and
their interactions. Contrary, however, to Luhmann’s strategy of exclu-
sion that bans individuals and everything that is not communication
from society, for ANT, the “collective” excludes no one and no thing.
Technologies, artifacts, all things material, and living are in some way
social partners.200 Whereas Luhmann’s concern is to limit the scope of
sociology to the study of forms of communication, for Latour, sociology
finally includes all those actors that contribute to and actively partici-
pate in the world of meaning. Of course, Latour intends to overcome
typically modern opposition between individuals and society. If society
cannot consist of human beings, neither does society consist of the
hypostasized relations among individuals. Such relations were thought
of as “social structures,” what Durkheim referred to as “social facts,” that
exist on a supposed “macro” level above the “micro” level of individual
interactions and which then control individual actions and thoughts, as
it were, from behind the backs of the individuals. Against traditional
sociology, Latour insists that no individuals can somehow exist on their
own in an a-social state of nature until they sign a social contract. Then,
once they have signed this mythical document, social structure, insti-
199 Philosophical hermeneutics, Heidegger, Gadamer, Ricoeur, has always claimed
that the social sciences in distinction to the explanatory methods of the natural
sciences were oriented toward “self-understanding.” For a discussion of the once
hotly debated issue see K.-O. Apel (1988).
200 This becomes apparent when Latour (2013) ocially includes not only technolo-
gy, but also material things and living organisms as a “mode of existence” within
the collective.
From Systems to Actor-Networks
234
tutions, and organizations arise, which make up what Durkheim saw
as the proper object of empirical sociology. For ANT, social space is not
divided into distinct domains of agency and structure; it is “flat.”201 The
distinction between individual and society is a modern myth disguising
the networking that is constructing the collective.
Second, just as for the theory of social systems, for ANT, consensus or
agreement cannot be the basis of meaning. On the contrary, meaning
is the basis for both consensus and conflict. Relations and associations,
which are the very stuff of the social, are not always peaceful and
hardly ever reflect agreement about values, world views, or who and
what and how things are supposed to be associated and what trajec-
tories networks are to follow. Disagreement is just as meaningful as
agreement; conflict is probably more prevalent than accord. For some-
thing to be social, there is only the demand that it, whether human or
nonhuman, enter into associations. And something that does not enter
into associations does not exist. Actor-networks are often conflictual,
continually in revision, pursuing different and contrary goals, and open
to transformation in all directions. The idea that people can only mean-
ingfully talk to each other if they agree on fundamental assumptions
about reality, values, etc., which Habermas proposes, rests upon a very
simplified understanding of communication. It is only when people
disagree on fundamental values and world views that communication
becomes crucial. Otherwise, we all know what is expected of us and
act accordingly, social life collapses into a sterile conformism, nothing
new and surprising happens, information is not constructed, and the
network does not grow.
Thirdly, for ANT, as for the theory of social systems, society cannot be
understood as being territorially bounded. First, networks, by nature,
do not have clear borders or boundaries, especially not territorial
ones. They have distributions, intensities, extensions, branches, hubs,
and peripheries, but nothing marks exactly where a network stops
and something else begins. The actor-network of our hominid hunter
quickly extended to a network of a builder or a warrior. As soon as the
201 See Latour (2005).
Network Science 235
stone ax became a status symbol, it became a network of social order and
advantages in finding a mate. The stone ax also became a ritual object in
a religious ceremony that served to maintain contact with transcendent
forces. The process of extending and transforming the network trajecto-
ries continues until today. Furthermore, actor-networks are constructed
of information; as a relation, all information is, in principle, linked to
all other information. Information cannot be territorially bounded, not
even by firewalls. Information cannot constitute a bounded individual
or what Western metaphysics has called a substance or an individual.
General systems theory also has no use for substances or individuals
since systems are, by definition, compositions of elements constructed
by the system. Systems theory, however, does require that systems be
bounded by the constitutive difference to an environment. We have seen
that this theoretical demand causes problems for Luhmann’s theory
of society as a system of meaning and communication, for meaning is
information, and information is relational. The boundary of a system of
meaning must itself be meaningful and thus fall within the system, or
to put it more precisely, link the outside to the inside such that no real
boundary exists. On the level of emergent order of meaning, relations
are real, and every boundary is a relation. Therefore, the necessary
exclusion of an environment constative of systemic order in all its forms
is lacking in meaning systems. But without this constitutive boundary,
there is no selecting, relating, and steering of system elements and oper-
ations. As argued above, this implies serious problems for a systems
model of meaning. Since meaning is all-inclusive, and there can be no
system of everything, it would seem that Luhmann’s attempt to under-
stand meaning as a form of systemic order should be revised. We have
argued above that the unique characteristic of meaning as constituted
by information makes it impossible to model meaning as a system
successfully. Meaning knows no boundary other than such that it itself
constructs for various purposes. We argue that conceiving of meaning
as a network and not as a system avoids this problem. Networks are not
constituted by boundaries, as are systems, but by associations, links, or
relations. Networks are scalable. They can expand almost indefinitely
or shrink to small groups of actors and few associations. Networks
differentiate, as we shall see below, but never by binary exclusion. For
From Systems to Actor-Networks
236
ANT, meaning is necessarily ordered as a network, and only under
certain circumstances, where boundaries can be temporarily drawn to
construct functional black-boxes, can it be modeled as a system.
Finally, the fourth assumption that traditional sociology makes when it
proceeds from the idea that society can be observed “objectively,” that
is, from “outside,” has been debunked by ANT’s close description of
how science actually works. Science in action, the title of one of Latour’s
first works, shows that the traditional epistemology, which understands
science as a disinterested knowing of supposed facts, is mostly myth.
ANT has for decades documented how many social processes construct
scientific facts. Society, no more than anything else, can be an “object”
for a” subject” in the sense that modern epistemology and philosophy
of science supposes. Not only ANT but Science and Technology Studies
have shown that science is not made up of subjects and objects. What
things are and how they appear in laboratories results from long and
complex constructions, interactions of humans and nonhumans, indeed
artificialities. The ideas of the subject and the object portrayed in modern
philosophy of science have nothing to do with science as it is practiced
and are largely modern myths. What philosophical hermeneutics has
always claimed turns out to be true not only for the social sciences but
for science in general. All knowledge is interpretation, and there is no
dichotomy between “understanding” and “explanation.” Different
kinds of networks construct information in different ways and thus
embody different kinds of knowing. But none of these “modes of exis-
tence,” as Latour (2013) calls them, are adequately modeled according
to the traditional distinction between subject and object and the idea of
a pure, value-free, and disinterested knowing.
At this point, we can ask what the rejection of these four traditional
principles of modern sociology by both Luhmann and Latour means
for the self-understanding of the social sciences. If both Luhmann and
Latour mark a clear transition from classical modern social science to
what could be called a “post-modern” theory, what is the “object” of
the social sciences, and what is the function of science in society? Let’s
look back at the development of the science of sociology. We see that it
Network Science 237
has been based on a fundamental assumption of the difference between
culture or society on the one side of an ontological divide and nature
on the other. Modernity divided reality into ontological domains, with
nature on one side, society on the other, the object on one side of an
ontological divide, and the knowing subject on the other. This distinc-
tion goes back to Descartes’ distinction between res cogitans and res
extensa, that is, thinking beings who are free and somehow cannot be
reduced to matter on the one side and natural beings who exist under
the rule of determinate causality on the other. This ontological divide
grounded the typically modern self-understanding of the human
(Humanism) and the later distinction between the natural sciences and
the social sciences. Sociology, as it was conceived at the end of the 19th
Century, was a science concerned with free subjects who enter into rela-
tions with one another to build society and culture. It was a science of
the relation between the one and the many. The fundamental question
was how individuals are integrated into society and how society condi-
tions individuals. Although it was admitted that society does not obey
the determinate laws of nature, if there is to be a “science” of society and
not merely philosophical speculation, it must be assumed that society
obeys its specific cultural and historical laws. The latest version of
these laws is what network science discovers, as well as the new social
physics.202 These were the laws that Durkheim, for example, claimed to
be discoverable by the newly founded empirical science of sociology.
Whatever these laws may be, they are what the social sciences have
been looking for ever since.
Both systems theory and ANT reject the traditional assumptions of
sociology and allow us to go directly to a different conception of society
and sociology. Although Luhmann must be credited for introducing
meaning as an independent level of emergent order beyond matter and
life, a level of order upon which society comes into being as a system
of communication, the ontological status of matter and life in relation
to meaning in Luhmann’s theory remains unclarified. For Luhmann,
as for all traditional sociology, the realm of nature, things, biological
202 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Social_physics.
From Systems to Actor-Networks
238
bodies, artifacts, and technology are not part of the social system. There
can be nothing in the social system except communications. Every-
thing else is banned into the environment, which is assumed to exist,
somehow beyond the city walls of the social system. Luhmann care-
fully maintained the fundamental distinction between society, nature,
and even society and technology. For Luhmann, all these things are,
in a certain sense, part of the social system, but only insofar as they
are the subject matter of communication. Society is communication, not
material things, living beings, artifacts, or even human individuals. The
social system consists of communication and not what is talked about,
except when communication talks about itself, which is a special case
whose difficulties have led to our criticisms. The things that are spoken
of are not part of the social system; only their meaning in the form of
selected information is part of society. When things become informa-
tion, they can be communicated and, in this sense, have meaning and
are part of the social system. But information is not what they are, their
mode of existence, as Latour will say. They exist in their own ontolog-
ical domain, whether or not they are being talked about. For the social
systems theory, the typically modern gap between society and nature
remains wholly intact and unquestioned. Physical and biological
systems do not exist on the same ontological level as meaning. There
is meaning and other kinds of systems, whatever that might “mean.”
For Latour, on the contrary, meaning is constructed not by commu-
nication but by networking or, as we have seen, technical mediation.
And since there is nothing that does not appear in an actor-network,
and thus as in some way or another technically mediated, meaning is
not in any sense of the word a “realm” somehow distinct from matter,
life, or artifacts. Meaning is not one domain of being alongside others.
Meaning is actor-networks, consisting of both (pre)humans and nonhu-
mans. The world, which is the only world we know, is always already
“social” because it and everything in it is constructed by relations. And
since relations are information, the world of relations is always already
meaning.203 Actor-networks are always at once natural and social. This
203 We should not forget that the moment one particular stone became an ax, all the
other stones became “not-axes” or “potential-axes.”
Network Science 239
is what Latour’s principle of irreduction implies. There is no nature “out
there” conceived of as a “thing in itself” that we cannot know – Kant’s
critics quickly asked how he could know about the noumenon – or a
realm of objective facts governed by determinate causality and ontolog-
ically distinct from society as a realm of freedom, subjective opinions,
cultural artifacts, and historically changing institutions.
Therefore, the sociology that ANT proposes represents a more radical
break with the tradition of modernity than Luhmann’s systems approach.
To mark this fundamental difference between ANT and traditional
sociology, Latour prefers not to speak of sociology at all but instead of
“associology,” a theory of the associations of humans and nonhumans
that make up the world of meaning. As Latour (Latour 2005:5, 9) puts it:
Even though most social scientists would prefer to call ‘social’
a homogeneous thing, it’s perfectly acceptable to designate by
the same word a trail of associations between heterogeneous
elements. Since in both cases the word retains the same origin—
from the Latin root socius—it is possible to remain faithful
to the original intuitions of the social sciences by redefining
sociology not as the ‘science of the social,’ but as the tracing of
associations. In this meaning of the adjective, social does not
designate a thing among other things, like a black sheep among
other white sheep, but a type of connection between things that
are not themselves social.
For Latour, the science of sociology cannot define itself in opposition
to or as somehow similar to the natural sciences, nor can the natural
sciences continue to assume they are discovering objective facts of
nature independent of all social concerns and embeddedness in soci-
ety.204 It can be said that ANT extends and completes the universaliza-
tion program of sociology that began with Spencer and Durkheim and
was continued by Luhmann. The idea of the sociology of knowledge
204 Latour (2004) proposes that one speak of “maers of concern” instead of mat-
ters of fact. The very idea of an “objective” fact of “disinterested” knowing is an
illusion of Western modernity and ignores how knowledge actually comes into
being.
From Systems to Actor-Networks
240
completed this program. But ANT goes even further in that it acknowl-
edges that sociology is no longer a science of society distinction from
the science of nature or even that such ontological domains as nature
and culture exist. Things, artifacts, and nature are no longer excluded
from the realm of the “social” understood as everything that is “asso-
ciated,” that is, everything that is networked and thus constituted by
information. The traditional distinctions between domains of scientific
investigation and disciplines require revision.205
Despite the theoretical advances of general systems theory and ANT,
the distinctions and assumptions typical of Western modernity remain
mostly unquestioned. Although for both Luhmann and Latour, tradi-
tional modernity has passed, the convictions of classical modern thought
have not lost their persuasiveness and influence. Latour suggests in a
provocative and radical statement that we should acknowledge that
“we have never been modern” and leave the conundrums of modern
philosophy, nature and society, subject and object, idealism and mate-
rialism behind as we move into a networked future.206 What does the
networked future that ANT envisions look like?
The relational ontology of ANT conceives the world as an assembly of
associations, replacing traditional substance ontology, which conceived
of the world as a totality of things. For ANT, meaning is not something
that is added on to things by cognitive acts, signs, mental states, or
language, but that which things are. The world of meaning is a universal
process of networking and information construction in which asso-
ciations are constantly being constructed, transformed, scaled up or
down, extended, connected, etc. Networks are not things, even ordered
compositions of things, as systems theory sees the world, and they are
205 The status of science itself is an important question in both systems theory and
ANT. Luhmann’s sociology hardly ts into the functional subsystem of science
whose true/false coding demands methods of verication. And Latour’s vision
of an associology raises the question of the status of the associations it itself is
making, whether they belong to what Latour (2013) calls the specic ”mode of
existence” of science (REF) or to some other mode of existence, for example that
of networking (NET), or are generally applicable to all modes of existence thus
forming a metalanguage.
206 This is the explicit program of Latour’s An Inquiry into Modes of Existence (2013).
Network Science 241
not cultural or social things as opposed to natural things. Networks
are processes of networking. Just as life may be seen as a process of
self-organizing autopoiesis, meaning is a process of networking. The
network is the actor in action. ANT offers an alternative to traditional
ontologies and general systems theory by proposing a theory of network
order according to which networks are not closed, functional entities
but open, flexible, scalable, and dynamic forms of order that can be
defined as information. This takes up, extends, and grounds Luhmann’s
notion of society as a system of meaning but interprets communication,
which for Luhmann is the operation of the social system, as processes of
information construction through networking. ANT does not claim that
Luhmann’s description of autopoietic, informationally and operation-
ally closed systems is false. Systems theory can usefully model certain
forms of functional networks within society. ANT refers to such struc-
tured networks as “black boxes,” that is, fixed input/output systems.207
But what is true of the functional subsystems need not be true, as we saw
above, of society as a whole. If the whole of society cannot adequately be
modeled as an automated input/output system. What is it then?
With the close empirical observation and description typical of
ethnological fieldwork, Latour meticulously followed the scientists’
day-to-day work in laboratories.208 He soon discovered that the idea of
science as a closed system in Luhmann’s sense, a domain that is clearly
distinguished from other social domains, did not stand up to scrutiny.
What the ethnologist investigating laboratory life and science in the
making sees is
… visits by a lawyer who has come to deal with patents, a
pastor who has come to discuss ethical issues, a technician who
has come to repair a new microscope, an elected official who
has come to talk about voting on a subsidy, a ‘business angel’
207 See Latour (1999:304) “When a machine runs eciently, when a maer of fact
is seled, one need focus only on its inputs and outputs and not on its internal
complexity.” This could be said of Luhmann’s functional subsystems. They are
eciently running communication machines.
208 See Latour for example Laboratory Life (….), Science in Action (…..), but also An
Inquiry into Modes of Existence (2013).
From Systems to Actor-Networks
242
who wants to discuss the launching of a new start-up, an indus-
trialist concerned about perfecting a new fermenting agent, and
so on. (Latour 2013:30)
To assume that all these people from different social subsystems
do not communicate with each other but go to great trouble to come
together only to “disturb” each other, as Luhmann’s theory would
have us believe, is, no matter how much “structural coupling” we
might assume, hardly credible. The scientists themselves insist that all
these different people with their very different goals and concerns are
necessary to the laboratory’s success, and they are not at all interested
in drawing boundaries or excluding everything that is supposedly not
“science.” On the contrary, the actual practice of science actively seeks
and maintains associations with many other different social activi-
ties. For a laboratory to function, contrary to Luhmann’s banning of
all non-scientific communication into the environment of the science
system, the laboratory depends on communication and close cooper-
ation with many others. For example, the operation of laboratories in
universities is closely linked to educational research programs, which
are not directly concerned with constructing scientific proofs but with
certifying skills, with business investors concerned with knowledge
transfer, innovation, and profits, or with regulators involved with
policies, and civil society actors motivated by public interests. Labora-
tories in the private sector operate in close cooperation with business,
market pressures, competition, government regulation, civil society
watchdogs, etc. Finally, religion and politics play a role in what science
investigates even down to methods and theories. Science, in short, is
not just science; it is business, education, politics, religion, law, and
even art, but of course, doing experiments in a laboratory is not the
same as contributing to a dossier on proposed regulation or designing
a curriculum for a masters degree in engineering. These are all different
forms of constructing different kinds of information. Nonetheless, the
realities of science in action imply that Luhmann’s view of society as
consisting of autonomous functional systems informationally closed
to one another does not tell the whole story. In practice, science and
society resemble much more closely an open network with many cross-
Network Science 243
overs and associations than a collection of autonomous, operationally
and informationally closed systems that are more or less successfully
adapted or structurally coupled to each other.
The ethnologist also soon discovers that scientific experimentation
cannot adequately be described as verification or falsification of hypoth-
eses about the objects under investigation. Bacteria or molecules are not
passive objects deterministically reacting to the activities of researchers.
The very meaning of the highly valued “objectivity” sought in science
consists of letting the objects of inquiry also have something to “say”
about the conditions under which they manifest themselves and the
roles they play.209 The actual work of experimentation is more like a set
of challenges in which the objects of study are artificially enabled to act
in specific, and often unexpected, ways by going through many “medi-
ators,” such as microscopes, particle colliders, or measurement devices,
and various kinds of inscription and documentation. The results of
an experiment are never immediately apparent, and no matter how
often they are repeated, the results and interpretation could always be
challenged and changed. The much-celebrated characteristic of “objec-
tivity,” what for Luhmann is the effect of the specific true/false coding
of the science system, itself implies the object does not deterministically
do what is expected of it, or what it is “told” to do. The object must be
brought to say something about itself through the mediation of many
artificial instruments and settings. Experimentation is, therefore, rather
like a process of negotiation by which roles are discovered and assigned
among human and nonhuman actors, negotiations of a specific kind
in which both sides set conditions, show activities, demonstrate
appearances and capabilities, and “mediate” the construction of infor-
mation in many different ways. For this reason, Latour suggests that
the traditional self-understanding of science as active subjects capable
of a pure knowing of passive objects or so-called “facts” does not do
justice to the empirical practice of scientific work, which is “coded”
by variegated practices of inscription, documentation, demonstration,
experimentation, etc. which Latour calls chains of “reference” or chains
209 Latour has demonstrated this clearly in his work on Pasteur, The Pasteurization
of France (1993).
From Systems to Actor-Networks
244
of “immutable mobiles.”210
Far from being a closed system processing only communications that
can be either true or false, science in action looks much more like an
open network of heterogeneous and hybrid actors pursuing different
and often conflicting goals. The laboratory is not defined as scientific
by means of excluding everything legal, educational, economic, polit-
ical, religious, artistic, etc., but by including actors that are scientific
AND legal AND economic AND educational AND religious, etc., but
in its unique way. The actors involved are not merely human but also
nonhuman, not only cognitive or communicative but also material, tech-
nical, and institutional. The network does not operate self-referentially
like a closed system but expands and retracts in all directions simulta-
neously, depending on the changing needs, purposes, and programs of
action of all the actors involved. In distinction from systems, networks
are principally open in all directions and tend towards increasing rather
than reducing complexity.
This brief review of how both Luhmann and Latour distance them-
selves from modern sociology raises the question of the status of the
social sciences, at least, the status of these new theories. Are these theo-
ries considered “science” in the sense of that particular social activity
concerned with generating “knowledge?” Are we still dealing with
“science” as it has been traditionally understood in Western modernity?
Or do the theories that Luhmann and Latour offer mark a paradigm
change in the social sciences such that one must ask anew what science
is and what role it plays in society? Indeed, in what kind of society does
science play the role it does? The critique of the modern philosophy of
science and epistemology, as well as the doubts both authors express
towards modern sociology, seem to indicate that the theory of social
systems and ANT demand a reappraisal of the status, function, and
nature of scientific knowledge, or at least the status of their metatheo-
ries. An answer to this question rests upon the answers these theories
give to how society is differentiated into various functional subsystems
or different forms of networks.
210 See Latour (2013).
Network Science 245
We have seen that, for Luhmann, science is a functional subsystem
delimited by its unique binary code of truth/falsity. The system of
science consists of only those communications that can be coded, with
appropriate methods, institutions, organizations, etc., as either true or
false. All other communications, such as those dealing with whether
something is legal or illegal, bought or sold, relevant for the certification
of skills or not, etc., are banned into the environment of the functional
subsystem of science. They perturb science, but they are not scientific.
Despite the fact that sociology is only one discipline within the func-
tional subsystem of science, Luhmann reserves a special place for what
he calls “supertheories,”211 that is, those foundational theories that serve
as paradigms in the Kuhnian sense and are not subject to falsification.
Supertheories are neither true nor false. They set the conditions or
parameters within which anything can count as science. They function
as the code, the borders of the science system.
As a supertheory, Luhmann’s theory of social systems is, therefore,
not strictly speaking science. It does not fall clearly within the bound-
aries of the system of science. Instead, it delimits and differentiates
social communications that can count as science. The differentiation of
communication into functional subsystems, one of which is science, is
an autocatalytic, self-organizing, and constructive process within the
social system. It is the self-differentiation of the meaning system. The
supertheory cannot be understood as transcendental knowledge of the
conditions of the possibility of science, such as the self-reflection of a
transcendental subject, as modern epistemology since Kant has dealt
with this issue.212 Science claims universality since there is nothing that
cannot be scientifically investigated, not even science itself. However, it
achieves universality only by differentiating itself from other functional
211 Luhmann (1995:4-5) “Supertheories are theories with claims to universality (that
is, to including both themselves and their opponents). If a supertheory achieves
a signicant centralization of dierence, then a paradigm change also becomes
possible. Systems theory is a particularly impressive supertheory.”
212 Luhmann (1992:13) rejects modern subjectivism and thus also the transcenden-
tal/empirical distinction. The meaning system aempts to make itself meaning-
ful as a system. We have argued that is questionable, since meaning may not be
a systemic form of order.
From Systems to Actor-Networks
246
subsystems and society as a whole. The same, of course, could be said of
all the functional subsystems. Luhmann can claim, and he does, that the
theory of social systems does what all scientific theories, at least foun-
dational theories, do: it offers a framework for heuristics, hypothesis
construction, and research programs. While this is, to a limited extent,
valid, the universal claims and the level of abstraction of the theory point
toward a paradigm going beyond the boundaries of normal science. If
the theory of social systems is not clearly located within the system of
science, where is it, and what function does it fulfill? Luhmann locates
his theory on the level of a supertheory that itself is neither within nor
without the system of scientific knowledge it supposedly explains. We
have argued above that this mysterious form of knowing, in all its forms,
cannot explain itself as long as it assumes that it is a system. We have
argued that with Luhmann, meaning, as it were, misunderstands itself
as a form of systemic order. In short, we claim that if meaning knew
what meaning is, it would understand that it is not a system. A theory
of meaning is meaning’s self-explication. Is a self-understanding of
meaning science? For general systems theory, this appears to remain an
open question. How do things stand with ANT?
At first, it seems that ANT is only concerned with how everything is
connected to everything in networks; that is, ANT seems to emphasize
the continuity of the social and not its differentiation. Latour is openly
critical of talk of “domains” that would break up the collective by
distinguishing society from nature, subjects from objects, knowledge
from opinion, micro form macro actors, structure from agency, and the
transcendental from the empirical.213 ANT emphasizes the continuity
between networks to clearly outline a theory of society as the universal
collective of humans and nonhumans. Society, or the collective, indeed,
the world, is through and through networking. There is nothing that is
not networks networking. How, then, does ANT answer the question
of social differentiation? Is the collective differentiated at all? And if so,
how does this influence the status of ANT and the social sciences?
213 Interestingly, Latour seems not directly to have commented on Luhmann’s theo-
ry of functional subsystems or taken up a dialogue with systems theory.
Network Science 247
3.6NetworkDierentiation
Latour’s description of society from the actor-network perspective is
presented in his major work, An Inquiry into Modes of Existence (AIME)
(2013).214 The work is subtitled An Anthropology of the Moderns, which
indicates an empirical and descriptive methodology and emphasizes
that the goal is not to offer a general theory of society but a descrip-
tion of modern Western society. Latour’s stated purpose is to describe
the “Moderns” much as an ethnologist doing fieldwork with a tribe
in Borneo would do, that is, to empirically describe the Moderns such
that what they truly value comes to the fore, apart from their often
misleading beliefs about themselves. An ethnology of the Moderns is
needed because it allows the Moderns to see themselves in a different
light, from the perspective of an “objective” description of their prac-
tices and beliefs.215 This is important for Latour because it supposedly
allows the Moderns to enter into a fruitful and promising dialog with
other cultures, a dialogue which the misleading and arrogant self-un-
derstanding of the Moderns has hindered for centuries. The unprec-
edented problems facing society today are global: climate change,
migration, pandemics, financial insecurity, etc. The age of global
dialogue on common issues, above all, the climate crisis, has come, and
modern Western society is no longer in a position to dictate the terms
upon which a global solution could emerge. The description of modern
society that Latour develops under the title of An Inquiry into Modes
of Existence (AIME) seems, therefore, not to have a purely “scientific”
motivation. It is not merely a question of whether the description of
the Moderns he offers corresponds to how the Moderns are, but it has
a normative purpose. It aims to enable cultural comparison, mutual
understanding, and global negotiation such that the problems facing
humanity today and in the future can be addressed cooperatively and
thus effectively. In addition to this explicitly normative purpose, the
main title of the work suggests that it is concerned with philosoph-
214 This is the main publication of an EU funded research project whose Website is
hp://modesofexistence.org/.
215 “Objective” in this context means that the self-description of the Moderns is not
taken at face value.
From Systems to Actor-Networks
248
ical questions and not merely empirical descriptions. An inquiry into
“modes of existence” suggests ontological and metaphysical concerns,
usually considered philosophical and not scientific. Therefore, the
reader is confronted right at the outset with whether Latour’s enterprise
is science or philosophy or something else altogether.
Latour quite openly admits that his enterprise is not normal science and
not even the attempt to establish a paradigm as Luhmann understands
his project of a “supertheory.” Latour himself expresses uncertainty
about the status of his theory. While he admits that his inquiry into the
beliefs and practices of the Moderns is not normal science, he has no
clear answer about what it is, even though the role it is intended to play
in society seems to be clear.
…it is not in the mode of knowledge that I claim to be
working. The term “inquiry” has to be taken in a plurimodal
sense whose object is to preserve the diversity of modes.
Can we call this approach “empirical philosophy”? I am not
sure, given how indifferent philosophy has become to the
tasks of description. Experimental metaphysics? Cosmopol-
itics? Comparative anthropology? Practical ontology?
(Latour 2013:481)
Whatever one calls it, it seems clear that for Latour, science in general
and his own theory is an endeavor that serves practical purposes. It
has a goal beyond objective knowledge of how modern society is
differentiated into what Latour calls “modes of existence.” Of course,
because it is “construction,” all information construction has a built-in
value judgment. The idea that information must be “constructed” in
the threefold sense of distributed agency, uncertainty of outcome, and
valuation mentioned above prohibits any assumptions about science
being a pure and disinterested knowing of objective facts. At the same
time, it implies that all the modes of information construction are
goal-directed in that they are “obliged” to construct well and not badly.
Science in action, for example, is a process of translating and enrolling
actors in actor-networks in ways directed to constructing objective
Network Science 249
knowledge. This goal is what Latour calls a “value,” and practices of
networking that are held together by being directed toward a particular
value show that we are dealing with a specific “mode of existence.” As
a mode of existence, science, for example, is not to be confused with
other networking practices such as law, religion, business, politics, etc.
These different practices of networking are directed toward their own
values. If we ask where in this framework of modes of existence we can
find the kind of science that Latour himself is doing, we must ask what
its goal and value are. The value that Latour strives for is global cooper-
ation based on mutual understanding. Is this science? If not, what is it?
And what is the connection between the objective knowledge about the
Moderns revealed by the ethnological investigation and the purpose
of global negotiation guiding the inquiry? One must ask of Latour the
same question he asks of the Moderns. How does the self-description
of ANT compare to the description ANT offers of science? Is ANT itself
not a form of networking? If so, how is it different from science? If it is
different from science, what is it, for it is certainly not religion, politics,
business, or any other of the various modes of existence?
For Latour, science consists of a particular kind of networking that mobi-
lizes artifacts or “inscriptions” in relation to each other such that the content
transported by these artifacts remains the same throughout long chains
of reference. The artifacts can be anything: a diagram, a measurement, a
graph, a sample, a demonstration, a textual document, images, models,
simulations, or anything that can move information without changing it.
Fittingly, Latour calls such artifacts or inscriptions “immutable mobiles”
and locates their purpose in making remote objects accessible. For
example, a carefully arranged set of earth samples in a wooden lattice that
is assembled in the field “represents” the exact relations and content of
that area of land under investigation when transported hundreds of miles
away into a laboratory. The assemblage, the artifact, is mobile, but it trans-
ports the original object, the relative composition of the soil, unchanged
or immutably to a distant place. In the laboratory, the soil samples are
chemically analyzed utilizing special instruments whose results are trans-
ported to a graph, which is then compared to other graphs to end up in a
scientific paper published in Nature or on a PowerPoint slide presented to
From Systems to Actor-Networks
250
students, government agencies, or business developers. Throughout this
entire chain of references, the remote object, that is, the composition of the
original soil, must be kept constant. One must be able to follow the chain
of references backward to the original land and “reproduce” the results.
Otherwise, what comes out at the end is not “objective.” Only by means
of such chains of references among immutable mobiles can “objective
knowledge” be constructed. Assembling long chains of such immutable
mobiles is the specific mode of networking (Latour speaks of “mode of
existence”) of science as distinct from law, religion, politics, art, etc., which
all have their own ways of constructing information. But is this the science
that Latour himself is doing?
On the one hand, Latour claims to offer careful empirical descriptions of
the experiences the Moderns make in their various modes of existence.
These phenomenological descriptions can be “verified” by all readers.
Every reader has the right and duty to ask: Does this description fit with
my experience? Science and Technology Studies has operated success-
fully with this methodology for decades, and Latour is convinced it has
passed the test of scientific verifiability. Luhmann would probably agree
and say that science is a functional subsystem operating according to
its own code of true/false, distinguishing scientific communication from
other kinds of communication in other social subsystems. But for Latour,
science is not an operationally and informationally closed system. It is
an actor-network and is open to information from all other networks.
The collective is not ordered in the same way that Luhmann describes in
terms of functional differentiation. There is more to the collective than
the functional “black boxes” that are autonomous social subsystems. In
order to express this idea, Latour (2004a; 2005:98ff.) has introduced the
term “matters of concern” in opposition to “matter of fact.”
Matters of concern are not the objective facts of modern science, where
the term “objective” means that nothing is coming from a subject, and
nothing is artificial, social, or constructed about them. The objective facts
of modern science exist simply as they are in the ontological domain
of nature until science somehow, without any mediation, construction,
or social effort, “discovers” them. Matters of concern, on the contrary,
Network Science 251
are those facts (matters) that only reveal themselves by mediation and
construction. Without “concern,” without the “gathering” of many
different actors, trajectories, contingencies, and uncertainties constantly
being disputed and revised, and without the unpredictable crossings
into different networks, no scientific fact could come into being. To be
precise, the objective knowledge Latour identifies as the “value” or
goal of the mode of existence called science is neither natural nor social,
neither subjective nor objective, neither artificial nor natural. These typi-
cally modern distinctions are what the ethnological description of how
science works places into question. According to Latour, they are to be
replaced by the question of whether knowledge has been constructed
well or badly, whether all relevant actors have been taken account of
and “gathered” together properly.
What does this mean for the status of science and its role in society?
As a practical activity, science, like all other practical activities, can be
considered from the perspective of “design.”216 Science, for Latour, is
not confined to an operationally and informationally closed subsystem
of society concerned exclusively with distinguishing between truth
and falsity. Recall the above-cited example of the laboratory director
who must deal with law, business, government regulations, educa-
tional curricula, private donors, logistic chains, machines of all kinds,
and much more to do science. The anthropologist of the Moderns sees
that science is intimately bound up with many other social activities
and with many nonhumans. Science is not pure knowing, somehow
performed by a disinterested observer, which is how the Moderns
mistakenly think of science. It is practically involved in constructing
the many programs of action that influence the conflicting trajectories
of the global network of humans and nonhumans, or the “collective,” as
Latour calls society. For Latour, the scientist, the matter of concern that
216 The knowledge discovered by science helps in the “design” of institutions. “…
the proposition I am exploring through this inquiry consists in using a series of
contrasts to distinguish the values that people are seeking to defend from the
account that has been given of them throughout history, so as to aempt to es-
tablish these values, or beer yet to install them, in institutions that might nally
be designed for them” (Latour 2013:7). For a reinterpretation of action in general
as “design” following Latour see Belliger/Krieger (2021).
From Systems to Actor-Networks
252
must be appropriately gathered is the collective as a whole. When it
comes to social differentiation, the question is not to describe functional
subsystems as Luhmann does, but to gather the collective in the right
way, to make the collective the most important “matter of concern.”
There is a direct connection between striving for objective knowledge
(the mode of existence of science), which distinguishes the various
modes of existence, and concern for the collective, which strives to
gather all things together.217
Corresponding to these two directions, it is not surprising that Latour
explicitly acknowledges the philosophical character of his project
despite the emphasis on empirical description. There is a tension
between title and subtitle because the title speaks of modes of existence
and awakens expectations of a philosophical ontology or metaphysics,
not merely empirical anthropology. Indeed, Latour claims to offer a
“pluralistic ontology.” This claim is directed against a certain kind of
metaphysics, which has always been interpreted as a general theory
of Being applicable to all beings no matter how they differ. According
to this view, reality is one, but there are many ways of speaking of it.
Latour, the empiricist, has never been comfortable with universalistic
claims of philosophy and especially metaphysics, which, as he puts it in
a well-known self-criticism (Latour 2013:35), ends up “saying the same
thing about everything.” If everything is a network, how are things
different? There are many concepts in ANT besides the concept of
“network” that seem to apply to all entities, for example, “translation,”
“enrollment,” “mediation,” and “agency.” In our reading of ANT as
a theory of meaning, we propose not only adding a term Latour does
not use but also openly admitting that this concept is metaphysical.
This is the concept of “information.” We use this concept to describe
what networks consist of, that is, relations/associations as defined by
Latour’s principle of irreduction. We interpret the principle of irre-
duction, admittedly against Latour’s will, as a metaphysical statement
saying the same thing about everything. What it says about everything
is that it is a relation, which we propose to call information or meaning.
217 One recalls the scholastic principle “distinguish in order to unite.”
Network Science 253
This position allows us to interpret Latour’s modes of existence as modes
of information construction or modes of networking. Therefore, the philo-
sophical program of ANT is a contradiction since it proposes at once to
say the same thing about everything while insisting that being is plural.
It would seem that, according to Latour, there are different modes of
existence, but no one Being.218 How can this contradiction be resolved?
Let us turn first to the differentiation of the collective and then return to
the problem of unity. For ANT, society is differentiated not by functional
subsystems but by different networks of information construction. They
are all networks and share the constitutive principles of networking, but
they are different networks insofar as they each have their own “mode
of existence.” As networks and not systems, they are all operationally
and informationally open to each other. In contrast to systems, networks
pursue multiple goals and diverse programs of action. In contrast to
systems, networks are constantly changing, scaling up or down, oper-
ating at both global and local levels. Instead of a society differentiated
into autonomous functional subsystems, which Luhmann describes,
we have a “collective” of humans and nonhumans that is differentiated
into different forms of networking and different modes of information
processing. Since information is Being, these networks can be called
“modes of existence.” Where Luhmann speaks of functional subsystems,
Latour speaks of modes of existence. Like Luhmann, for whom there is
only one world society, for Latour, the collective is global. But against
Luhmann, this “post-modern” society is characterized by “network
differentiation” instead of functional differentiation. Luhmann often
seems not to fully realize how far the social systems theory has moved
beyond modernity, for he characterizes “modern” society by functional
differentiation. For Latour, “post-modern” society is defined by network
differentiation. What is network differentiation from the point of view of
Latour’s description of modes of existence?
There are many points of comparison between Luhmann’s functional
218 We do not aempt to resolve this paradox but by means of our interpretation
of ANT as a theory of meaning and by emphasizing information we propose to
simply side step it.
From Systems to Actor-Networks
254
subsystems and Latour’s modes of existence. The vocabulary is different,
but often something quite similar is intended. Instead of functions, Latour
speaks of values. For example, the “value” embedded in scientific prac-
tice is “objective knowledge.”219 The ways in which objective knowledge
is obtained are called the “conditions of veridiction,” the specific ways
scientific truth is constructed by means of chains of reference. These
conditions of veridiction are different for law, religion, or politics, just
as the programs of action, values, or goals these networks pursue are
different. For religion – Latour speaks only of Christianity – the mode of
veridiction is the experience of being personally addressed by a saving
power. For law, establishing a legal judgment depends on marshaling
the necessary “means.” For politics, it is the continuously renewed
assembly of all participants negotiating an issue. For art, what Latour
calls “fiction,” there is the appearance of form in a material of some kind.
For organizations, there is the narrative ordering of actors, roles, and
purposes by means of at once authoring and being subject to “scripts.”
For individuals, there are the many ways of dealing with those trans-
formative forces guiding or hindering psychological development. For
technology, there are many devices, detours, and delegations of action
onto artifacts of all kinds. For social structures, there is what Latour
refers to as “habit,” the automatisms of behavior, belief, and action that
take over as soon as one does anything on a regular or continuous basis
and conforms to expectations. For business or the economy, Latour
speaks of “attachment” or “passionate desires” for things. Interestingly,
he does not mention money or anything like Luhmann’s symbolically
generalized media. For morality, Latour identifies “scruples” about
whether any information construction has been done well or badly.
Since construction necessarily involves the value question of whether
something has been constructed well or badly, morality arises as a mode
of existence from the scruples one must have about all actions.
In all of these modes of existence, we have no difficulty recognizing
topics that have long preoccupied sociology, psychology, economics,
219 Latour reserves the concept of “truth” to describe the “felicity conditions,” or the
way in which each mode of existence retains its unique practices in the achieve-
ment of its goals or it value. All modes of existence have their own “truth.”
Network Science 255
political science, ethics, and other human sciences in one way or
another. What strikes the reader as unusual in Latour’s description of
these well-known phenomena is that although he focuses on how we
subjectively experience them, the phenomena themselves are granted
a peculiar ontological status as distinct modes of existence. There is
a strange identity between experience and being; for example, moral
scruples are not merely subjective experiences people “have” but beings
that have a grip on people. Forces of psychological development and
transformation are not internal, mental experiences that people “have”
but beings that hold sway over people. The passionate interests humans
have for things of all kinds that drive production and consumption in
the economy are not merely strong emotions but beings in their own
right. This surprising and disconcerting confusion of subjectivity and
objectivity is, of course, intentional. It aims to overcome the Modern’s
false understanding of themselves as subjects distinct from objects,
agents separate from a passive world of things, and belonging to a
social and cultural realm distinct from material nature.
The investigation into modes of existence ends with a list of fifteen
modes or different kinds of beings.220 The fifteen modes, a list which
Latour admits is provisional, are divided into four groups. Latour
distinguishes between modes of existence that, in his view, are more
related to “quasi-objects,” such as Technology, Fiction (Art), and Refer-
ence (Science), and those related to “quasi-subjects,” such as Politics,
Law, and Religion. In this context, the term “quasi,” which Latour
takes over from Serres, refers to the assertion that these beings are
neither subjects nor objects, neither purely social nor purely natural,
but somehow both together. Of course, this can be said of all the modes
of existence. Nevertheless, there are also modes of existence that are
unrelated to subjects or objects. These are Reproduction (i.e., the subsis-
tence of material things and viability of living organisms), Metamor-
phosis (i.e., forces of psychic development and transformation), and
Habit (i.e., behavioral automatisms, social structures, frames). There
is a fourth group of beings that link quasi-subjects and quasi-objects.
220 See Latour (2013:488) for a table listing all the modes and their respective charac-
teristics.
From Systems to Actor-Networks
256
These are Attachment (i.e., business, the economy), Organization (i.e.,
the narrative ordering of roles, times, and places), and Morality (i.e.,
scruples about whether construction is well done or not). And finally,
there are modes of existence that apply to all others and, thus, to all
beings. These are Network (i.e., entering into associations), Preposition
(i.e., marking the distinctions between modes), and Double Click (i.e.,
that mode which covers over the networking to which all beings owe
their existence as well disguising thereby the distinctions between
modes). Double Click, a reference to a computer mouse’s commands
that initiate and conceal complex processes, is responsible for the typi-
cally modern self-misunderstandings of Subject/Object, Nature/Society,
Agency/Structure, Individual/Society, Fact/Value, etc.
Latour’s ethnology of the Moderns and their practices has revealed
more and different modes of existence than the functional subsys-
tems Luhmann describes. For Luhmann, modern society consists of
the functional subsystems of science, law, politics, religion, business,
education, media, art, and healthcare. Luhmann also considers orga-
nizations, which are found within all subsystems, to be systems in
their own right, and finally, even face-to-face interactions are consid-
ered a form of systemic order. Latour remarkably omits healthcare,
education, and media but characteristically adds the way of being of
material things, such as stones, mountains, rivers, etc., and of living
organisms, which Luhmann explicitly excludes from the social system.
Luhmann also excludes technology or any artifact from social partic-
ipation. After all, at least until recently, tools and machines do not
communicate. For Latour, all tools and machines, indeed all artifacts
such as the technologies of language, writing, imaging, and so on, are
social partners contributing in their own ways to networking, that is,
to the construction of information. Networking itself is dealt with as a
separate mode of existence, but it is a special way of being since it is
the basis of all modes. The collective as a whole can only be “gathered
together,” which is the very purpose of conducting the inquiry into
modes of existence through networking. We need not go into each of
these modes of existence in detail. As we noted, meaning can under-
stand itself and differentiate itself as it wishes. Our concern is whether
Network Science 257
meaning understands itself better as a system or as a network.
It will come as no surprise that everything hinges on what is meant by
the term “network.” Let us cite a long passage, which we have partially
quoted above that sets the stage for understanding this all-important term.
Instead of wondering, for example, if Science is a domain distinct
from Politics or The Economy or Religion, the investigator will
be content to start with some arbitrary sequence of practices.
For example, she goes into a laboratory: there she finds white
lab coats, glass test tubes, microbe cultures, articles with foot-
notes—everything indicates that she is really “in Science.” But
then, with a certain obstinacy, she begins to note the origins
of the successive ingredients that her informants need in order
to carry out their work. Proceeding this way, she very quickly
reconstitutes a list of ingredients characterized by the fact (in
contradiction with the notion of domain) that they contain ever
more heterogeneous elements. In a single day, she may have
noted visits by a lawyer who has come to deal with patents, a
pastor who has come to discuss ethical issues, a technician who
has come to repair a new microscope, an elected official who
has come to talk about voting on a subsidy, a “business angel”
who wants to discuss the launching of a new start-up, an indus-
trialist concerned about perfecting a new fermenting agent, and
so on. Since her informants assure her that all these actors are
necessary for the success of the laboratory, instead of seeking to
identify domain boundaries, which are constantly challenged
by innumerable erasures, nothing prevents her any longer from
following the connections of a given element, it hardly matters
which one, and finding out where it leads. (Latour 2013:30)
From the network perspective, science does not consist of the scientific
method and its epistemological assumptions or any theory of science
as the truth about facts that speak for themselves. Besides the careful
inscription of experimental results and measurements, science consists
of many other things: artifacts, machines, institutions, regulations,
From Systems to Actor-Networks
258
supply chains, funding procedures and conditions, social concerns,
business applications, etc. All these things and artifacts are subsumed
under the mode of existence of Technology. Interestingly, any of these
“things” can become the starting point of a network of associations
that branches off into all directions into society; indeed, if one goes far
enough, they come to participate in all other modes. If one follows the
trail long enough, anything can be associated with almost anything else.
Let’s go back to our assumption that this is all information construction,
and it is the nature of information to be able to be linked with all other
information. Networking is what the collective in all its forms consists
of. For Luhmann, this situation creates the problem of over-complexity
of communication that must be reduced or channeled in specific ways.
Luhmann solves this problem by introducing the idea of symbolically
generalized media and functional subsystems. For Latour, there arises a
similar problem of describing how the global network of all things and
practices is reduced to forms recognizable as particular social practices
such as science, law, politics, or religion. How is the unity of the global
network, the “world,” differentiated into specific forms of networks?
Latour makes the theoretical move that makes this question possible at
the outset by distinguishing two meanings of the term network.
On the one hand, there is the network of things and practices that are
all associated in ever-widening circles that the anthropologist describes.
This network is global and cannot be limited to any “domain” of
society, whether law, politics, economy, religion, etc. But on the other
hand, there is the network of what flows through any particular setup
of things and practices. Focusing on what is flowing rather than the
fact that it is flowing allows Latour to make an important distinction.
This distinction is borrowed from that which serves as the basis for his
pluralistic ontology, namely, the distinction between how a network
is set up and what flows through the setup of the network. Latour
(2013:32) cites the example of a mobile telephone network:
The fact that information can circulate by means of a cell-phone
network tells us nothing about the way the network has been
put together so as to work, right now, without a hitch: when all
Network Science 259
the elements are in place and everything is working well, in the
digital window of our cell phones what we can track is only the
quality of a signal marked by a certain number of rising vertical
bars (by convention, from one to five). The “network” in the
usual sense of technological network is thus the belated result
of the “network” in the sense that interests our investigator.
And he goes on to add:
The distinction between the two senses of the word “network”
would be the same if she were interested in railroads: following
the tracks is not the same as investigating the French national
railroad company. And it would still be the same if, taking the
word more metaphorically, she wanted to investigate “networks
of influence”: here, too, what circulates when everything is in
place cannot be confused with the setups that make circulation
possible. (Ibid)221
And he concludes:
So under the word “network” we must be careful not to confuse
what circulates once everything is in place with the setups
involving the heterogeneous set of elements that allow circula-
tion to occur. (Ibid)
What the Moderns have done is systematically ignore (Double Click)
the “setup” and create theories explaining what flows, whether it be
the flow of truth, of legal decisions, of contact with transcendence in
religion, of collectively binding decisions in politics, of originality in art,
or of business in the economy. The setup, the network of heterogeneous
associations, the myriad of things, artifacts, institutions, organizations,
etc., all linked together such that a particular kind of flow is possible
(Latour speaks of “pass” through the gaps between the heterogeneous
221 It should be noted that for network science, for example, in the investigation of
internet email networks, the nodes, whether persons or servers, as uniformly
dened at the outset and not further tracked into all the many other associations
and components make constitute them and same is for the links.
From Systems to Actor-Networks
260
entities, or “continuity” preserved through discontinuity), has always
been there as the condition of the possibility of what flows.222 Throughout
modernity, the first meaning of the term network, the setup, has been
enhanced, expanded, and deployed globally. What the Moderns have
actually done with their science, technology, and business is to increase
the number of hybrids enormously. The networks of humans and
nonhumans have dramatically expanded. This is the setup. It is through
these networks that everything, money, goods, knowledge, people,
etc., flows. But the account of the flow, the second meaning of the term
network, has skipped over the setup and gone directly to the results
of the flow. To a certain extent, the setup is itself responsible for its
own forgetfulness. It is because the network of associations of hetero-
geneous things has been painstakingly and laboriously set up so well
and efficiently that it disappears, and only the results are visible. One
sees, for example, that there is knowledge of objective facts, but one has
lost sight of all the detours via things, instruments, inscriptions of all
kinds, and the institutions necessary to arrive at this knowledge. If one
leaves all the detours of the setup out of account, there appears only a
subject knowing an object, and there appears a material world of nature
as opposed to a world of signs, meaning, culture, and society. Out of
this forgetting of networking, a distinctively “modern” worldview
has been created with clear distinctions between society and nature,
subject and object, mind and matter, language and reality, individual
and society, freedom and determinism, rationality and belief, etc. Once
the project of the anthropology of the Moderns has been successful, and
the dependence of the flow upon the setup has come into view, the
Moderns should recognize, as Latour provocatively puts it, that “we
have never been modern.”223
222 See Latour’s (2013:33) denition of the term network: “The notion of network can
now be made a lile more specic: it designates a series of associations revealed
thanks to a trial—consisting in the surprises of the ethnographic investigation—
that makes it possible to understand through what series of small discontinuities
it is appropriate to pass in order to obtain a certain continuity of action.”
223 We are not concerned in this book with Latour’s “archeology” of modern ideas, but
with the network paradigm or order that emerges from ANT and will therefore not
rehearse the archeology of modernity that takes up most of the AIME project.
Network Science 261
What distinguishes ANT from systems theory concerning social differ-
entiation is the fact that the various networks, because they are all
interconnected, “cross” each other, creating thereby a tapestry or tissue
of trajectories of information construction that pursue multiple goals,
are open on all sides and in all directions, expand and contract, and are
constantly changing. This network of networks is society, or, following
Latour, the collective. To be “collected” differs from being selected,
related, controlled, and excluding an environment. Being collected is
different from being selected, whether as an element of a system or a
system under the selective pressure of the environment. Networks do
not arise in an ever more complex environment to which they must
adapt by viably reducing complexity. Comparing Luhmann’s theory
of social systems and ANT, we see that in the place of a view of society
as consisting of autonomous functional subsystems that are merely
structurally coupled or adapted to each other, there appears the collec-
tive of crossing and linking networks. It is these crossing networks
that constitute society as a whole. A global network society, however,
is not bounded or contained within any system boundaries. ANT has
no problem with an unbounded collective. However, the theory of
social systems runs into conceptual difficulties when society must be a
bounded system.
For Luhmann’s theory of social systems, there are no processes of selec-
tion, relationing, and steering that order society as a whole, for there
can be no system of everything. On the other hand, Latour’s technical
mediation and processes of translation and enrollment are ubiquitous.
They must include everything, or at least they should. Other forces and
norms weave the social fabric together rather than merely functional
dependence, understood as the need for an autonomous system to
adapt to its environment. For Luhmann’s functional subsystems, other
systems are only environment, to which any system must be structur-
ally coupled to continue its autopoiesis successfully. This is what it
means for a social system to be “functional.” A network must not be
functional. It can be. Functional networks are referred to in ANT as a
“black box.” Black boxes can be opened and transformed when actors
become mediators again, and information construction begins anew.
Functionalism implies that a social system can become dysfunctional,
as, for example, is increasingly the case for the political system, which is
supposed to provide collectively binding decisions but can do this only
within a nation-state, which is no longer what the “collective” is. When
the collective has become global but organized on the level of territori-
ally bounded nation-states, politics can no longer fulfill its function of
steering society. Instead, we are witnessing a Hobbesian a-social state of
nature and a war or all against all on the international level. It is increas-
ingly apparent that the political system cannot effectively address any
present-day global problems. The climate crisis is a case in point. On
the contrary, the networks described by ANT are unbounded, scalable,
and continuously crossing each other, participating in each other, thus
making up the social understood and experienced as the “world.” For
this reason, Latour is convinced that the network view of society can
lead to a new, or “ecological,” form of global cooperation. Politics, for
example, can become “political ecology.”224 Global governance need not
be a utopian dream.
The network view of society comes to the fore in the example cited
above of the laboratory, where, as we saw, not only science but law,
education, business, etc., come together to make the laboratory a
successful scientific institution. The laboratory illustrates what this
change of perspective from systems to networks means. Even though
the description focuses on science, there is communication and coop-
eration with lawyers, business people, regulators, religious believers,
students, technologists, etc. This complex, multi-network communi-
cation is constantly contributing to and shaping the research that is
going on in the laboratory. Furthermore, there is an ongoing interaction
with nonhumans of all kinds whose contributions are necessary for the
construction of information. What is going on at the laboratory cannot
be adequately modeled as an operationally and informationally closed
communication system being perturbed and disturbed by meaning-
less interventions from various environments, forcing the scientists to
construct the information necessary to adapt to these environments
224 See Latour (2004b) for an exposition of his political theory.
Network Science 263
internally. Latour’s detailed description of what actually happens in
society is much closer to reality than Luhmann’s view through the lens
of the abstract theory of systemic order and the exclusion of nonhu-
mans. The collective as a whole with all its different networks comes
together in the laboratory, but of course, in a special way, different from
how it comes together on Wall Street, or in Congress, or church, or at an
art museum, or in the editorial office of a media company, etc.
The laboratory is indeed society, but society is “articulated” in a partic-
ular way such that the network setup has its own character, way of
translating and enrolling actors, and methods of deciding what flows
through the setup, that is, its specific value. Latour seems to be more
concerned with the diversity of networks than their unity. The Inquiry
into Modes of Existence ends with a list of fifteen different kinds of beings
and not a unified vision of a global society. Of the two principal ways
of ordering information, the list or the narrative, Latour concludes his
investigation with a list. But we must not forget that this list appears
within a broader narrative, the story of “gathering together” the global
collective with the purpose of addressing the climate problem. In the
end, despite the list, we are offered a unifying narrative. In this story,
the fifteen different kinds of beings the investigation discovers take
on roles and pursue goals beyond their individual networking forms.
This is the story of a global network society. It is the story of a special
kind of science dedicated to gathering together the collective. The social
sciences, indeed, all forms of knowing, are taken to task for what and
how they construct information. The supposedly value-free science of
the Moderns is confronted with an ethical imperative to construct well.
In this new narrative, the role of science in society changes. But it is
not only science that is changing. New roles and new narratives are
appearing throughout society. This is what we will examine below.
Part III
From Systems to Networks –
A Conclusion and a New Beginning
Chapter 4
From Systems to Networks – A Conclusion
and a New Beginning
Latour often speaks of the passing of Western modernity. He speaks of
the closing of a parenthesis; of the end of a historical moment in which
the Moderns not only realize they have never been modern but in which
“society” becomes the “collective.” We propose referring to this histor-
ical, cultural, and social event as the “digital transformation.” Although
this term normally designates the changes an organization goes through
when adopting digital technologies, we use it in a much broader sense
to designate the global, social, and cultural changes brought about by
the technological advances of the 21st century. We are witnessing and
participating in the emergence of a global network society. In saying
that technology changes society, we are not returning to theories of
technological determinism but referring to Latour’s notion of technical
mediation, which we have interpreted in this book as information
construction. When our poor hominin, more than three million years
ago, picked up a stone that became an ax and which transformed him
or her into a hunter or a warrior, neither the hominin nor the stone was
“determining” anything. There was no determination or causation from
the stone or the hominin. Instead, there was a “gathering” of actors into
networks and the construction of information or meaning. Instead of
saying that technology changes society, we should follow Latour, who
says that technology is society.
Luhmann describes a society as differentiated into functional subsys-
tems. For Luhmann, this is a description of modern society precisely
because functional differentiation succeeds stratified and hierarchical
forms of social differentiations that characterize pre-modern societies.
Latour describes a “collective” that is no longer modern Western society
but a networked world, including humans and nonhumans, which,
as we have seen, can no longer be adequately understood in modern
terms. Latour attempts to describe this new, but, as we argued, very
From Systems to Actor-Networks
266
old, collective in terms of modes of existence. Latour’s modes of exis-
tence are in some ways similar to Luhmann’s functional subsystems.
Both Luhmann and Latour are talking about a society in which science,
business, politics, religion, education, etc. characterize social order and
human endeavor. But just as Luhmann’s functionally differentiated
society, so too, Latour’s pluralistic ontology of modes of existence does
not tell us about society as a whole or what kind of society the collective
is becoming. Luhmann leaves us with a society fragmented into autono-
mous, functional subsystems that do not communicate with each other
but somehow exist in a state of mutual adaptation, and Latour leaves
us with a plurality of modes of existence whose crossings and overlap-
pings tend to cause more harm than good. Following our information
theoretical interpretation of Latour, we propose to speak of a global
network society. And we ask, what are this now emerging world’s
guiding concepts, norms, values, and practices?
The global network society can be described neither in terms of
functional subsystems nor as consisting of different modes of exis-
tence, but instead, by reference to the guiding norms of networking.
Networking, or the construction of meaning through translating and
enrolling actors into networks, follows certain norms that guide how
it is to be done well. Construction, as noted above, always involves a
value judgment, the judgment of whether what has been constructed
has been constructed well or badly. Latour argues that a scientific fact,
just as much as a legal judgment, business deal, or work of art, can
be constructed well or badly. This applies to all of Latour’s modes of
existence since all are forms of networking and information construc-
tion. We will argue that by focusing on the norms of good networking
instead of the various modes of existence or functional subsystems of
which society at any time might consist, we can understand society
as a whole and attain a vision of where society is going. Our present
historical moment calls for a vision of a global future. We assume this
is what the entire “ecological” shift and the increasingly urgent call for
global governance is all about. We, therefore, embark upon the admit-
tedly dangerous and uncertain path. Still, it is a path that lies at the
heart of Latour’s notion of what social science today should be about,
From Systems to Networks – A Conclusion and a New Beginning 267
namely, to contribute to gathering together the global collective.
Instead of the functional subsystems that Luhmann assumes, or the
modes of existence Latour describes, we propose focusing on what
could be called “network norms.” Speaking of network norms does not
imply that society, or the collective, no longer consists of organizations,
institutions, or systems such as education, business, science, healthcare,
art, religion, etc. In Luhmann’s view, what are necessary functions for
social order and modes of existence for Latour are all still part of the
global network society, but they are ordered differently. They arise
from the constructive activities of networking that follow, more or less,
the network norms we shall describe below.
We claim that the stone that became an ax in the hand of our ancient
hominin appeared with the emergence of meaning. It appeared for the
first time as something that “subsists” in its own way (what Latour calls
the mode of Reproduction); it appears as something to which activities
can be delegated (Mode: Technology); it appears as something whose
repeated use becomes “second nature” (Mode: Habit), as something
formed of material (Mode: Fiction/Art), as something that can carry
knowledge (Mode: Reference/Science), as something about which all
stakeholders dispute and negotiate (Mode: Politics), as something that
testifies and bears witness to events (Mode: Law), as something to which
many hominins (and humans as well) become passionately attached
(Mode: Economy), as a bearer of scripts and narratives assigning roles to
animals, enemies, wood, etc. (Mode: Organization), as a source of scru-
ples about whether the ax has been constructed well or badly (Mode:
Morality), as a ritual object mediating contact with transcendent beings
(Mode: Religion), and as linked in many unforeseeable ways to more
and more things as time goes by (Mode: Network). The emergence of
the stone ax, which illustrates the emergence of meaning having many
modes, is a “gathering” that happened way back then for the first time
and has been going on ever since. This gathering is what it means to talk
about the emergence of meaning.
Gathering is another word for constructing relations, that is, constructing
From Systems to Actor-Networks
268
information. Meaning is a level of emergent order that integrates both
humans and nonhumans into actor-networks that exist as information,
associations, or relations. Just as our hominin became a warrior or a
hunter associated with the stone ax, we are becoming informational
beings. Floridi (2014) speaks of “Inforgs” to designate humans in asso-
ciation with what could broadly be called digital technologies. Our lives
no longer revolve around stone axes, which was the case for hundreds
of thousands of years, but around computers. It is technical mediation that
makes meaning; this is why technology, in the broadest sense, plays a
major role in how the collective is gathered. The gathering is always tech-
nology, and technology is always a gathering. Discovering and describing
the ways in which the affordances of digital technologies normatively
influence and guide the activities of networking that are going on today
is, therefore, the task at hand. It is the task that ANT has bequeathed us.
It is what is left after sociology has ceased being one of the disciplines of
modern science and become responsible for gathering the collective. It is
what the paradigm shift from systems to networks is all about.
4.1 Digital Transformation
The digital transformation has disrupted Western modernity in at least
three important ways. The first disruption is the posthumanist revision
characteristic of both systems theory and ANT. The posthumanist revi-
sion no longer puts the autonomous rational subject at the center of
history and society. The myth of humanism has become problematic
in many ways, not the least of which is that it seems to have taken its
last stand on the individual. In the digital age, individual freedom is
often interpreted as an issue of privacy. Privacy is an attempt to block
connectivity and the free flow of information characteristic of the
digital transformation and its new norms and values. Privacy is how
the bounded individual attempts to resist being transformed into
information, that is, becoming a being that is no longer defined by indi-
viduality and freedom but by relationality and association.225 Despite
225 For a discussion of the meaning of privacy in ethics and regulation, see Belliger/
Krieger (2018).
From Systems to Networks – A Conclusion and a New Beginning 269
Luhmann’s posthumanism, individuals are considered as systems with
constitutive boundaries. For ANT, on the contrary, networks have no
constitutive boundaries. Instead of privacy, the global network society
values “publicy.” Publicy is the default condition of the informational
self, that is, the form of identity that has emerged from the affordances
of information technologies. The informational self is essentially related,
connected, and open. The informational self is not a bounded indi-
vidual but a network of relations, that is, a being that exists embedded
in actor-networks and, therefore, as information and not as substance.226
The attempt to save the autonomous rational subject of humanism using
privacy as a fundamental and inalienable right derived from an ahistor-
ical human nature fails to account for history and, more importantly,
the relational nature of information. Informational selves cannot be
bounded individuals constituted, as a recent interpretation would have
it, by information that can in no way be shared, communicated, or used
in social interaction.227 The informational self exists in the condition of
publicy and not privacy; that is, it exists as connected to many other
actors, both human and nonhuman, in many different actor-networks.
The disappearance of the autonomous rational subject of humanism
is the first disruption characterizing the digital transformation and
the advent of a global network society. It is the disruption of human
self-understanding, at least for Western modernity.
If the first disruption is the loss of the myth of humanism, which, as
Latour might say, could be called the discovery that we have never been
modern, the second is the loss of an age-old principle of social organiza-
tion, namely, hierarchy. Since the earliest times, we have grown accus-
tomed to organizing cooperative action among large groups by means
of hierarchy. As soon as more than twenty or thirty people get together,
communication’s spatial and temporal constraints require that someone
be the boss, the chief, the king, the leader, the president, etc. As Shirky
(2008) has noted, face-to-face communication does not allow large
groups to coordinate group activity effectively. Someone must put an
226 For a discussion of the informational self see Belliger/Krieger (2018).
227 See Floridi (2005).
From Systems to Actor-Networks
270
end to the discussion and give orders that others carry out. Throughout
history, the much-discussed and not well-defined concept of “power” is
always illustrated by hierarchies of one kind or another. Governmental
bodies, businesses, educational institutions, religious communities,
indeed, every form of cooperative endeavor among people is organized
through power, which seems inevitably to flow from the top down in
the form of command-and-control communication. It is no accident that
the commands of God come from above.
Although for Luhmann (1975), power represents that symbolically
generalized medium that is the basis of the functional subsystem of
politics, he admits that power is constitutive of social order in general
and, therefore, embedded in all communication. All communication, in
one way or another, depends on a functional subsystem organized by
power. The extent to which systems are necessarily hierarchically orga-
nized or what systemic hierarchy means must be answered regarding
the systems theoretical constitutive principles of control and steering.
An autonomous, self-referential, operationally and informationally
closed system must be able to control its operations to maintain its auto-
poiesis. Control and steering is what agency means in systems theory.
Agency manifests itself as autonomous, adaptive behavior. Networks,
on the contrary, are not constructed with regard to any kind of self-ref-
erential operations and the maintenance of setpoints. Consequently,
networks do not need a “controller.” Networks are neither machines
nor organisms. Networks do not operate to maintain the operations
and organization of a limited set of elements. Networks know no
homeostasis. Nonetheless, networks exhibit durability and resistance to
change based on the quantity of links that must be differently translated
and enrolled if the network is to be changed. We are not claiming that
there is no such thing as power in network order. In ANT, power is
usually understood in terms of constructive agency, which is the ability
to be a mediator and construct information through translation and
enrollment processes. Agency is an inherent characteristic of all actors.
Even when actors cease to be mediators and become intermediaries
in black boxes, they remain capable of regaining “power” to change
the network. ANT is a theory of distributed agency, whereas systems
From Systems to Networks – A Conclusion and a New Beginning 271
theory must centralize control within a system and assume an ecological
“balance of power” among structurally coupled autonomous systems.
Even if one does not go so far as ANT when decentralizing power, it is
clear that for modernity and most of Western history, the pyramid could
be seen as the best visualization of power.228 It is a structure that can be
found in the organogram of almost every organization in society. There-
fore, popular resistance to power or attempts to gain power are always
interpreted as bottom-up movements, whereas attempts to maintain
power are always top-down. This assumption underlies what has come
to be known as “critique” in the modern age. Critique is based on the
belief that social communication has always been vertical; orders come
from the top, whereas compliance or resistance comes from below. The
digital transformation disrupts hierarchical power by allowing effec-
tive communication and cooperation not only from one-to-many but
from many-to-many. The affordances of digital technologies encourage
lateral and distributed communication. If social space, as Latour never
tires of saying, is indeed “flat,” then power cannot be hierarchical. As
we will see below, it has become clear in today’s globally networked
society that traditional bureaucratic organizations are becoming
increasingly dysfunctional and uncompetitive. Hierarchies are every-
where being replaced by network organizations based upon distributed
decision-making and self-organization.229
Finally, the third disruption that characterizes the digital transformation
is the transformation of the order of knowledge. Knowledge depends
on media. If what we know cannot be communicated through some
durable medium, it disappears with us and does not become part of
society. Without media of some kind, there is no knowledge. Social and
political revolutions have always accompanied revolutions in media.
The invention of writing replaced oral tradition and changed society.
The invention of the printing press and the electronic mass media
changed society. The invention of digital media is currently changing
228 For Latour (2013), power, at least political power, is best visualized as a circle
and not a pyramid.
229 See Belliger/Krieger (2016) for a discussion of networked organizations.
From Systems to Actor-Networks
272
society. The so-called “new media” have introduced a new order of
knowledge that is non-hierarchical, unlimited, connected, inclusive,
complex, and open to everyone.230 The new media are non-hierarchical
because they give everyone the means of producing and distributing
information. They are unlimited because the cost of producing and
distributing information has drastically been reduced, thus eliminating
the old economy of scarcity in information and knowledge. The new
media are connected and, therefore, inclusive. Nobody, or at least very
few, do not have access to all the information on the Web. It is almost
impossible to block or limit access to information, as all the leaks,
whistleblowers, hacks, and disclosures demonstrate. The new order of
knowledge is complex since there are no longer gatekeepers, authori-
ties, institutions, or trusted sources of knowledge. Knowledge comes
from anywhere, indeed, everywhere, as the acceptance of so-called
“citizen journalism” by mainstream media shows. And finally, the new
order of knowledge is public in a way in which knowledge previously
has never been. The very idea of a “public sphere” arose in the modern
period because of media proliferation and the increasing availability
of information. But the public, in the traditional sense of an arena in
which citizens participate in democratic deliberation, has long become
a global socio-sphere in which politics is only one of many forms of
communication and participation.231
These three disruptions, the dissolution of the bounded individual,
non-hierarchical network organizations, and the new order of knowl-
edge that dismantles traditional roles and functions, are emerging from
the affordances of digital information technologies and are creating a
global network society. This contemporary society, or as Latour would
say, the collective, can no longer be understood or regulated based
on the values, norms, and forms of power typical of modern Western
industrial society. New norms and new forms of organization and
regulation are required if we are to move into the global future.
230 See Weinberger (2012).
231 For a discussion of ANT and new media see Krieger/Belliger (2014).
From Systems to Networks – A Conclusion and a New Beginning 273
4.2 Network Norms
Just as the prehistoric stone ax placed certain constraints on what it could
be used for and what it meant to become a hunter, warrior, or builder,
so do the new information and communications technologies shaping
our world have their own affordances. The symmetrical and distributed
agency ANT proposes obliges us to understand technologies and arti-
facts as social partners contributing in their own ways to what and who
we are and what kind of society we live in. We, therefore, assume that
digital technologies shape and condition what we can do and who we
can become in the actor-networks they allow. Shaping and conditioning
are not to be understood as irresistible forces or determining constraints
but as normative guidelines. Affordances have a normative character.
They influence, suggest, encourage, dispose, and nudge but do not
determine. They guide behavior and experience. It can, therefore, be
said that they condition the kind of beings we are and the society we
live in. Since, as noted above, all construction is subject to the question
of value, of whether what is being constructed is being constructed well
or badly, the guidelines derived from the affordances of digital technol-
ogies can be considered normative. The network norms describe what it
means to construct something well in the digital age. In this sense, one
could also understand them as the normative framework of the global
network society. In other words, if networking is the game, both human
and nonhuman actors are the players, and there are rules in the form
of normative guidelines about how to play the game well. The most
important of these guidelines, which we propose to call network norms,
are listed briefly below.232
4.2.1 Connectivity
Networking tends to connect everything to everything. This is generally
the case, no matter what technology is being used. As soon as the stone
became an ax and the hominin became a hunter or a warrior, the network
232 This list is neither exhaustive nor exclusive. More or other norms, just as with
Latour’s modes of existence, can be described and most probably will be.
From Systems to Actor-Networks
274
began to grow, expand, branch out, and scale up indefinitely. It is of the
nature of meaning, the collective, to gather all things together. Whether
we are wielding stone axes or programming computers, we are doing
the same thing, only on a different scale and with different technolo-
gies. Networking in the digital age makes this inherent tendency of all
networking explicit in a unique way. Indeed, it is probably no accident
that a theory of network order has arisen at this point in history. In a
sense, our technology is telling us who we are.
Generally, network connectivity is precisely the opposite of that which
closed systems do. The “relationing” of elements, which, along with
selection and steering, is one of the foundations of systemic order, is not
the same as connectivity. Where closed systems are characterized by
hierarchy, limitation, exclusion, and reduction, networks are non-hier-
archical, inclusive, connected, complex, and public. On the most funda-
mental theoretical level, this follows from the assumption that Being
is information and information is relation, not substance. To be, in the
perspective of ANT, is to be related. Therefore, the inherent dynamic
of networks is for everything to become ever more related, to gather
more associations, and to integrate more actors in more complex narra-
tive scripts. Whereas systems arise as attempts to reduce complexity,
networks arise to increase the number of associations, relations, links,
and connections between things. Whereas there can be no system of
everything, networks strive to become universally inclusive. Whereas
the basic principle of systemic order is exclusion, the basic principle of
network order is inclusion.
Connectivity is the network norm that accounts for the necessarily
global nature of network society. Luhmann also emphasized the global
nature of the functional subsystems, but only at the cost of excluding
everything that did not fit into the constitutive binary codes of these
systems. Systemic order is indeed universally inclusive, but only
because it excludes everything outside the system’s boundary. The
economic system, for example, extends globally, but only with respect
to things that can be bought or sold. The legal system extends glob-
ally, but only concerning what can be judged as legal or illegal. The
From Systems to Networks – A Conclusion and a New Beginning 275
binary coding of the functional subsystems is an inclusion/exclusion
mechanism. Networks know no such principle of exclusion and are
not coded or organized by such binary distinctions. Networks are not
only not territorially bounded, as is Luhmann’s society, but they are
also not differentiated based on inclusion/exclusion codes. Networks,
of course, do have boundaries, but such boundaries are accidental and
not essential and are, therefore, not constitutive of the network. The
networks’ borders are fuzzy and always provisional, open to renegoti-
ation, constantly changing, porous, and, as we shall see below, flexible.
It is important to note that connectivity is not a weakness but a strength.
It is not a bug but a feature. Let us recall that Paul Barron’s original
concept of a distributed network – which became the basis of the
Internet – was a solution to a security problem. It was the only form of
order that was designed to withstand a nuclear attack. All those who
equate security with impenetrable boundaries of some kind, whether
it be city walls, motes around castles, closely guarded borders, or even
The Great Wall, do not understand that the distributed network is a
security strategy. Contrary to systemic order, which is founded on
drawing clear boundaries, building, so to speak, walls around systems,
networks are permeable and base their strength on scalability, decen-
tralization, openness, and flexibility.
The more extensive the network, the more connections it has, the more
resilient it is. ANT has an interesting and suggestive correlation between
reality and the quantity of associations in a network. The more links a
network has gathered, the more “real” one can say the network is. This
follows directly from the relational ontology of ANT. If Being consists
of relations, the more connections there are in a network, the more
effort it takes to change the program of action of the network, the more
resistance the network offers to change, and thus the more durable and
“real,” it is. When the network is large, there are many more nodes
and links that must be moved. When the network is highly connected,
it requires more effort to divert its trajectory or subvert its program of
action. Concerning the digital transformation, it has become apparent
that the so-called “virtual” world is becoming more real and effective
From Systems to Actor-Networks
276
than what was previously thought to be solid, physical reality.233
4.2.2 Flow
A second network norm can be termed “flow.” Networks allow,
encourage, and predispose information to flow freely through all
the nodes. Flow is not merely passage or movement like water flows
through a pipe. As a network norm, flow is derived from the fact that all
actors in a network are, in principle, not mere intermediaries through
which information moves but mediators who can change how infor-
mation in the network is being constructed. Therefore, this notion of
flow is not to be equated with Latour’s distinction between the setup of
a network and what flows through it. When actors are mediators and
not mere intermediaries, they change not only what flows but the setup
of the network. Flow in the sense of the word that we intend means
that information and everything else, such as people, money, goods,
etc., are always uncontrollably and unpredictably making associations
and configuring and reconfiguring the network. Technical mediation,
processes of translation and enrollment, construct information. Infor-
mation implies that there will always be surprises, there will always
be unforeseen, uncontrolled, unchanneled associations, unexpected
relations, and disruptive connections. Unlike the carefully limited and
controlled communication within a functional subsystem, networks do
not know, cannot know, and cannot control information construction.
No more than our prehistoric hominin could tell in advance what the
stone would tell it about using an ax and becoming a hunter, so we
today cannot in advance know what AI will mean for human existence
and society. Regarding the construction of information, we are only one
player in the game that all beings are playing. As we shall see below,
there is ample evidence that digital technologies have disrupted tradi-
tional forms of information control. Despite untiring calls for regula-
tion and a return to centralization, gatekeeping, censoring, etc., it has
become clear that attempts to control or limit information flows weaken
233 For a discussion of the “mixed reality” of digitally mediated society see Krieger/
Belliger (2014).
From Systems to Networks – A Conclusion and a New Beginning 277
the network or invite workarounds and extensions into other networks
– the darknet is a case in point.234
4.2.3 Participation
As mentioned above, all actors in a network are potential mediators,
that is, agents of information construction. Participation as a network
norm means that all nodes in the network are not mere intermedi-
aries or passageways through which information – or anything else
– moves, but every node also has the ability, and even the duty, to
change, improve, transform, and repurpose information. Every node
in the network is an actor, a source of information, a contributor to the
whole, and not a mere function, a cog in the machine. Participation is
what distinguishes the functional elements of a system from actors in a
network. Participation makes not only the flow but also the content of
information unpredictable and uncontrollable. Participation is agency.
It could be considered as the normative force of Being, or one could say,
the “obligation” to exist, to actively construct meaning. As Shakespeare
put it, it is the question to be or not to be. The answer to this question
is participation.
4.2.4 Transparency
Transparency means that the sources, reliability, and uses of information
are known by all. Nothing is hidden; this is the famous “glass human
being.” Publicy, not privacy, is the default condition today.235 Transpar-
ency, however, must be symmetrical. The network does not condone or
facilitate asymmetrical transparency. The futility of attempts to main-
tain asymmetrical transparency is what all the leaks, hacks, disclosures,
and whistle-blowing teach us. The network norms of connectivity,
flow, and participation oblige that everyone and everything be equally
transparent. Transparency is the opposite of anonymity. Those actors
234 Another case in point is the current proliferation of AI and the many haphazard
aempts by regulators, both private and public, to keep AI under control.
235 For a discussion of “publicy” as the new norm replacing privacy see Belliger/
Krieger (2018).
From Systems to Actor-Networks
278
who depend upon anonymity, such as criminals or state actors, and all
the spies, bad actors, trolls, etc., whatever their interests and whoever
they serve, are contributing to the general insecurity of the network and
not making the world safer. Many of the dangers and harms of digital
technologies, including recently AI, can be traced back to anonymity,
disguise, deception, and strategies to avoid identification and account-
ability. Lack of transparency fosters not only opportunities for misuse of
technologies in all forms but also fosters mistrust. Transparency means
that networks are based on trust. It is no secret that secrecy has always
been a means of power, advantage, inequality, and misuse. The digital
transformation and the advent of a global network society present the
opportunity to develop new forms of regulation other than those arising
from traditional power asymmetries that depend on privileged access
to and control of information.
4.2.5 Authenticity
Authenticity is related to transparency but designates the network
affordance of truth. It is commonplace that networks function based
on trust. In networked organizations, mistrust is dysfunctional. In the
Machiavellian world of bureaucratic hierarchies and informational
scarcity, mistrust is normal and perhaps even necessary for survival.
This is no longer the case in connected, participatory, and transparent
networked organizations. Authenticity takes up but transforms the
modern Western ideal of the free, self-determined individual, the
autonomous, rational subject. Against Western individualism, the
network norm of authenticity does not mean that the individual must
free itself from the pressures of social conformity. Authenticity does not
mean that one must strive to discover one’s true self or that the true
self is an individual. It does not mean one must become the architect
of one’s destiny. It does not subscribe to any notion of human nature,
freedom, or self-determination. To say that networking obliges one to
be authentic means that individuality is experienced as a specific kind
of agency, such as mediating associations, constructing information,
and becoming networked. Authenticity means that one accepts that the
From Systems to Networks – A Conclusion and a New Beginning 279
actor is the network and that agency should always be in the service of
gathering the collective or, in other words, playing the game of meaning
well.
4.2.6 Flexibility
Networks are never stable, even if they may appear so. Unlike closed
systems, open networks do not consist of elements subordinated to
clearly defined functions within a whole. The table system constructs
the legs of a table for one purpose. The system maintains its organi-
zation, integrity, identity, and autopoiesis only so long as all elements
are related to each other in such a way as to ensure that the goals of the
system can be attained. On the contrary, every “actor” in a network
can forge new links, change old ones, assume different roles, and either
expand or retract the network. Networks do not have clear borders that
define what they are and what they do. They do not have fixed functions
and roles, as do the elements of a system. Networks are open to change
and redesign in many directions. They often integrate different identi-
ties and purposes. Flexibility means that innovation and transformation
are more valued than sustainability and fixed programs of action. It is
not what you are but what you can become that guides and motivates
action. As a network norm, flexibility replaces sustainability, a systemic
concept based on maintaining a system’s specific values, constants, or
setpoints. What sustainability means for systems, flexibility means for
networks.
4.3 From Systems to Networks
If the above-listed network norms do influence how the networks that we
live in are constructed, this should be visible. There should be apparent
differences between traditional forms of organizing cooperative action
in society and the forms of organizing emerging based on networking.
The first two parts of this book were dedicated to theory. We now turn
to empirical reality. The question is no longer which scientific paradigm
is more coherent, heuristically powerful, universal, or explanatory. We
From Systems to Actor-Networks
280
now ask: What does society look like if and when systems become
networks? In the following, we will briefly examine some of the more
significant trends that characterize organizing in some of the major
areas of the global network society. These are the areas of business,
education, and healthcare. We assert that the trends becoming apparent
in these areas are also typical for other areas of society. Be that as it
may, these examples should be sufficient to illustrate how networking
is shaping society today. Although the examples we have chosen are
neither exhaustive nor even representative, we hope that the reader
can understand how the network norms briefly described above are
concretely influencing society. They are intended to illustrate, even if
only superficially, how systems today are becoming networks. Leaving
the terrain of pure theory and venturing into empirical description has
solely the purpose of a modest illustration. The discussion below is not
meant to be empirical sociology but merely to show how the digital
transformation is ushering in a global network society.
4.3.1NetworksDoNotDeneFixedRolesandFunctions
Networking in Business
Let us begin by focusing on business. The influence of the network
norms on organizing in the private sector and beyond can be seen in
the many new management practices that disrupt traditional organi-
zational roles and functions. This is the case today, even for not explic-
itly networked organizations.236 A new understanding of leadership
and new employee roles are being implemented in organizations of
all kinds, both production and services. Before command-and-control
communication was standard, employees on all levels now contribute
to and influence decision-making. One speaks of “servant leadership”237
236 An organization can be considered “networked” when informal, non-hierarchi-
cal structures have precedence over traditional bureaucracy, high clustering and
short paths of communication, connectivity and opportunities for random en-
counters, and context steering. See hps://hbr.org/2015/06/what-makes-an-orga-
nization-networked, and Wikipedia hps://en.wikipedia.org/wiki/Network-cen-
tric_organization.
237 See Wikipedia hps://en.wikipedia.org/wiki/Servant_leadership.
From Systems to Networks – A Conclusion and a New Beginning 281
to designate new forms of management that are increasingly replacing
traditional management. Furthermore, employee roles are becoming
more participative and self-directed. Wüthrich et al. (2009) and Kaduk
et al. (2015) document and describe how managers today are called
upon to “break patterns” that have long been defined as best practices.
Notably, they do not attempt to describe networked organizations but
successful management practices in today’s world, regardless of what
kind of organization. The organizations cited as examples of “pattern-
breaking management” are not at all the typical examples cited in the
literature, that is, high-tech and internet companies. These examples
show that breaking traditional roles of industrial management is
imperative for management in all kinds of organizations today, even
traditional businesses.
The typical management patterns in the industrial age were control,
standardization, rational decision-making, optimizing for short-term
gains, efficiency, command and control communication, well-defined
and routine processes, predefined structures, clear organizational
boundaries, risk avoidance, compliance, and clearly defined job specifi-
cations. All these management practices can be modeled in terms of the
selection, relationing, and control functions characteristic of systems.
Indeed, systems management arose almost simultaneously with
cybernetics and has been part of systems science ever since.238 Faced
with the growing complexity and dynamic of today’s VUCA world,
many of these traditional pillars of management wisdom and prac-
tice have dissolved into paradoxes and contradictions.239 Formulated
as paradoxes, these new trends in management could be described
as attempts to steer unsteerability, to mistrust trusted control mecha-
nisms, to standardize diversity and innovation, to shortsightedly see
238 See for example Boulding (1956); Beer (1959); Kast/Rosenzweig (1972). See also
Wikipedia hps://en.wikipedia.org/wiki/Management_cybernetics.
239 There is a vast knowledge management literature which documents these trends.
See for an overview Wikipedia hps://en.wikipedia.org/wiki/Knowledge_man-
agement. VUCA stands for volatility, uncertainty, complexity, and ambiguity
and was coined in 1987 based on the leadership theories of Bennis and Nanus to
describe or to describe the conditions under which organizing in today’s world
must be done. See Wikipedia, hps://en.wikipedia.org/wiki/VUCA.
From Systems to Actor-Networks
282
into the future, to decelerate acceleration, to know the unknown, to risk
risk-tolerance, and to freely choose and vary constraints. Where once
order and stability were valued, now organizations are striving to be
resilient on the edge of chaos. Once command and control communica-
tion, hierarchy, clearly defined processes, and structures were valued,
organizations now attempt to foster knowledge networks, teams, inter-
connectedness, open communication, collaboration, and transparency.
Management is no longer able to rely upon strategies for making things
simple. It is a fundamental principle of systems theory that complexity
cannot be managed with simplicity but only with more complexity.
What we are witnessing today, however, is that at a certain point, the
increase of internal complexity in a system transforms the system into
a network. The increasing internal complexity of many contemporary
organizations at once dissociates and decentralizes teams and sections
and increases their independent contact with the environment. In addi-
tion, changing roles for consumers who, via the widespread adaptation
of “agile” methods for product development and independent access
to information, have become “prosumers” tend to erase the fixed roles
and functions in organizations of all kinds.240
The tendency of organizations to shift from systems to networks is
nowhere more apparent than in the omnipresent trend toward decen-
tralization. Decentralization is a management strategy that has become a
model for managing networked organizations. In response to the chal-
lenges of an ever more complex environment, systems react by building
up more and more internal complexity. This is done by decentralizing
decision-making. Large organizations delegate more autonomy and
independence to smaller working units. This fosters trust, enables
flexibility, brings decision-making closer to the practitioners and the
customers, and motivates employees at all levels to take responsibility
for the whole organization. Traditional management roles are trans-
formed thereby into what is known as “servant leadership.”241 Leaders
240 Agil methods for project management began in the software industry and have
since spread to all sectors. See Wikipedia hps://en.wikipedia.org/wiki/Agile_
software_development.
241 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Servant_leadership
From Systems to Networks – A Conclusion and a New Beginning 283
become more of a support or a coach for the smaller units instead of
using budgeting and controlling to dictate what should be done on
lower levels. Fostering open communication, trust, transparency, and
self-organization are becoming acknowledged and even required
management skills. The end of this process is marked by the transfor-
mation of the closed system into an open network.
Other important pattern-breaking management strategies could be
self-management, cooperative, or bottom-up management. Wüthrich
et al. cite the well-known Orpheus Chamber Orchestra that operates
without a conductor in that it “rotates musical leadership roles for each
work and strives to perform diverse repertoire through collaboration and
open dialogue.”242 Bottom-up decisions about roles and functions are a
typical scenario for Holocracy and Sociocracy, where workers choose
their own roles and can flexibly change positions, and there is no longer
a CEO.243 The guiding principles of this kind of management are to give
power to those who do the work, encourage personal responsibility,
enable flexible role definitions, distribute decision-making, support
cooperation, and promote a culture of collaboration, listening as well as
speaking, valuing consensus, and devoting oneself to the task at hand.
A further example is the Brazilian city of Curitiba, which managed to
deal with many of the typical urban problems, such as unemployment,
health care, education, environmental pollution, transportation bottle-
necks, etc., through encouraging and enabling citizen participation.244
What is significant about Curitiba is that instead of top-down projects,
the city government supported the initiative and self-responsibility of
its citizens to find and implement creative solutions. In this way, plan-
ning and decision-making became a social process enacted in non-hier-
archical, inter-organizational, and networked communication.
A further management strategy that is typical of networking is trust-
242 According to Wikipedia hps://de.wikipedia.org/wiki/Orpheus_Chamber_Or-
chestra.
243 See Wikipedia hps://en.wikipedia.org/wiki/Holacracy; and hps://www.soci-
ocracyforall.org/sociocracy/.
244 See Wikipedia hps://en.wikipedia.org/wiki/Curitiba.
From Systems to Actor-Networks
284
based leadership. Hierarchical organizations are characterized by
mistrust and rely extensively on bureaucratic rules, rigid systems of
control, and discipline. Transparency and open communication are not
characteristics of traditional management. On the contrary, trust-based
leadership enables all co-workers to access information about the entire
organization. Furthermore, all employees are encouraged to participate
in contributing new ideas and participating in making decisions. These
are issues usually discussed under the rubric of organizational culture.
Trust, transparency, participation, open communication, and information
sharing are values and attitudes, not merely processes. They are based on
the network norms of flow, participation, authenticity, and transparency
and have become, in many organizations today, more or less explicit
expectations. Norms, however, should not be confused with standards
and standardization. Traditionally, management was concerned with
setting up standardized processes, job routines, measurement criteria
and methods, qualifications, assessments, incentive programs, and other
instruments to allow for prediction and control.
Wüthrich and Kaduk cite management strategies in human relations
and human development in which the usual laid-out career paths with
indicators and milestones are explicitly avoided. Instead, individual
preferences, personal interests, abilities, an open horizon of possible
paths to take within an organization, self-assessment, and similar
options replace standard qualifications and assessments. Where once
there was uniformity and standardization, now there is an emphasis
on diversity, heterogeneity, and innovation. It has become almost
commonplace in management theory to emphasize the importance of
diversity. Heterogeneous teams, for example, have demonstrated better
problem-solving results than groups with similar backgrounds, experi-
ence, and knowledge. This is a cornerstone of Surowiecki’s Wisdom of
the Crowds. Surowiecki (2004) showed many examples of how heteroge-
neous groups are more effective at problem-solving than experts with
similar backgrounds. In general, it can be argued that the participa-
tory culture of networking fosters diversity because it normalizes the
abnormal and gives equal opportunities for expression to all. In strategic
management, Wüthrich and his colleagues point out the importance of
From Systems to Networks – A Conclusion and a New Beginning 285
flexibility and innovation. Flexibility and an emphasis on innovation
foster and support resilience. The network norm of flexibility allows for
unforeseen goals and avoids fixed setpoints. Organizing in such a way
as to enable many possible goals or programs of action, the free flow
of information, and encouragement of flexibility and innovation is not
only to be found in the business world. Organizing based on openness
to changing roles and functions can also be seen in education.
Networking in Education
The advent of networking as the most important form of constructing
social order has become apparent not only in business but also in
education. The influence of networking has been documented by early
studies from The World Bank (2003). Collis (2005) has noted that in
all nations, education must create a workforce that can “derive local
value from information often in creative ways that go beyond expected
performances,” that is able “to work in multidisciplinary and distrib-
uted teams,” that can “use information technology (IT) for knowledge
management, sharing, and creation,” and that can meet “the need to
‘act autonomously and reflectively; joining and functioning in social
heterogeneous groups’” (Collis 2005: 215). Digital technologies have
brought forth an entire pallet of education tools and methods that are
disrupting traditional educational practices.245 Among others, these are
e-learning, knowledge networks and communities of practice, open
educational resources (OER), learner-centered instruction, mobile
learning, social learning or learning 2.0, workplace learning, MOOCs,
learning analytics, big data in education and AI, personal learning envi-
ronments (PLE), partnerships among educational institutions, flipped
classroom, and connectivism as a new didactic and pedagogical theory.
245 There are many trend reports for education. See for example the Horizon Re-
ports hp://www.nmc.org/publication-type/horizon-report/ , or the OECD Ta-
lis Report hp://www.oecd.org/edu/school/talis.htm , or the European School-
net Brieng Papers hp://www.eun.org/observatory/surveyofschools , or the
Partnership for 21st Century Learning hp://www.p21.org/ , or the Education
section of the Digital Agenda for Europe hp://ec.europa.eu/digital-agenda/en/
ict-education, and in the area of learning and development the Cegos Study 2015
hps://www.integrata.de/leadmin/Bilder/Service/Downloads/2015_CEGOS_
Research_-_LD_Globalization.Final_Report.pdf.
From Systems to Actor-Networks
286
This list is not exhaustive and perhaps not even representative since
new media in education, or technology-assisted education, or whatever
word or phrase is being currently used cannot possibly address the
entire scope of theories, practices, experiments, pilot projects, strategies,
programs, initiatives, etc. that characterize the digital transformation of
education on all levels and in both private and public sectors today. The
administration of educational organizations has much the same tasks
as management in any organization. Organizational development,
personal management, finances, marketing, quality management,
controlling, etc., are similar in most organizations. The affordances of
ICTs in the global network society demand educational decision-makers
to break out of the patterns that have guided education for a hundred
years or more. Decentralization and the practices of self-management,
trust-based management, and bottom-up management are also to be
found in education. One will find many of the new functions and roles
for management that we have already noted for business in educational
administration as well. However, curriculum development, course
administration, assessments, knowledge transfer, and learning evalua-
tion are specific to education.
Applied to education, these norms do not merely change traditional
administrative roles, but instructors and teachers are becoming coaches
who guide the self-directed learning processes of students and trainees.
New forms of curricula such as MOOCs, social and mobile learning,
open educational resources, learning analytics, and personal learning
environments, to name only a few, transform the traditional roles of
both teachers and students. We will discuss these new educational
developments below from the perspective of how networking pushes
educational institutions and organizations beyond conventional bound-
aries and forms of management.
Networking in Healthcare
The digital transformation of healthcare from the perspective of the
disruption of traditional roles and functions can perhaps most clearly
From Systems to Networks – A Conclusion and a New Beginning 287
be seen in what has come to be known as the e-patient movement.246
The e-patient movement is formally organized around the non-profit
Society for Participatory Medicine.247 The e-patient movement defines
itself as “a movement in which networked patients shift from mere
passengers to responsible drivers of their health.”248 In contrast to the
Quantified Self movement, which we will discuss below and operates
mainly in the secondary healthcare sector of fitness and prevention,
the e-patient movement operates within the primary healthcare sector
of diagnosis, therapy, and rehabilitation. A new role for patients and
new relationships between patients and healthcare providers such as
doctors, clinics, hospitals, insurers, regulative agencies, pharmaceu-
tical companies, and the medical technology industry characterizes
networked healthcare in the primary sector. The driving force in this
new constellation of roles and relationships are the patients, who, based
on newly gained access to medical information, increasing demand to
play a more participatory role in healthcare, and connection to other
patients are becoming more active and self-determined. The explicit
goals of the Society for Participatory Medicine are: 1) “to guide patients
and caregivers to be actively engaged in their health and health care
experiences;” 2) “to guide health professional practices where patient
experience and contribution is an integral goal of excellence;” and 3) “to
encourage mutual collaboration among patients, health professionals,
caregivers, and others allowing them to partner in determining care.”249
Examples of networked communities are Patients Like Me,250 Cure-
Together,251 and Acor.252 In these communities, patients are linked to
peers, information throughout the Web, and medical professionals. Not
only are patients connected to each other and healthcare professionals,
but the flow of information, including access to medical research,
246 In general see Wikipedia hps://en.wikipedia.org/wiki/E-patient.
247 hp://participatorymedicine.org/.
248 hp://participatorymedicine.org/about/.
249 hp://participatorymedicine.org/about/.
250 hps://www.patientslikeme.com/.
251 hp://curetogether.com/.
252 hp://www.acor.org/.
From Systems to Actor-Networks
288
makes it possible for some patients to become better informed about
their specific diseases than many doctors. Another example of changing
roles and functions is that patients can publicly rate the performance
of medical service providers. CureTogether.com, for example, offers
access to millions of ratings not only of doctors and hospitals but also
treatments and medications.253 Regarding participation, e-patients
can donate their data to medical research, whereas doctors share data
via such programs as “Open Notes.”254 An example of data donation
that the US government supports is the Blue Button Movement. The
Blue Botton makes it easy for patients to access their medical records.255
Another example is 23andMe, a web-based personal genome service,
which allows individuals to get information about their genetic dispo-
sition for certain diseases and donate this data to research.256 Echoing
the explicit goal of the Society for Participatory Medicine in the USA,
“letting patients help,” the EU Action Plan 2012-2020 for using digital
solutions in healthcare has the slogan: “Putting patients in the driving
seat: A digital future for healthcare.” The traditional healthcare system
did not allow, encourage, and enable patients to play an active and
constructive role in medicine. The shift from systems to networks
in healthcare changes this situation radically. Healthcare networks
empower patients to responsibly and effectively use their “health
literacy” and new networked technologies in ways that were utterly
unforeseen in traditional healthcare.257
A further example of changing roles in medicine is the practice of
“shared decision-making.” In shared decision-making, the interaction
between doctors and patients becomes cooperation that is based on
access to shared information and the explicit acceptance of equal deci-
sion-making rights. Diagnosis and therapy cease to be a one-way street
253 hp://curetogether.com/.
254 hp://www.opennotes.org/.
255 hps://www.healthit.gov/patients-families/join-blue-buon-movement.
256 hps://www.23andme.com/en-int/.
257 “Health literacy is the ability to obtain, read, understand and use healthcare in-
formation to make appropriate health decisions and follow instructions for treat-
ment.” hps://en.wikipedia.org/wiki/Health_literacy. See Bessiere et al. (2010)
for a discussion of the positive eects of networked health.
From Systems to Networks – A Conclusion and a New Beginning 289
and become a collaborative endeavor in which doctors and patients
work together as a team. Networking allows patients to gain access to
information, medical records, peer communities, and other resources
that were not available before networked forms of healthcare arose.
Furthermore, e-patients can now contribute to research by donating
data and collaborating with scientists on agenda-setting for clinical
studies and research programs.
These examples show that connectivity, flow, communication, and
participation have enabled normal people to access, understand, and use
health-related information in new and unforeseen ways. This changes
patients’ expectations toward healthcare professionals and institutions,
as well as the possibilities that these established players have in dealing
with their “customers” and “clientele.” Patients and caregivers can now
easily obtain information on all aspects of their health concerns, commu-
nicate, and get advice and support from others in similar situations and
healthcare professionals. Patients can participate constructively in diag-
nosis, therapy, and rehabilitation. As in other areas of society, health-
care is becoming a part of the “sharing economy.”258 Challenges for all
those who take on new roles in healthcare lie in managing the networks
that are emerging from the digital transformation in such a way that
costs are lowered, the effectiveness of interventions and prevention is
increased, communication between doctors and patients is improved,
and extended, medicine becomes more personalized, and transparency
and trust replace hierarchical relationships. In short, one could say that
the democratization of healthcare is an effect of networking.
4.3.2 Networks Have No Constitutive Boundaries
Business
One of the most significant changes that is taking place because of
the digital transformation is the breaking down of institutional and
organizational boundaries. In comparison with systems, networks do
not have constitutive boundaries. All of the functional subsystems of
258 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Sharing_economy.
From Systems to Actor-Networks
290
society that Luhmann describes are constituted as systems by a specific
binary coding such that communications can be ascribed to one system
or another and processed accordingly. Business, for example, is coded
by buying/not-buying (paying/not-paying). If you go to the market and
don’t buy anything, there has been no financial transaction; therefore,
whatever you do at the market is not business. The same is true for
education, which is coded by the binary difference of certification/
non-certification. If you are an autodidact without any means or inten-
tion of certifying your knowledge and skills, the learning you are doing
is not part of the educational system. Healthcare is also a functional
subsystem of society, constituted by the difference between sick/well. If
you are well, you do not participate in the healthcare system. No insur-
ance will cover the costs of your wellness, and no doctor will treat you
for it. When systems become networks, these constitutive boundaries
become porous and diffuse.
Networked organizations are constituted not by binary rules of inclu-
sion/exclusion and clear boundaries but by attempts to establish and
maintain various cooperative relations. If one looks at the problem
of competitive advantage in a global market facing many companies
today, closed systems are not well-positioned for success. Lavie (2008)
notes that the “emergence of interconnected firms is a contemporary
phenomenon” (325) rooted in the economic realities of the global
network society. To explain why networked firms have competitive
advantages, he introduces the notion of “network resources” (325).
“Network resources reside in alliances in which the interconnected
firm is involved rather than within the scope of the firm’s organiza-
tional boundaries” (2008:325). Alliances serve various goals. They allow
“flexible and more rapid adjustment to changing market conditions,”
“reduce time to market in response to shortened product life cycles,”
“assist in bridging national boundaries,” and “reduce market uncer-
tainty and stabilize the firm’s competitive environment by forming
norms of reciprocity that establish commitment and regulate exchange
transactions” (325). Furthermore, alliances are not restricted to any one
branch or sector. They have proven efficient in research and devel-
opment, production, and marketing “in almost any industry” (326).
From Systems to Networks – A Conclusion and a New Beginning 291
As Saz-Carranza and colleagues pointed out, and for Lavie as well,
the networked firm is involved in associations that are “dynamically
evolving yet durable” (2008: 326).
Whereas traditional theories of the firm emphasize interfirm competi-
tion, theories of networked organizations emphasize alliances, partner-
ships, strategies for managing associations, inter-organizational gover-
nance structures, organizational learning, or knowledge management,
and they focus on the achievements of the entire network. This new
networked understanding of organizing departs from traditional views
that see the appropriation of value as exclusively based upon owner-
ship or disposal over resources such as capital, raw materials, and labor.
Against this conventional view, Lavie points out that for networked
organizations, there is a “direct sharing of resources” and an “indirect
transferability of benefits associated with these resources” (328). He
concludes that the “proprietary assumption of the resource-based view
prevents an accurate evaluation of an interconnected firm’s competitive
advantage” (328). Networked forms of organization can create and
appropriate value more efficiently and competitively because of their
ability “to maintain valuable interaction with their partners” (332) and
less because of their inner-organizational assets. Lavie concludes:
At the turn of the 21st century, alliances have emerged as a
primary vehicle for conducting economic transactions. Once
conceived of as independent entities, firms are now consid-
ered interconnected in the sense that they engage in multiple
alliances with counterpart firms. This phenomenon necessi-
tates a new theory of the firm that incorporates simultaneous
competition and collaboration as drivers of value creation and
appropriation. […] Managers need to pay attention not only
to the value of resources nurtured by their own firms but also
to the resources possessed by their firms’ partners. They must
evaluate the value of network resources as well as the value
of potential combinations of network resources and internal
resources. They must collaborate to create value and compete
to appropriate their relative share of that value. They must look
From Systems to Actor-Networks
292
beyond the immediate scope of alliances and seek to leverage
network resources while finding the right balance between
resource sharing and the protection of proprietary assets. (Lavie
2008:333)
Understanding the advantages of networked organizing from the point
of view of transcending traditional boundaries and cooperatively using
resources also explains the emergence of global projects as an important
way networked organizations operate today. Ainamo (2008) claims
that “a new ‘logic of organizing’ has been spreading in market econ-
omies” (482). This new organizational logic is based on “flattening out
organizational hierarchies, weakening of firms’ boundaries in favor of
networks of collaborations, and restructuring of competition between
firms within and across industries” (482). This has led to “global proj-
ects as new organizational forms” (482). Ainamo notes that “more and
more companies are now becoming structured around distinctly global
projects” (483).
Organizing highly skilled workers dealing with complex
problems to create novel outputs by integrating varied forms
of expertise – also across cultures and institutional systems –
represents a significant new organizational form. (483)
Barney (1991:101) has defined resources as “all assets, capabilities,
organizational processes, firm attributes, information, knowledge, etc.,
controlled by the firm.” According to Lavie (2008), sharing resources
adds value and competitive advantage. From this perspective, global
projects can be seen as a characteristic phenomenon of the global
network society. Ainamo comments upon this situation by claiming
that in today’s world, “firms are transforming into fluid, overlapping
organizational arrangements in which both internal and external
boundaries break down, and cooperation with other firms grow”
(482). Cooperative activities between organizations can take many
forms. Instead of alliances or partnerships, projects have emerged as a
common form of inter-organizational cooperation. Castells (2005) also
emphasizes the project-based dynamic of networking:
From Systems to Networks – A Conclusion and a New Beginning 293
[…] large corporations, and their ancillary networks, engage in
strategic partnerships on various projects concerning products,
processes, markets, functions, resources, each one of these proj-
ects being specific, and thus building a specific network around
such a project, so that at the end of the project, the network
dissolves and its components form other networks around
other projects. (Castells 2005:9)
Projects, of course, are not organizations. They are usually distinguished
from organizations in that they are directed to specific goals to be
attained within a limited time and with limited resources. Nonetheless,
projects must be managed by organizations. Typical project manage-
ment tasks are initiating, planning, executing, controlling, monitoring,
and closing the project.259 Even though such activities usually take
place within organizations, in many projects from industries such as
the film industry, the software industry, the construction industry,
biotech, research consortia, and consulting and marketing, it is often the
case that experts from outside any one organization or from different
organizations are explicitly contracted for certain projects. They work
together on specific tasks whose completion also ends the contractual
cooperation.260 Ainamo has identified two ways organizations extend
their borders to form projects. Some organizations can operate as agen-
cies specialized in carrying out projects within a specific area for other
organizations. The experts on these projects remain within the agency
but move from one project to another or even work simultaneously
on several projects in various organizations. Organizations can also
cooperate or contract with independent external experts. The project is
a heterogeneous or hybrid form of organization that consists of flexible,
ad hoc teams that are free to regroup or move on to other projects once
any single project is completed. Along with such global networked
projects, there comes a transformation of what work means and what
259 See the Guide to the Project Management Body of Knowledge for a represen-
tative statement hps://en.wikipedia.org/wiki/Project_Management_Body_of_
Knowledge.
260 See the discussion of various project-based industries in Ainamo (2008).
From Systems to Actor-Networks
294
demands are being made upon labor.261
It is an often-noted characteristic of labor in the knowledge society that
knowledge workers, who traditionally were regularly employed by one
organization, become independent contractors with the companies they
once worked for and other companies, including competitors. Castells
(2000: 12) distinguishes between “self-programmable” and “generic”
labor. According to Castells, “self-programmable labor is equipped
with the ability to retrain itself and adapt to new tasks, processes,
and sources of information, as technology, demand, and management
speed up their rate of change” (12). Labor is no longer an organizational
role or something that only has its place within the boundaries of an
organization. Knowledge workers are experts who often make up the
flexible and heterogeneous project teams responsible for business activ-
ities in many areas today.262 What project forms of organizing show is
that flexibility with regard to organizational boundaries has become
imperative for both employers and employees. The major advantage
of project-based organizing is adapting and reconfiguring operations,
participants, goals, and processes quickly and globally. Of course, not
everyone is a knowledge worker and finds themselves able to profit
from the freedom implied by variable and open networks. For those
whom Castells puts into the category of “generic” labor and those who
fall out of any form of employability, new networked forms of educa-
tion and civil society can help alleviate many problems.
Another important example of networks going beyond the boundaries of
traditional organizations is platforms. Everyone knows how Microsoft,
in the early years, took over the PC Market from Apple. Apple’s prod-
ucts were locked into a proprietary pipeline, a one-sided value chain
in which consumers were offered products from a producer. Microsoft
changed the rules of the market by making their operating system open
261 We will return to this aspect of the network society when discussing challenges
for education below.
262 Castells (2000) speaks of the “individualization of labour” as the key transfor-
mative factor and characterizes this workforce as, “part-time work, temporary
work, self-employment, work by contract, informal or semi-formal labour ar-
rangements, and relentless occupational mobility” (12).
From Systems to Networks – A Conclusion and a New Beginning 295
for all to use as a platform upon which independent producers could
offer their products, whether hardware or software. Microsoft quickly
dominated the PC market, and Apple was on the brink of bankruptcy.
Apple, of course, learned the lesson. When the iPhone came onto the
market, Nokia, Motorola, and other manufacturers of telephones were
dominant. The iPhone, however, was not another telephone; it was a
platform that networked app developers and consumers like Micro-
soft did with Windows. Google soon followed with Android. Today,
there are other examples of platforms such as Airbnb, eBay, Uber, and
Amazon, not to mention the many innovative companies from China.
These new forms of doing business are networked markets. Like tradi-
tional markets, they bring buyers and sellers together. But platforms
bring many sellers and buyers in such a way together that all profit
from synergy, growth, innovation, and exploration of mutual opportu-
nities. This extraordinary dynamic is called “network effects.”263
Various kinds of network effects have been documented. One such
effect of networking occurs when an increase in product usage leads
to a rise in product value. For example, the more people who use an
iPhone, the more app developers are interested in creating apps for the
iPhone, and the more valuable the iPhone becomes for users. Facebook
is another example. The more people use Facebook, the more people it
attracts since everyone is there. This is a positive feedback loop and is
called a direct network effect.264 There are also indirect network effects,
such as when the increased usage of a product leads to innovation and
production of complementary or associated products. Again, Microsoft
is a good example insofar as the Windows platform, just as with Google
Android later, led to the development of many compatible products.
Another network effect can be called “two-sided” because the devel-
oper/producer and the consumer gain the value. The iPhone becomes
more valuable for consumers because it creates value for developers.
Uber or Airbnb can be seen as platforms with two-sided network effects.
263 See Liebowi hp://wwwpub.utdallas.edu/~liebowit/palgrave/network.html
and the discussion and list of literature by Arun Sundararajan on network eects
see hp://oz.stern.nyu.edu/io/network.html
264 See the classic work by Ka and Shapiro (1985; 1994).
From Systems to Actor-Networks
296
Uber creates value for both drivers and passengers. It can even be seen
as a “multi-sided” platform because it generates opportunities for new
services such as the delivery of food, doctors’ visits, etc.265
Parker et al. (2016) have recently summarized some of the disruptions
of traditional organizations introduced by platforms. In contrast to
traditional organizations, platforms do not produce and market partic-
ular goods and services. They enable an entire “ecosystem” consisting
not only of producers but also prosumers who are connected so that
all gain value from the free flow of information, open communication,
low barriers to participation, transparency, authenticity, and flexibility.
Whatever products and services are being offered, platforms are orga-
nized as networks and not as markets. They focus on the synergies,
cooperations, connectivity, and the optimal uses of generated data.
Data is a resource that is not primarily based on ownership but on use.
Sharing data in networks creates value for all participants. The only
thing that needs to be organized and managed are the transactions.
The participants contribute resources themselves. The major advantage
of the network is the connectivity, flow, and participation of all stake-
holders.
Education
Educational institutions are also benefiting from partnerships and
collaborative communities of practice that go beyond traditional
boundaries regarding problems and opportunities in the areas of infra-
structure and pedagogy. Schools and educational institutions on all
levels, such as primary or secondary schools, colleges, and universities,
are connecting and building networks. They enter into network-type
cooperations and communities of practice to exchange knowledge and
best practices. This is taking place in both administration and teaching.
Schools cooperate to develop better teaching materials, infrastructural
solutions, and similar “products.” In addition to sharing knowledge
and developing new solutions, educational institutions enter into part-
265 See the many articles on platforms and network economics by Economides
hp://www.stern.nyu.edu/networks/.
From Systems to Networks – A Conclusion and a New Beginning 297
nerships to carry out projects to distribute and implement ideas and
solutions beyond the participants in the network. Finally, such educa-
tional networks work with organizations such as businesses, museums,
cultural organizations, etc., beyond the boundaries of their communi-
ties, cities, and regions. The goal is to make an entire community or
region more attractive and competitive.266
Partnerships, networks, and new forms of inter-organizational learning,
such as MOOCs, offer significant management challenges. The same can
be said of Open Educational Resources (OER). We will discuss MOOCs
below. Let us turn to Open Educational Resources (OER), which can be
understood to illustrate the network norm of participation. New media
studies (Jenkins et al. 2005) speak of “participatory culture.” Forms of
participatory culture are communities of practice, collaborative prob-
lem-solving, user-generated content, crowdsourcing, social networking,
and many similar networked forms of collaborative work. In education,
participation fosters peer-to-peer learning and a very untraditional
attitude toward intellectual property. All participants produce informa-
tion, which is placed in a commons (often under a Creative Commons
license) so all can use it. Open Educational Resources are defined as
[…] teaching, learning, and research resources that reside in
the public domain or have been released under an intellectual
property license that permits their free use and re-purposing by
others. Open educational resources include full courses, course
materials, modules, textbooks, streaming videos, tests, software,
and any other tools, materials, or techniques used to support
access to knowledge. (Open Education Resources 2013)267
Properly connecting learners and learning resources is one of the most
important tasks of educational management and teaching. As Jones
(2015) points out, the affordances of ICTs “have the potential to change
the relationship between learners and the institutions of learning to
266 For a typology of school networks see Muijs et al. (2011:37.).
267 See also hps://en.wikipedia.org/wiki/Open_educational_resources; and Jones
(2015).
From Systems to Actor-Networks
298
their resources.” This potential, however, can only be realized when it is
“institutionally enacted” (120). This means that educational institutions
or content providers “make their materials freely available online for
anyone to use” (121). The usual way to do this is to license texts, images,
videos, etc., under the Creative Commons license, which puts content
into the public domain that does not stand under the usual copyright
restrictions. The OECD defines such freely accessible educational mate-
rials as
[…] digitized materials offered freely and openly for educa-
tors, students, and self-learners to use and reuse for teaching,
learning, and research. OER includes learning content, software
tools to develop, use, and distribute content, and implementa-
tion resources such as open licenses. (OECD 2007:29)
It is at this point that OER and MOOCs meet. A MOOC (Massive Open
Online Course) is a form of networked learning that, in many ways, epit-
omizes the global network society in the area of education. MOOCs are
forms of teaching and learning that encourage inter-organizational part-
nerships and user-generated content. MOOCs can be seen as an embod-
iment of a “connectivist” learning theory.268 Connectivism is a theory of
learning that goes beyond the psychological orientation of traditional
learning theories such as behaviorism, cognitivism, and constructivism
by claiming that knowing is not only a social activity but is distributed
in networks. Connectivism shifts learning away from individuals and
organizations and emphasizes the network as the primary subject and
object of learning. Learning happens when information moves in all
directions over networks rather than in the heads of individuals who are
then required to prove knowledge acquisition by testing. MOOCs often
use Open Educational Resources (OER), emphasizing user-generated
content and participatory information production. MOOCs are educa-
tional platforms since they orchestrate resources, focus on interaction
and sharing, and aim to improve the learning community.
268 See Siemens (2005), Downes (2005, 2012), AlDahdouh et al. (2015) for discussions
of connectivism. See also for an overview Wikipedia hps://en.wikipedia.org/
wiki/Connectivism.
From Systems to Networks – A Conclusion and a New Beginning 299
There are different forms of MOOCs. Siemens and Downes base
MOOCs on social learning, participation, user-generated content, and
connectivism and find them in corporate training and normal certifica-
tion programs of universities worldwide. Important consortia of insti-
tutions are Udacity, Coursera, edX, FUN, Iversity, and others. These
MOOCs require educational management based on the open flow of
information in networks.269 Setting up and managing a MOOC requires
that educational administration moves away from the traditional forms
of information control that in the past have made it possible to design
curricula, perform standard assessments, and carry out acknowledged
certification.
Another example of how networking moves education beyond tradi-
tional boundaries is Personal Learning Environments (PLE). The idea
of a PLE arose as an alternative to the centrally designed and top-down
managed learning management systems (LMS), which often reflected
institutional forms of organizing. In one form or another, LMS have
become standard teaching tools in most educational institutions. Orig-
inally designed as a “virtual” classroom or virtual school, they carried
the same centralism into the digital realm, top-down administrative
procedures, teacher-centered pedagogical practices, and standard
curricula and assessments typical of traditional schools.270 However, the
affordances of digital media and the tendency toward networking miti-
gate against conventional ways of organizing education. Consequently,
new forms of delivering education that encouraged learner control,
diversity of practices, and self-directed learning and communication
were sought. The learners themselves and not the educational institu-
tions should be making decisions about how and even what they learn,
about processes, content, and information resources. The PLE also
reflects the conviction that learning is a lifelong endeavor, extending
beyond any institution, program of studies, or career qualification
and including informal and formal learning, schooling, and on-the-job
269 See the contribution in the edutechwiki for a review of literature hp://edutech-
wiki.unige.ch/en/E-learning_2.0.
270 For literature see the curated list of publications on PLE by Ilona Buchem hps://
ibuchem.wordpress.com/ple/.
From Systems to Actor-Networks
300
training. However, lifelong learning is only practicable when learners
themselves have something to say about the educational processes.
Educational institutions can no longer assume that one-size-fits-all and
that standardized instruction and assessment are the way to guarantee
equality as well as quality in learning.
PLEs exemplify networking as opposed to systems in that they allow
individuals access to a wide range of information resources, educational
services, peer communities, aggregators, publication and distribution
tools, such as blogs and wikis, and content management systems, such
as e-portfolios. The PLE goes beyond any educational institution in that
it constitutes a kind of hub for learner-centered, lifelong networked
learning, both in regular schooling and in professional training, both
formal and informal, both public and private. The PLE is multipur-
pose, heterogeneous, integrative, and adaptive. It is not an educational
organization but a learning network no longer bound to a particular
organization, institution, or degree program. It is learner-centered,
open to many different applications, adaptable to other purposes and
situations, and principally without limitations coming from external
organizational constraints. The advent of PLEs has important conse-
quences for educators and training administrators at all levels. Speaking
for educational professionals and administrators, Leone writes, “The
introduction of lifelong learning objectives and policies has poised us
on the threshold of major change in education and society” (Leone
2010:30). Therefore, there is a “need for greater emphasis to be put
upon flexibility, transferability, individualization, modularization, and
mobility in education” (31). It has become a task for educational leaders
to change from a system to a network.
Healthcare
If we look at healthcare, we can see how networks are coming into being
that are not characterized by clear constitutive boundaries. Examples
of this can be seen in the Quantified Self Movement and the various
Platform or Health Ecosystem initiatives. One of the most interesting
recent developments in networked healthcare, e-health, or connected
From Systems to Networks – A Conclusion and a New Beginning 301
health is the so-called “Quantified Self” (QS) movement.271 Quantified
self designates a broad range of new tracking technologies, devices,
apps, services, practices, and values that all aim at “quantifying” health
data: collecting, aggregating, evaluating, sharing, and commercializing
personal health and wellness-related data. Other terms for quantified
self are life-logging, body tracking, self-tracking, auto analytics, body
hacking, and self-surveillance. A well-known slogan of the movement
is “self-knowledge through numbers.”272 It is important to empha-
size that this trend to track, quantify, and evaluate personal data is
not limited to healthcare. Similar technologies and practices can be
found in other areas. We have already discussed applications of this
kind in education, for example, personal learning environments and
learning analytics. The broader concept for personal data collection and
analytics is “personal informatics.” Personal informatics has become a
general term describing many different kinds of digital self-monitoring
concerning the body, health, and wellness, as well as other areas of life
such as work, hobbies, and education. The official Personal Informatics
website defines the term as follows:
Personal informatics is a class of tools that help people collect
personally relevant information for the purpose of self-re-
flection and self-monitoring. These tools help people gain
self-knowledge about one’s behaviors, habits, and thoughts.
(www.personalinformatics.org)
Personal informatics includes not only the tracking of vital functions
but also the monitoring of emotional and psychological states such as
stress, happiness, sleep, and relaxation. Furthermore, there are many
different technologies and services that can be classified as personal
informatics, which relate to productivity at work, consumer behavior,
monitoring learning processes, financial activities, hobbies, sports activ-
ities, etc. Within this broad understanding of self-monitoring, many
different kinds of networks and organizations are involved. The data
271 See hps://en.wikipedia.org/wiki/Quantied_Self as well as the ocial website
hp://quantiedself.com/.
272 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Quantied_self.
From Systems to Actor-Networks
302
that is gathered enters into and creates networks. For example, individ-
uals are networked with providers, developers, employers, social and
professional communities, educational institutions, businesses, govern-
ment, healthcare organizations, and patient communities.
In these networks, individuals are not integrated into closed systems.
They become administrators of their own information and make inde-
pendent decisions about how this information is to be used. Personal
informatics, therefore, offers many new opportunities, dangers, and
challenges for decision-makers in organizations that are interested in
using this data in different ways. Personal informatics is also closely
linked to developments such as the Internet of Things, Ambient
Assisted Living, smart buildings, smart neighborhoods, and smart
cities. Making homes and cities smart is not simply building networked
infrastructures but transforming closed systems into networked orga-
nizations. Smart government, smart education, smart healthcare, and
smart organizations of all kinds relate to each other in ways that go
beyond the mere adaptive behavior of systems.273 All of these networks
are connected to each other. No smart cities are without smart govern-
ments, smart education, smart businesses, etc. These classic domains
of society can only become smart when they connect to each other and
form new kinds of networks. Organizing the global network society
necessarily crosses boundaries and links activities that were tradi-
tionally isolated in different domains, departments, jurisdictions, and
social systems. From this perspective, the quantified self movement has
many similarities to other societal changes. In business, education, and
healthcare, networks cross institutional boundaries and branch out into
many different organizational fields.274
Technically the quantified self movement relies on smartphones,
smart watches, armbands, smart scales, wearables, smart clothes, and
273 We see an example of this in the discussion of the Brazilian City of Curitiba.
274 The digitalization of healthcare goes, of course, beyond QS and includes tele-
health, telemedicine, medical e-learning, electronic patient records, hospital
and clinical information systems, health data analytics, e-patient communities,
shared decision making, and much more. For an overview see Belliger/Krieger
(2014b),
From Systems to Networks – A Conclusion and a New Beginning 303
implants, also referred to as “insideables,” including injected or swal-
lowed microchips and similar technologies. These devices are equipped
with sensors that measure chemical, magnetic, mechanical, visual,
acoustic, and thermal changes in the body or near environment. The
data is registered and transmitted, usually via wireless or NFC technol-
ogies, to a smartphone or computer, which then evaluates it and sends
it to the Internet, service providers, portals, community websites, etc.
In this process, specially designed software or apps for mobile devices
play an important role. Hundreds of thousands of health-related apps
are available in the major app stores maintained by Google, Apple, and
Microsoft.275 These devices and apps continually monitor vital functions
such as blood pressure, heart rate, calorie consumption, sleep quality,
movement, glucose levels, stress, emotional states, and posture, as well
as environmental influences such as UV radiation, air quality, etc. The
evaluation, interpretation, and presentation of this data are done by
many different services that the device or app producers usually offer.
Some typical services are the aggregation of data, visualization of data,
evaluation, comparison, sharing, recommendations and tips, health-re-
lated advice, e-consulting, referral to consultation by medical profes-
sionals and peers in patient communities, and personalized advertising
by sellers of health-related products.
Self-tracking can become part of diagnosis, therapy, and rehabilitation
processes. At this point, doctors, clinics, hospitals, laboratories, health
insurance companies, and regulatory agencies also become involved
in the network.276 The same blurring of traditional boundaries can
be observed in the areas of fitness, well-being, and prevention. Not
only individuals are involved but also health insurers, public health
programs, and prevention-related organizations, including employers
concerned with having a healthy workforce. Quantified self is a new kind
of network involving individuals across all demographic distinctions
and many different kinds of organizations both within and without the
traditional healthcare sector. Quantified self is changing the traditional
275 Lutpon (2014a: 608) found over 100,000 health and medical apps mid 2014 in the
Apple App Store and Google Play.
276 See Lupton (2014c) for a discussion of medical apps and usage practices.
From Systems to Actor-Networks
304
landscape of healthcare by blurring the boundary between therapy and
prevention and empowering individuals to gather and evaluate their
medical data, providing new data for medical and clinical research,
enabling citizen science, and effectively disrupting the traditional, hier-
archical healthcare system.
The dismantling of closed system boundaries is also clearly visible
in the tendency to establish healthcare platforms. One can speak of a
significant trend in healthcare: moving away from isolated, individual
healthcare providers to networks of providers and patients linked
together over platforms.277 Benedict at al. (2018: 241) surveyed twen-
ty-three eHealth platforms in Europe and concluded, “Within eHealth
infrastructures we observe a predominant role of eHealth -platforms.”
These platforms bring together many different healthcare providers
and offer customers various services in the areas of information, insur-
ance, diagnosis, therapy, and prevention. Platforms usually consist of
four different “roles.” There is first and second the platform provider
and or sponsor who set up and maintain the digital infrastructure.
Third, the providers of products and services link together over the
platform and establish a market for healthcare consumers. Fourth,
there are the healthcare consumers, who are often prosumers in that
they can sometimes actively participate in developing and testing
products and services. The platform mediates transactions between all
parties. Providers use the synergies of the cooperation via the platform
to optimize business plans, avoid unfruitful competition, gather and
use data, and maintain customer contact. The major advantages of
eHealth platforms over the traditional healthcare system are efficiency,
better quality of care, simple usability for all participants, accessibility
for many different stakeholders, and empowerment of health services
consumers. In addition, one could add that the greater volume and
quality of the data generated by the platform enables the development
of better services and new business models. Platforms, therefore, help
reduce the time and effort required to access and manage health infor-
mation, streamline administrative tasks, and thus reduce costs in many
277 For a recent studies on eHealth platforms see Benedict et al. (2018); and Bernardo
et al. (2022).
From Systems to Networks – A Conclusion and a New Beginning 305
areas. Platforms are an integral part of the digital transformation of
healthcare and a clear example of how networks are breaking out of the
boundaries imposed by closed systems and even replacing systems in
today’s world.
4.3.3 Networks Cannot be Controlled Top-Down
Business
Systems are based on selection, relationing, and steering or control
processes. No cybernetic system can exist without a “controller” of
some kind. It is the reference to a controller that gives cybernetics its
name. The kybernetes was the helmsman of a ship. Another word for
the controller is governor. Turning to living systems, although there
are some exceptions, the “central nervous system” is an excellent
example of how living systems most efficiently steer their processes
and operations. The same is true on the level of meaning. For centu-
ries, if not thousands of years, centralized or hierarchical control has
been a primary form of establishing and maintaining social order. It
is hard to find any society that does not have a “political” function.
The digital transformation and the advent of a global network society
place this honored tradition in question. However, this is not a new
form of anarchy. As the complexity of the world increases, hierarchies
everywhere are becoming less efficient. In a complex, quickly changing
environment, centralized decision-making brings many problems.
These are 1) long and slow communication channels, 2) all the disad-
vantages of bureaucracy which hinder quick responses to problems and
innovative solutions, 3) inflexibility regarding changing demands from
within and outside the organization, 4) inefficient use of knowledge
and information, 5) tendency to disempower people at lower levels of
the organization, 6) lack of diversity and emphasis on standardization,
7) fostering of a culture of conformity, 8) mistrust and power struggles,
8) inability to manage remote work and dynamic teams which hinders
collaboration, knowledge sharing, and collective decision-making.
Given these well-known problems facing traditional organizations,
From Systems to Actor-Networks
306
much effort has been made in the last decades to describe the challenges
of “21st-century management.”278 A central topic that has emerged from
this research is the position and function of management in networked
organizations.279 Although explicitly concerned with inter-organiza-
tional networks, the framework of network leadership and management
developed by Saz-Carranza et al. (2008) can be applied to networks of
all kinds. As often remarked (Powel 1990), networks appear as a form of
order somewhere between the diversity and heterogeneity of markets
and the unity of hierarchical, bureaucratic organizations. Viewing
networks as a middle ground between markets and hierarchies with
characteristics drawn from both may seem helpful. However, this view
is still bound to the dualistic assumptions of agency and structure, or
individual and organization typical of traditional organization theory.
To say that networks are neither markets nor hierarchies does not neces-
sarily imply that they are some mixture of the two, a middle ground
between the opposing forces of agency and structure.
On the contrary, networks are neither markets, hierarchies, or a mixture
of the two. They are something entirely different. This does not mean,
however, that the task of managing networks successfully has nothing
to do with forces of diversity and unity. Saz-Carranza et al. locate
network management problems in the “inherent paradoxes implied
by networks, in particular, the need to be simultaneously united and
diverse” (2008: 292). This implies that “successful networks are simul-
taneously united and diverse” (297). Instead of finding a normative
basis in contracts and property rights, as do markets or employer/
employee relations, as in hierarchies, networks are based on comple-
mentarity, mutual adjustment, and reciprocity. They emerge and act
within a “collaborative context” (Chrislip/Larson 1994) characterized
by the specific “boundary-crossing nature of collaborations […], the
lack of formal authority and hierarchy, and the blurriness of strate-
gies” (Saz-Carranza et al. 2008:296). A collaborative context demands
278 See the classic work of Peter Drucker (1999) and the extensive handbook by Wan-
kel (2008).
279 See Powel (1990), Jarrillo (1993), Huxham/Vangen (2000), Kickert et al. (1997),
Koppenjan/Klijn (2004), and Saz-Carranza et al. (2008).
From Systems to Networks – A Conclusion and a New Beginning 307
different leadership and management practices than those established
in the industrial age.
Leadership in collaborative contexts must be necessarily
different, focusing largely on process, and has similarities to
facilitative, transformative, and servant leadership – that is, to
inspire commitment and action, to lead as peer problem solver,
to build broad-based involvement, and to sustain hope and
participation. (Saz-Carranza et al. 2008:296)
Citing Huxham and Vangen (2000), Saz-Carranza and colleagues define
leadership broadly as “mechanisms that make things happen in a collab-
oration” (2008:296). Contrary to much research on leadership, which
focuses on individuals conceived of as either leaders or followers, that
is, on their skills, roles, competencies, talents, or authority, Saz-Car-
ranza et al. propose focusing on “activities and actions that make things
happen” (296). Therefore, they do not need to draw a clear boundary
between management and leadership since bureaucratic routines and
authoritative decisions no longer play important roles. Furthermore,
the framework they propose does not focus on leadership and manage-
ment within networks, that is, within network nodes or network part-
ners, but instead describes leadership and management of the entire
network. Leadership is a “collective phenomenon” (297) that distributes
among all participants the task of “generating unity among the network
members while preserving their diversity” (297). Enabling networks to
be at once diverse and unified gives them an advantage over traditional
forms of organization. Diversity fosters innovation, adaptability, flexi-
bility, and resilience, while unity allows the various partners to coordi-
nate action and to perform cooperatively and effectively.
In sum, the unity/diversity paradox is inherent to networks,
which must be diverse to have an added value with respect
to hierarchies but united to allow for concerted action of any
kind, unlike markets. Leadership is about avoiding diversity
and unity to undermine each other by respectively generating
disunity and similarity. (297)
From Systems to Actor-Networks
308
Based on this understanding of the task that network management
must accomplish, Saz-Carranza et al. develop a framework according
to which network management practices fall into four categories: facili-
tating, framing, activating, and mobilizing.280
Facilitating includes all those management practices and activities that
enable and support interaction, equality, motivation, and participation
among network members. It is not merely a matter of exchange of infor-
mation but of inclusive and participatory decision-making. Otherwise,
certain members may feel disadvantaged, and the network would tend
to disintegrate. Facilitating is a management practice derived from the
network norms of connectivity, flow, and communication. These norms
guide networking toward inclusivity while maintaining diversity,
heterogeneity, open communication, and participatory action.
Framing is the second category of management practices Saz-Carranza
and colleagues identify as peculiar to networked organizations. As the
name suggests, framing includes all activities that enable cooperative
action with regard to specific goals. This term strongly recalls Goffman’s
staging and Weick’s sensemaking. It is also closely related to Latour’s
notion of “localizing” (Latour 1996:234) since framing sets up struc-
tures, rules, boundaries, etc., that focus, channel, partition, and reduce
cooperative action. Framing consists of those management practices
that operate not primarily for the network as a whole but for decentral-
ized units, nodes, and partners, who can manage their unique contribu-
tions to the entire network in this way. Without framing, the diversity
of the network would not be able to become operative since individual
units would not have sufficient freedom and decision-making powers
to organize their respective activities. Framing can be understood as
guided by the network norms of participation and authenticity. Each
partner can and should be responsible to the network in ways that they
are themselves able to determine, at least partly. This demands manage-
ment practices that make participation possible without unnecessarily
limiting participants’ authentic expression.
280 Saz-Carranza et al. (2008: 297.).
From Systems to Networks – A Conclusion and a New Beginning 309
Activating is the third form of management practice found in networks.
It refers to all management practices that seek to integrate actors into
the network. Activating is similar to what ANT calls translation and
enrollment. The network becomes more powerful and effective to the
extent that it can enroll new members who contribute new skills, compe-
tencies, ideas, and impulses. Just as translating and enrolling in ANT,
activating in the network management framework of Saz-Carranza and
colleagues is concerned with network goals or programs of action into
which members can be integrated, and thus, it is also concerned with
network identity, reputation, and legitimacy. The network norms of
connectivity, participation, authenticity, and transparency play roles in
successful activation since it is only when all participants in the network
are accepted for their unique contributions and information is available
to all that the network as a whole can become a collective actor capable
of cooperative action concerning consensual goals.
Finally, mobilization refers to management practices that can bring
resources and support into the network. The term “mobilization” in
the Saz-Carranza framework means much the same as it does for ANT.
For ANT, actor-networks “mobilize” as many human and non-human
actors as possible to create stability, durability, social presence, and
power. Translating and enrolling actors into a network becomes mobi-
lizing when the goal is to generate network power. For Saz-Carranza
and colleagues, “Mobilizing essentially builds network power […]”
(298). Summarizing the four aspects of network management, they
write:
Sustaining the unity/diversity paradox generates power, which
increases the network’s effectiveness. The networks sustain
the unity/diversity paradox by activating members, facilitating
interaction, and framing the structure (procedures, rules, and
values). In addition, networks mobilize support. When acti-
vating, the network selects and attracts members who share
certain experiences, values, and principles but who are diverse
regarding other organizational characteristics. Interaction and
open decision making among the diverse members must be
From Systems to Actor-Networks
310
facilitated, and members united by framing common proce-
dures, rules, and values. (2008:298-299)
This framework for understanding network management practices
distinguishes itself from traditional management models because
network management is not about decision-making but facilitating,
framing, activating, and mobilizing decision-making, which no longer lies
exclusively at the upper levels of the hierarchy. Decisions are distrib-
uted throughout the network and are made by network members.
Management does not reduce complexity or uncertainty but establishes
procedures for openness and inclusiveness. Furthermore, manage-
ment does not control outcomes but facilitates processes. It does this
through what Weick (1995) would undoubtedly call sensemaking. In
words similar to how Weick describes sensemaking, Saz-Carranza et al.
understand framing as what “ascribes, interprets, and makes meaning
of the outside reality; it constructs the perceptions of actors regarding
reality” (299).
The description of network management practices in facilitating,
framing, activating, and mobilizing enables actors to “go through a
naming and framing process, which is, basically, meaning making” (299).
As we have shown in detail above, networking is a matter of translating,
enrolling, and mobilizing actors into networks guided by programs of
action and concerned with allowing all actors to have their own voice.
In the global network society characterized by the affordances of new
digital technologies, management practices of facilitating, framing, acti-
vating, and mobilizing can be derived from the norms of connectivity,
flow, communication, participation, transparency, authenticity, and
flexibility. It would be a mistake to view networking and networked
forms of organizing as idealistic or ideologically motivated activism
that ignores the realities of economics and market constraints. On the
contrary, networked forms of organizing in all areas, whether public or
private, profit or non-profit, can be said to have a competitive advan-
tage over traditional organizations. Lavie (2008:324) notes that “firms
can no longer be considered simply as independent entities competing
for favorable market positions and protecting their core assets from
From Systems to Networks – A Conclusion and a New Beginning 311
imitation and appropriation.” The global network society creates a situ-
ation where “firms have become interconnected in the sense that they
engage in multiple simultaneous alliances” (324), contributing directly
to economic success.
The fact that networks, as opposed to systems, are not steered effec-
tively top-down is usually discussed in terms of flexibility. However,
speaking of organizations as “flexible” often occurs within the frame-
work offered by evolutionary theory and the theory of complex adap-
tive systems. Organizations are compared to biological organisms that
can maintain their operations in the face of environmental changes by
building up internal complexity. The more complex a system becomes,
the more possible states it has, and thus, the more options it has to
react to environmental changes. Since it became apparent that bureau-
cratic, hierarchical organization does not favor internal complexity,
management theories first looked to systems theory for a model of how
organizations can foster innovation. This model is adaptive behavior
and learning, and a typical example of this theoretical development is
Stafford Beer’s Viable Systems Model of management.281 Viable organi-
zations tolerate internal complexity to adapt to a complex environment.
Herbert Simon, for example, speaks of “nearly decomposable systems”
(1996:193) that consist of relatively independent specialized elements.
“The claim is that the potential for rapid evolution exists in any
complex system that consists of a set of stable subsystems, each oper-
ating nearly independently of the detailed processes going on within
the other subsystems” (193). It is well known that systems can only
increase internal complexity to a certain extent or otherwise risk losing
the central control that they need to pursue fixed goals and coordinate
operations. The systems model has two drawbacks in this context. First,
there is the issue of whether models of adaptive behavior do not lead
to goals of sustainability, which are necessarily conservative and do
not foster innovation. Second, we have noted above that the increasing
complexity of systems tends to transform them into networks. From
the point of view of network theory, there are no inherent limitations
281 For an overview see Wikipedia hps://en.wikipedia.org/wiki/Viable_system_
theory.
From Systems to Actor-Networks
312
on internal complexity. In networks, flexibility does not need to be
achieved by special measures to increase internal complexity since
networks do not arise as a complexity reduction in the first place. Flex-
ibility is a characteristic of networks as such.
Networks are structurally opposed to hierarchies because every node
in the network has a certain freedom to enter into connections to other
nodes, bring new nodes into the network, and branch out into new
directions with new goals. Of course, nodes in a network can be black-
boxed into functional units with fixed input/output mechanisms. The
prevalence of such black boxes everywhere is indisputable. This is why
the systems model appears so plausible and convincing. Functional
networks are needed and convenient for many reasons: efficiency,
predictability, and control. Before the digital transformation, limitations
on communication made it necessary to construct black boxes within
organizations and sharp boundaries between organizations. Black
boxes, however, can be opened, and any element of a functional unit
can become a mediator again and enter into new and unforeseen associ-
ations. The affordances of digital technologies tend to open black boxes
and deconstruct system boundaries. This characteristic of network
structure makes complexity and flexibility into “normal” attributes of
networks. Networks do not need to be made flexible by extraordinary
measures. They are, by nature, flexible. What from the systems point
of view is a complex and challenging task is for networks a matter of
the normal dynamic of connectivity, flow, and participation. The more
connectivity a network has, the more nodes it has and the more infor-
mation flows through the nodes in unforeseeable ways, leading to more
participation and unexpected developments, that is, to innovation and
change.
Innovation from the systems theory point of view is a response to
environmental changes that perturb the system but do not “inform”
the system. Innovation for systems is always adaptation. Information
construction is internal to the system. In a certain sense, it can be said
that environmental changes are always threats to the system. If the
system does not construct information adequately, it ceases to be viable.
From Systems to Networks – A Conclusion and a New Beginning 313
System viability depends on being able to adapt quickly to environ-
mental changes. Much of systems-based management theory is devoted
to understanding how systems can become more adaptable, which is
today called “resilience.”282 Risk management and innovation have
become central topics in today’s business world. Innovation, however,
can be conceived as a synergy of organization and environment, a
mutual and interdependent exchange.283 In this view, both the organi-
zation and its environment mutually construct each other, which, we
would argue, is another way of speaking about networks. Networks do
not have an “environment” in the specific sense that complex adaptive
systems do. Networks do not “adapt” to an environment in the same
way that biological systems do. This has consequences not only for how
innovation and organizational change are theorized but also for what
has been called “organizational learning.”284
Education
Trust-based management, decentralization, leveling out hierarchies,
and involving stakeholders in decision-making are new forms of orga-
nizing that also have an important role in education. In administration,
teaching, and learning, education is a cooperative effort in which
everyone plays active roles. Therefore, it is not surprising that school
administrators are also decentralizing, turning more to self-manage-
ment, bottom-up management, and collaborative management to profit
from experiences and resources available beyond the reach of one single
school or group of instructors and teachers.285 Outcomes of partnerships
and networks are often uncertain and cannot be easily incorporated
into budgets and quality control processes that can be centrally steered.
282 A recent McKinsey survey underlines the importance that the term as gained in
the business world. See hps://www.mckinsey.com/capabilities/risk-and-resil-
ience/our-insights/from-risk-management-to-strategic-resilience
283 “In the 21st century, organizations should not solely respond to preordained en-
vironmental conditions, but should instead inuence and actually create their
environment by innovation” (Kollman/Stöckmann 2008:12).
284 For an overview of research on organizational learning and knowledge manage-
ment, see Easterby-Smith/Lyles (2011).
285 See for example the discussion on school partnerships in Muijs et al. (2011).
From Systems to Actor-Networks
314
Accountability, as crucial as it is, above all in the public sector where
much education is located, must be balanced against innovation and
risk-taking for long-term improvements in teaching, development of
learning materials, learning success, assessment, and curricula reform.
These processes show that partnerships and networking can make
resources and infrastructures accessible so that information is distrib-
uted to all participants and instruments and procedures are set up to
transfer knowledge throughout and beyond the network.
One of the most important trends in the use of digital information
today in all areas is so-called “big data.” Big data refers to the avail-
ability of very large data sets (volume) from diverse sources and in
many different formats (variety), which can be “analyzed” quickly,
if not in real-time (velocity), by special analytic software.286 Analytics
aims to discover trends and correlations in large datasets. Only when
data becomes big in volume, variety, and velocity, do correlations
and patterns that can otherwise not be found become apparent. This
information can be useful for descriptive, predictive, preventive, and
prescriptive analytics. These new possibilities of using data affect two
significant areas in educational management. These are networked
learning and the digitalization of educational administration. One
speaks of “learning analytics” in both areas to describe all aspects of big
data in education. Jones (2015) speaks of learner analytics to identify
“those aspects of learning analytics that focus on the learner or student
as opposed to the institutional or business aspects of analytics” (114).
We have spoken above about how personal learning environments and
MOOCs extend teaching and learning beyond the borders of traditional
institutions.
Learning analytics can illustrate how top-down management in educa-
tion is becoming inefficient and questionable. Learning analytics appli-
cations aim to predict and identify which students need support and
allow instructors to intervene to optimize learning outcomes.287 It has
286 See hps://en.wikipedia.org/wiki/Big_data for an overview.
287 See the discussion and examples of applications in EduFutures.net hp://edfu-
tures.net/Learning_Analytics.
From Systems to Networks – A Conclusion and a New Beginning 315
become possible to track not only online activities within a traditional
LMS but also to monitor social media usage, search activities on the
Internet, emails, and physical activities such as using the library, sports,
or entertainment facilities on and off campus. This kind of tracking of
student activities, as with other big data applications outside the educa-
tional context, creates an entirely new situation. The learner becomes
a “glass human being” who is completely transparent to those with
access to the data and analytic applications. This results in a situation
in which administrators are forced to include all stakeholders in deci-
sion-making. Jones (2015) cites the Director of Teaching and Learning at
the Open University UK, who voiced his concern as follows:
With these possibilities come dangers that the data could be
used in ways undesirable to students. These include invading
their privacy, exploiting them commercially by selling their
data to third parties or targeted marketing of further educa-
tional products. (120)
In the area of enterprise management of educational institutions, the
same problems occur. Analytics allows for a kind of educational “busi-
ness intelligence” in which not only students are being monitored but
teaching staff, non-teaching employees, and other operations as well.
Big data is used for ERP systems to monitor not only libraries, labo-
ratories, and materials suppliers for teaching and learning, but also
customer relations, housing, parking, catering, athletics, etc. There are
no technical barriers to data use in all areas, leading to the problem
that business imperatives could be applied to teaching and learning. As
Jones points out,
If a broad administrative view of analytics is taken, then
analytics can be just another management and administra-
tive tool that will shift the balance in institutions away from
academic and pedagogic concerns towards market and business
concerns with measurable performance and value for money.
Measurements of quality, progression and dropout when seen
in terms of students as units of resource are quite different from
From Systems to Actor-Networks
316
measures of quality understood in terms of human develop-
ment and citizenship. (Jones 2015:115)
The challenge to education management coming from big data and
learning analytics is to break the patterns of traditional business
management and adopt new symmetrical transparency regarding the
uses of data. In the long run, analytics cannot succeed as a top-down
administrative process. All stakeholders should be involved in deci-
sions about what data will be collected and analyzed for what purpose,
and all stakeholders should have equal access to the results of analytics.
This implies that these results be made understandable and useable by
all stakeholders, including learners. The unavoidable transparency of
all users, whether students or staff, should be compensated by symmet-
rical transparency on the side of administration, decision-makers,
and institutional leaders. Transforming top-down transparency into
bottom-up transparency can only be successful based on open commu-
nication and trust. Jones (116) discusses the framework developed by
De Laat in which users are themselves responsible for “tagging prob-
lems that are part of their learning,” “adding people they collaborate
with,” and “linking themselves with problems others have already
described.” Networking transforms education into a self-organizing
process and forces a revision of traditional hierarchical management
practices. Within a framework such as this, learning analytics is actively
implemented by users so that more and better information is collected,
leading to more reliable and more valuable results. Instead of being
monitored behind their backs, as it were, all participants actively
contribute to data aggregation and consent to data use.
Healthcare
It is well known that healthcare is one of society’s most heavily regu-
lated sectors. It is also a well-established tradition that doctors are “gods
in white” who control all health-related information about patients and
often subscribe to a paternalistic model in which “the doctor knows
best,” and only the doctor decides what information to share with
patients. There is hardly any other subsystem of society that is more
From Systems to Networks – A Conclusion and a New Beginning 317
characterized by hierarchy than the healthcare system. We have already
discussed how the digital transformation and the shift from systems
to networks are changing healthcare. E-Patients are now everywhere
acknowledged and often welcomed. Medical information is now avail-
able to patients as never before. Patients are networked with other
patients and medical professionals via platforms and communities.
The quantified self movement, which is more active in prevention than
primary care, also shows how networking and digital technologies
empower people to take control of their health and well-being instead
of relying on medical professionals. These developments challenge,
disrupt, and transform the traditional hierarchical healthcare organi-
zation. We have discussed many of these developments above and will
limit our discussion to only a few examples.
As an example of a bottom-up healthcare organization, the Dutch
Buurtzorg model for patient-centered care is often cited and discussed
as an exemplary solution to home-based healthcare.288 The Dutch
word Buurtzorg means “neighborhood care” and focuses on nursing
and community health, not hospital treatment. It was founded in 2007
by a Dutch nurse, Jos de Blok, and has since become internationally
recognized for its innovative and practical approach to delivering
home healthcare services. The major characteristics of the Buurtzog
model are self-managing teams of nurses consisting of ten to twelve,
who are responsible for providing a wide range of healthcare services
to patients that are located within a particular area. The teams come to
know the patients personally and care not only for their physical health
but also for the emotional, psychological, and social well-being of their
patients. The emphasis is on the overall quality of life and not merely on
whether someone can be classified as either sick or well. This concept
of encompassing care assists patients in becoming more self-sufficient
and responsible regarding their health. For this reason, much effort is
devoted to personalized care, that is, working closely with patients and
supporting family members to develop approaches to care that fit the
personal needs of patients and their families.
288 See hps://www.buurorg.com/about-us/buurorgmodel/.
From Systems to Actor-Networks
318
With regard to administration, Buurtzorg strives to reduce costs and
bureaucracy. The aim is to allow nurses to spend more time with patients
and less on paperwork, thus enhancing the quality of care. Further-
more, the teams emphasize preventive measures and health education
to help patients maintain their health and well-being, reducing the need
for more intensive interventions in the future. The nurses work collab-
oratively with other healthcare professionals and social workers and
use available community resources to ensure comprehensive and coor-
dinated care for their patients. Another important characteristic of this
bottom-up approach to healthcare is increased flexibility in scheduling
and care delivery. Nurses can decide how they manage their sched-
ules to accommodate patients’ needs and provide care when it’s most
convenient and effective.
Illustrating the digital transformation’s values, Buurtzorg utilizes
technologies such as a custom-built electronic health record system,
digital support for communication and coordination, and electronic
documentation. These measures empower the nurses to make decisions
about patient care, how to manage the teams, and how to set up opera-
tional processes of their work. From this empowerment and autonomy,
which one does not have in hierarchical organizations, the nurses share
a sense of ownership and engagement lacking in traditional top-down
organizations. It has been shown that the model effectively improves
patient outcomes, reduces hospital admissions, and enhances patient
and staff satisfaction. It is based on patient-centered care, decentralized
decision-making, and community engagement.
Similar to Buurtzorg, many Community Health Worker programs have
been initiated all over the world. These programs train and deploy indi-
viduals from within local communities to provide needed healthcare
services and health education to people who would otherwise not have
easy access to these services. Some examples, which are only mentioned
here, are: the BRAC Manoshi in Bangladesh employs female commu-
nity health workers to provide maternal and child healthcare services,
family planning education, and health counseling to women and fami-
lies in urban slums; the Treatment and Care for Everyone from South
From Systems to Networks – A Conclusion and a New Beginning 319
Africa which is a community-based HIV prevention and care program;
the Accredited Social Health Activists (ASHAs) program which is part
of India’s National Rural Health Mission; the Promotoras de Salud in the
USA and Latin America, which employs community health promoters to
improve healthcare in Latin American countries and Hispanic commu-
nities in the USA; the Mitanin Program in Chhattisgarh, India, which
trains women from tribal and rural areas as community health workers;
the Community Health Extension Workers in Nigeria who are trained
to provide healthcare services such as maternal and child health, immu-
nizations, family planning, and health education in communities. There
are many more examples of this kind that could be cited here. The point
is that there are clear tendencies to support bottom-up healthcare initia-
tives and thus shift decision-making to lower levels of the hierarchy.
Summary
We have attempted to give some concrete examples of the shift from
systems to networks in contemporary society. It is admitted that the
above discussion is neither exhaustive nor representative of how
networking is changing society. And, of course, there can be no question
that the examples we have described in business, education, and health-
care in any way “prove” the thesis that we are currently going through
a paradigm change in our understanding of the world and how we
should organize our lives. Nonetheless, we hope to have convincingly
shown that something is indeed changing, that what can plausibly be
described as networked forms of social order are being adopted, and
that they represent a significant departure from what could be called
modern Western industrial society. This is the society that Luhmann’s
theory of social systems claimed to have theoretically modeled. Modern
society, according to Luhmann, is a functionally differentiated society.
Despite Luhmann’s posthumanism and acceptance of globalization,
systems theory is a theory, at least on the level of the social sciences, of
Western modernity. Differently, Latour’s actor-network theory is also a
theory, or rather, an ethnology of the “Moderns.” Latour, however, is
convinced, and this is what he attempts to show in his empirical studies,
that modernity has come to an end and that the task of thinking, of
theory, of science has now become the task of assembling the collective.
We have called this new collective the global network society and have
proposed in this book to have gone beyond Latour in formulating the
values that guide networking. Unlike Latour, who does not address the
question of the digital transformation, we have derived the values of
the global network society from the affordances of digital technologies.
After all, if society, as Latour says, is technology, it makes a difference
for society what that technology is. The digital transformation is not
only a technological change but a transformation of society. When
society changes, values change. How are these changes to be theoret-
ically modeled? What concepts, foundational assumptions, and future
perspectives are needed to move constructively into the global future,
wherein digital technologies, among them AI, will play a major role
Summary 321
in how we live our lives, understand ourselves, and construct social
order? These were the questions that motivated this book, and which
have guided our discussion. We hope they are the right questions and
that we have offered answers that at least are a small contribution to
answering them.
Literature
Adams, K. M., Hester, P. T., Bradley, J. M. (2013): A Historical Perspective
of Systems Theory. In: Proceedings of the 2013 Industrial and Systems
Engineering Research Conference, A. Krishnamurthy and W.K.V. Chan,
eds.
Ainamo. A. (2008): “Global Projects as New Organizational Form,” in:
Wankel 2008, Vol. 1, 482-489.
AlDahdouh, A. A., Osorio, A. J., Caires, S. (2015): “Understanding Knowl-
edge Network Learning and Connectivism.” in: International Journal
of Instruction Technology and Distance Learning, Oct. Vol. 12. No. 10,
3-46.
Andersen, B.P., Miller, M., Vervaeke, J. (2022): Predictive processing and
relevance realization: exploring convergent solutions to the frame
problem. Phenom Cogn Sci.
Apel, K.-O. (19760): Transformation der Philosophie, 1. vol., Frankfurt:
Suhrkamp.
Apel, K.-O. (1976): Transformation der Philosophie, 2. vol., Frankfurt:
Suhrkamp.
Apel, K.-O. (1988): Understanding and Explanation: A Transcenden-
tal-Pragmatic Perspective, Boston: MIT Press.
Arsenault, A. (2011): “Networks: The Technological and the Social,” in:
Dealnty, G./Turner, S. P. (eds) Routledge International Handbook of
Contemporary Social and Political Theory, 259-269. UK: Routledge.
Asbhy W.R. (1956): An Introduction to Cybernetics. London: Chapman &
Hall.
Autio, T., Ropo, E. (2005): “Networking as a Strategy for Restructuring
Basic and Further Teacher Education,” in: Wiel, V., O’Hair, M. J. (eds)
Network Learning for Educational Change. UK: Open University Press.
Baecker, D. (ed) (1993a): Problems of Form, Frankfurt: Suhrkamp.
Baecker, D. (1993b): Calculus of Form, Frankfurt: Suhrkamp.
Barabasi, L.-A. (2002): Linked. The New Science of Networks. Perseus
Publishing: Cambridge MA.
Barabasi, L.-A. (2016): Network Science. Cambridge University Press.
Barney, J. (1991): “Firm Resources and Sustained Competitive Advantage,”
Literature... 323
Journal of Management, 17(1), 99-117.
Bateson, G. (1972): Steps to an Ecology of Mind, New York: Ballentine.
Bateson, G. (1979): Mind and Nature, A Necessary Unity. New York:
Dutton.
Beer, S. (1959): Cybernetics and Management, English University Press.
Belliger, A. (2008): “Wissensmanagement – ein Instrument auf dem Weg
zur Lernenden Organisation,” in: Schweizerische Zeitschrift für Heil-
pädgogik 9.
Belliger, A. (2010a): “Enterprise Social Networking. Kollaboration im
Unternehmen 2.0: Organisation oder Netzwerk?,” in: Wissensmanage-
ment. Zeitschrift für Führungskräfte, 5/2010.
Belliger, A. (2010b): “Human Resource Management in Zeiten von Face-
book, Twitter, iPhone & Co. Von Unternehmen als Netzwerken,
Kontextsteuerung und Grenzmanagement,” in: Risikofaktor Personal.
Wie viel Kontrolle ist nötig. Management von Unsicherheit. Zeitschrift
HR Today.
Belliger, A. (2010c): “eLearning und Wissensmanagement: Chancen und
Herausforderungen von elektronischen Lern-Hilfsmitteln für Nutzer
und Anbieter; wie eLearning effizient eingesetzt und genutzt werden
kann,” in: Organisator. Das Magazin für KMU, 7-8/2010.
Belliger, A. (2010d): “Die Akteur-Netzwerk-Theorie – oder: Wie sozial sind
soziale Netzwerke?” in: Wissensmanagement. Zeitschrift für Führung-
skräfte, 6/2010
Belliger, A. (2011a): “Lernen und Lehren mit Technologien aus der Sicht
der Akteur-Netzwerk-Theorie,” in: Lehrbuch für Lernen und Lehren
mit Technologien (L3T), TU Graz.
Belliger, A. (2011b): Social Media und Wissensnetzwerke in der Perso-
nalentwicklung, in: Jahrbuch Personalentwicklung 2012, Luchterhand.
Belliger, A. (2016): Digitale Transformation im Gesundheitswesen, in:
Tagungsband zum 5. Nationalen Biobanken-Symposium der Technol-
ogie– und Methodenplattform für die vernetzte medizinische Forschung
TMF, Berlin, Bd. 14 Schriftenreihe der TMF.
Belliger, A. (2017): Connected Health – Connected Architecture, in:
Tagungsband 6. Symposium Health Care der Zukunft – Livability of
Health, Technical University Berlin.
From Systems to Actor-Networks
324
Belliger, A. (2013a): “Werte und Normen vernetzter Organisationen oder
Digitale Geschäftsprozesse sind nur der Anfang,” in: Wirtschaftsmag-
azin 1/2013.
Belliger, A. (2013b): “Social Media und Wissensnetzwerke in der Perso-
nalentwicklung, in: Handbuch eLearning,” Wolters Kluwer Verlag
(gemeinsam mit D. Krieger), 6.25, 1-10.
Belliger, A. (2014): “Vernetzte Gesundheit,“ in: Belliger/Krieger (eds)
Gesundheit 2.0. Das ePatienten-Handbuch. Bielfelfd: Transcipt, 97-135.
Belliger, A., Krieger, D. J. (eds) (2006): ANThology: Ein einführendes Hand-
buch zur Akteur-Netzwerk-Theorie, transcript Verlag: Bielefeld,
Belliger, A., Krieger, D. J. (2008): “Netzwerk,” in: Lexikon Soziologie und
Sozialtheorie. Hundert Grundbegriffe, Stuttgart: Reclam.
Belliger, A., Krieger, D. J. (2009a): “ANT: Nichts Neues? Nichts Nützliches?
Eine Replik zu Gesa Lindemann,” in: Zeitschrift für Kulturwissen-
schaften.
Belliger, A., Krieger, D. J. (2009b): „Virtuelle Grenzen und Räume in Organ-
isationen,“ in: G. Thomann, G./Bucher, Th/Hagmann, Th./Kuhn, R.
(eds) Grenzmanagement, Resonanz Bd. 1,71-84, Bern: hep-verlag.
Belliger, A., Krieger, D. J. (eds) (2014): Gesundheit 2.0. Das ePatienten
Handbuch, transcript Verlag: Bielefeld.
Belliger, A., Krieger, D. J. (2014a): Interpreting Networks – Hermeneutics,
Actor-Network Theory & New Media, transcript Verlag: Bielefeld.
Belliger, A., Krieger, D. J. (2014b): “Netzwerke der Dinge (aus der Sicht
der Akteur-Netzwerk-Theorie),” in: Samida, M. K. H./Eggert, H. P/
Hahn, S. (eds) Handbuch Materielle Kultur. Bedeutungen, Konzepte,
Disziplinen, J.B. Metzler.
Belliger, A., Krieger, D. J. (2015): Die Selbstquantifizierung als Ritual virtu-
alisierter Körperlichkeit, in: Körper und Ritual. Sozial- und kulturwis-
senschaftliche Zugänge und Analysen, Gugutzer, R., Staack, M. (eds),
Springer: Wiesbaden, 389–404.
Belliger, A., Krieger, D. J. (2016a): From Quantified Self to Qualified Self –
A Fictional Dialogue at the Mall, in: Quantified Selves Statistic Bodies,
Digital Culture & Society Journal. P. Abend, M. Fuchs (eds), transcript
Verlag: Bielefeld, S. 25–40.
Belliger, A., Krieger, D. J. (2016b): Organizing Networks – An Actor-Net-
work Theory of Organizations, transcript Verlag: Bielefeld.
Literature... 325
Belliger, A., Krieger, D. J. (2016c): The End of Media – Reconstructing
Media Studies on the Basis of Actor-Network Theory, in: Applying the
Actor-Network Theory in Media Studies, B. Ochsner, M. Spöhrer (Eds.),
IGI Global, Hershey, PA, S. 20–37 http://www.igi-global.com/book/
applying-actor-network-theory-media/147546
Belliger, A., Krieger, D. J. (2018): Network Publicy Governance – On Privacy
and the Informational Self, transcript Verlag: Bielefeld.
Belliger, A., Krieger, D. J. (2021): Hacking Digital Ethics, Anthem Press. UK.
Belliger, A., Krieger, D. J. (2022): New Directions in Digital Ethics, in
Casas-Roma, J., Conesa, J., Caballé, S. (eds) Technology Users and Uses
– Ethics and Human Interaction Through Technology and AI, Ethics
International Press: Cambridge UK.
Benedict M, Herrmann H, Esswein W. (2018): eHealth-Platforms – The
Case of Europe. Stud Health Technol Inform. 247:241-245.
Berger, P. L. and T. Luckmann (1966), The Social Construction of Reality:
A Treatise in the Sociology of Knowledge, Garden City, NY: Anchor
Books.
Bernardo, J.; Apóstolo, J.; Loureiro, R.; Santana, E.; Yaylagul, N.K.; Dantas,
C.; Ventura, F.; Duque, F.M.; Jøranson, N.; Zechner, M.; et al. (2022):
eHealth Platforms to Promote Autonomous Life and Active Aging.
Bertalanffy, L. v. (1968): General Systems Theory, New York: George
Braziller.
Bijker, W. E. (1995): Of Bicycles, Bakelites and Bulbs. Toward a Theory of
Sociotechnical Change. Cambridge, MA: MIT Press.
Bijker, W. E./Hughes, Th. P./Pinch, T. (eds) (1987): The Social Construction
of Technological Systems: New Directions in the Sociology and History
of Technology. Cambridge, MA: MIT Press.
Boulding, K. E. (1956): General Systems Theory – Applications for Organi-
zation and Management. Management science, 2(3), 197-208.
Callon, M. (1987): Society in the Making: The Study of Technology as a Tool
For Sociological Analysis, in Bijker, W./Hughes, T./Pinch, T. (eds) The
Social Construction of Technological Systems. Cambridge: MIT Press.
Callon, M. (1991): Techno-Economic Networks and Irreversibility, in Law,
J. (ed) A Sociology of Monsters: Essays on Power, Technology and
Domination. London: Routledge.
From Systems to Actor-Networks
326
Callon, M. (1986): “Elements of a sociology of translation: Domestication
of the Scallops and the Fishermen of St Brieuc Bay”. In John Law (Ed.),
Power, Action and Belief: A New Sociology of Knowledge? London,
Routledge: 196-233.
Callon, M., & Law, J. (1982). On interests and their transformation: enrol-
ment and counter-enrolment. Social Studies of Science, 12(4), 615-625.
Callon, M., Law, J., & Rip, A. (1986). Mapping the dynamics of science and
technology. Sociology of Science, 16(3), 397-441.
Callon, M., & Latour, B. (1987). Unscrewing the big Leviathan: How actors
macro-structure reality and how sociologists help them to do so. In K.
Knorr-Cetina & A. Cicourel (Eds.), Advances in social theory and meth-
odology: Toward an integration of micro-and macro-sociologies (pp.
277-303). Routledge.
Callon, M. (1988). The role of lay people in the production and dissemi-
nation of scientific knowledge. Science, Technology, & Human Values,
13(1), 3-28.
Callon, M., & Law, J. (1995). Agency and the hybrid collectif. South Atlantic
Quarterly, 94(2), 481-507.
Castells, M. (1996): The Rise of the Network Society. The Information Age:
Economy, Society and Culture Vol. I, 2nd ed. 2000. Oxford: Blackwell.
Castells, M (1997a): The Power of Identity. The Information Age: Economy,
Society and Culture Vol. II, 2nd ed. 2004. Oxford: Blackwell.
Castells, M. (1997b): An Introduction to the Information Age, in City 7, 6-16.
Castells, M (1998): End of Millennium. The Information Age: Economy,
Society and Culture Vol. III. 2nd ed. 2000. Oxford: Blackwell.
Castells, M. (2000): Materials for an Exploratory Theory of the Network
Society, in British Journal of Sociology Vol. No. 51 Issue No. 1 (January/
March 2000) pp. 5–24.
Castells, M. (2001): The Internet Galaxy. Oxford: Oxford University Press.
Castells, M. (2004): Informationalism, Networks, and the Network Society:
A Theoretical Blueprint, in Castells, M. (ed) (2004): The Network Society.
A Cross-cultural Perspective. UK: Edward Elgar, 3-45.
Castells, M. (2005): The Network Society: From Knowledge to Policy, in
Castells, M./Cardoso, G. (eds) The Network Society from Knowledge to
Policy. Washington, DC: Center for Transatlantic Relations.
Literature... 327
Castells, M. (2009): Communication Power. New York: Oxford University
Press.
Castells, M. (2011): A Network Theory of Power, in International Journal of
Communication 5, 773-787.
Collis, B. (2005): E-Learning and the Transformation of Education for a
Knowledge Economy, in Castells, M., Cordosa G. 2005, 215-223.
Cronin, L., Walker, S. (2023): Assembly Theory Explains and Quantifies
the Emergence of Selection and Evolution. arXiv:2206.02279v3 [physics.
bio-ph].
Czarniawska, B., Hernes, T. (eds) (2005): Actor-Network Theory and Orga-
nizing. Liber
Czarniawska, B. (1998): Narrative approach in organization studies. UK:
Sage.
Czarniawska, B. (2004): Narratives in Social Science Research. UK: Sage.
Czarniawska, B. (2008): A Theory of Organizing. London. UK: Edward
Elgar.
Deely, J. (1990): Basics of Semiotics, Bloomington: Indiana University Press.
Deplazes A, Huppenbauer M. (2009): Synthetic organisms and living
machines: Positioning the products of synthetic biology at the border-
line between living and non-living matter. Syst Synth Biol. 2009 Dec;3(1-
4):55-63.
d’Espagnat, B. (2006) On Physics and Philosophy. Princeton University
Press. Princeton & Oxford.
DiMaggio, P. J./Powell, W. (eds) (1991): The New Institutionalism in Orga-
nization Analysis. Chicago: University of Chicago Press.
Drucker, P. F. (1999): Management Challenges for the 21st Century. New
York: Haprer.
Ebeling, W. (1989): Chaos – Order – Information. Selbstorganisation in
Natur und Technik, Leipzig: Urania Verlag.
Ehlers, U. D. (2011): From open educational resources to open educational
practices, in Elearning Papers, 23, 1–8. http://openeducationeuropa.eu/
en/paper/open-educationchanging-educational-practices.
Fields, C., Friston, K., Glazebrook, JF., Levin, M. (2022): A free energy prin-
ciple for generic quantum systems 1 Sep 2022, in Progress in Biophysics
and Molecular Biology173:36-59.
From Systems to Actor-Networks
328
Flechtner, H.-J. (1970): Grundbegriffe der Kybernetik. Stuttgart: Wissen-
schaftliche Verlagsgesellschaft.
Floridi, L, (2005): The Ontological Interpretation of Informational Privacy,
in Ethics and Informational Technology, vol. 7, no. 4, pp. 185-220.
Floridi, L. (2014). The Fourth Revolution: How the Infosphere is Reshaping
Human Reality, Oxford: Oxford University Press.
Floridi, L. (2016): On Human Dignity as a Foundation for the Right to
Privacy, in: Philos. Technol. 29, S. 307-312.
Foerster, H. von 1992: Einführung in den Konstruktivismus. Munich: Piper.
Foerster, H. von 1993: Wissen und Gewissen, Frankfurt: Suhrkamp.
Friston, K. (2010): The free-energy principle: a unified brain theory?. Nat
Rev Neurosci 11, 127–138. https://doi.org/10.1038/nrn2787
Friston K, Frith C. D. (2015): Active inference, communication and herme-
neutics. Cortex. 2015 Jul; 68:129-43. doi: 10.1016/j.cortex.2015.03.025.
Gadamer, H.-G. 1989: Truth and Method, 2nd edn. London: Sheed and Ward.
Gamble, C., Poor, M. (eds) (2005): The Hominid Individual in Context.
Archeological Investigations of Lowe and Middle Paleolithic Land-
scape, Locales and Artefacts.
Gerok, W. (1990): Ordnung und Chaos in der unbelebten und belebten
Natur, Stuttgart: Wissenschaftliche Buchgesellschaft.
Gibson, J. (1977): The Theory of Affordances, in Shaw, R./Bransford, J. (eds)
Perceiving, Acting, and Knowing: Toward an Ecological Psychology.
New York: Wiley.
Giddens, A. (1984): The constitution of society: Outline of the theory of
structuration. Cambridge: Polity Press.
Glasersfeld, E. von (1987a): Knowledge, Language and Reality. Works on
radical constructivism, Braunschweig: Vieweg.
Glasersfeld, E. von (1987b): Siegener Gespräche über Radikalen Konstruk-
tivismus, in: Schmidt 1987.
Glasersfeld, E. von (1992): Construction of Reality and the Concept of
Objectivity, in: Introduction to Constructivism, Munich: Piper.
Gumbrecht, H. U., Pfeiffer, K. L. (eds) (1994): Materialities of Communica-
tion, Stanford U. Press: Stanford CA.
Habermas, J. (1971): Theory of Society or Social Technology?, in Habermas/
Luhmann 1971.
Literature... 329
Habermas, J. (1976): Legitimation Crisis, T. McCarthy (trans), Cambridge:
Politiy Press.
Habermas, J. (1981): Theory of Communicative Action. 2 Volumes. McCa-
rthy, Th. (trans). Boston, MA: Beacon Press
Habermas, J. (1989): The Structural Transformation of the Public Sphere: An
Inquiry into a Category of Bourgeois Society, Cambridge: Polity Press.
Habermas, J. (1984): Preliminary Studies and Additions to the Theory of
Communicative Action, Frankfurt: Suhrkamp.
Habermas, J. (1985): The Philosophical Discourse of Modernity, Frankfurt:
Suhrkamp.
Habermas, J., Luhmann, N. (eds) (1971): Theory of Society or Social Tech-
nology. What does systems research achieve? Frankfurt: Suhrkamp.
Haferkamp, H., Schmid, M. (eds) (1987): Meaning, Communication and
Social Differentiation. Beiträge zu Luhmanns Theorie sozialer Systeme,
Frankfurt: Suhrkamp.
Haken, H., Haken-Krell, M. (1989): Emergence of Biological Information
and Order, Darmstadt: Wissenschaftliche Buchgesellschaft.
Haken, H., Wunderlin, A. (1991): The Self-Structuring of Matter. Synerc-
getics in the Inanimate World, Braunschweig: Vieweg.
Heidegger, M. [1927] (1962): Being and Time. Macquarie, J. (trans). Oxford:
Blackwell.
Heidegger, M. (1977): The Question Concerning Technology, in Basic Writ-
ings Krell, D. F. (ed), New York: Harper & Row, 287.
Hidalgo, C. A. (2016): Disconnected, Fragmented, or United? A Trans-Dis-
ciplanary Review of Network Science. Appl Netw Sci 1, 6.
Hoffman, D. (2019): The Case Against Reality: How Evolution Hid the
Truth from Our Eyes, Allen Lane.
Hutchins, E. (1995): Cognition in the Wild. Cambridge, MA: MIT Press.
Jantsch, E. (1980): The Self-Organizing Universe. Pergamon Press Oxford/
New York.
Jantsch, E. (1987): Epistemological Aspects of the Self-Organization of
Natural Systems, in Schmidt 1987.
Jarrillo, J. C. (1993): Strategic Networks: The Borderless Organisation.
Oxford: Butterworth-Heinemann.
Jones, Ch. (2015): Networked Learning: An Educational Paradigm for the
From Systems to Actor-Networks
330
Age of Digital Networks. Switzerland: Springer.
Kaduk, S., Osmetz, D., Wüthrich, H. A., Hammer, D. (2015): Müsterbrecher.
Die Kunst, das Spiel zu drehen. Hamburg: Murmann.
Katz, J., Gartner, W. B. (1988): Properties of Emerging Organizations, in
Academy of Management Review, 13(3), 429-441.
Katz, J. A., Shepherd, D. A. (2004): Corporate Entrepreneurship. New York:
Elsevier.
Katz M. L., Shapiro C. (1985): Network Externalities, Competition, and
Compatibility, in American Economic Review 75: 424-440.
Katz, M. L., Shapiro, C. (1994): Systems Competition and Network Effects,
in Journal of Economic Perspectives 8: 93-115.
Kalpokas, I. (2919): Algorithmic Governance. Politics and Law, in the Post-
Human Era. Palgrave Macmillan: London.
Kast, F. E., Rosenzweig, J. E. (1972): General systems theory: Applications
for organization and management, in Academy of management journal,
15(4), 447-465.
Kiani, N., Gomez-Cabrero, D., Bianconi, G. (eds) (2021): Networks of
Networks, in Biology: Concepts, Tools and Applications. Cambridge:
Cambridge University Press.
Kickert, W., Klijn, E.-H., Koopenjan, J. (1997): Managing Complex Networks.
London: Sage.
Klaus, H., Liebscher, G. (1979): Wörterbuch der Kybernetik. Fischer Verlag.
Kneer, G. (1992): Conservation and Reflection. On the critical reformulation
of social rationality, in Krawietz / Welker 1992.
Kneer, G., Nassehi, A. (eds) (1993): Niklas Luhmanns Theorie sozialer
Systeme. Eine Einführung, München: Fink.
Kratky, K. W., Wallner, F. (eds) (1990): Grundprinzipien der Selstorganisa-
tion, Darmstadt: Wissenschaftliche Buchgesellschaft.
Krawietz, W., Welker, M. (eds) (1992): Kritik der Theorie sozialer Systeme.
Auseinandersetzungen mit Luhmanns Hauptwerk, Frankfurt:
Suhrkamp.
Krieger, D. J., Belliger, A. (2014): Interpreting Networks – Hermeneutics,
Actor-Networ Theory and New Media. Bielefeld: transcript Verlag.
Krohn, W., Küppers, G. (eds) (1989): Die Selbstorganisation der Wissen-
schaft. Frankfurt: Suhrkamp.
Literature... 331
Krohn, W., Küppers, G. (1992): Emergenz. Die Entstehung von Ordnung,
Organisation und Bedeutung, Frankfurt: Suhrkamp.
Latour, B. (1987): Science in Action: How to Follow Scientists and Engineers
Through Society. Cambridge, MA: Harvard University Press.
Latour, B. (1992): Where Are the Missing Masses? The Sociology of a Few
Mundane Artifacts, in Bijker, W. E., Law, J. (eds) Shaping Technology/
Building Society: Studies in Sociotechnical Change. Cambridge, MA:
MIT Press.
Latour, B. (1993a): We Have Never Been Modern. Cambridge, MA: Harvard
University Press.
Latour, B. (1993b): Pasteurization of France. Cambridge, MA: Harvard U.
Press.
Latour, B. (1994): On Technical Mediation, in Common Knowledge Vol.3,
n2: 29-64.
Latour, B. (1996a): On Interobjectivity, Mind, Culture, and Activity. Vol. 3,
No. 4: 228-245.
Latour, B. (1996b): Review of Ed Hutchins’ Cognition in the Wild, in Mind,
Culture, and Activity. Vol. 3, No. 1: 54-63.
Latour, B. (1999): Pandora’s Hope: Essays on the Reality of Science Studies,
Cambridge MA: Harvard University Press.
Latour, B. (2002): War of the Worlds: What about Pace? Chicago: Prickly
Paradigm Press.
Latour, B. (2004a): Why Has Critique Run out of Steam? From Matters
of Fact ot Matters of Concern, in Critical Inquiry 30 (Winter): 225-248.
Chicago: University of Chicago Press.
Latour, B. (2004b). Politics of Nature. How to Bring the Sciences into
Democracy. Cambridge MA: Harvard University Press.
Latour, B. (2005). Reassembling the social: An introduction to actor-net-
work theory. Oxford University Press.
Latour, B. (2008): A Cautious Prometheus? A Few Steps Toward a Philos-
ophy of Design (with Special Attention to Peter Sloterdijk), in Proceed-
ings of the 2008 Annual International Conference of the Design History
Society, (ed) F. Hackne, J. Glynne, V. Minto, 2-10. Falmouth: Universal
Publications.
Latour, B. (2010): On the Modern Cult of the Factish Gods. Durham NC:
From Systems to Actor-Networks
332
Duke University Press.
Latour, B. (2012): The Whole is Always Smaller Than Its Parts. A Digital
Test of Gabriel Tarde’s Monads, in British Journal of Sociology 63(4):
591-615.
Latour, B. (2013a): What’s the story? Organizing as a Mode of Existence, in
Organization and Organizing: Materiality, Agency, and Discourse, (ed)
D. Robichaud and F. Cooren. New York: Routledge.
Latour, B. (2013): An Enquiry into Modes of Existence. An Anthropology of
the Moderns. Cambridge, MA: Harvard University Press.
Latour, B. (2014): What is the Style of Matters of Concern, in (ed) Nicholas
Gaskill & A.J. Nocek The Lure of Whitehead. University of Minnesota
Press, pp. 92-127.
Latour, B. (2017): Facing Gaia. Eight Lectures on the New Climatic Regime.
Cambridge, MA: Polity Press.
Latour, B., Woolgar, S. (1979): Laboratory Life. The Construction of Scien-
tific Facts. Beverly Hills, CA: Sage Publications.
László, E. (1974): Preface, in Perspectives on General System Theory, by L.
von Bertalanffy, (ed) E. Taschdjian. New York: George Braziller.
Lavie, D. (2008): The Competitive Advantage of Interconnected Firms, in
Wankel (2008), Vol 1, 324-333.
Law, J. (1992): Notes on the theory of the actor-network: Ordering, strategy,
and heterogeneity. Systems Practice, 5(4), 379-393.
Law, J. (1999): After ANT: Complexity, naming and topology. In J. Law &
J. Hassard (Eds.), Actor network theory and after (pp. 1-14). Blackwell.
Law, J. (1987): Technology and heterogeneous engineering: The case of
Portuguese expansion. In W. Bijker, T. Hughes, & T. Pinch (Eds.), The
social construction of technological systems: New directions in the
sociology and history of technology (pp. 111-134). MIT Press.
Law, J. (2019): Material Semiotics. heterogeneities.net/publications/
Law2019MaterialSemiotics.pdf
Levin, M. (2022): Technological Approach to Mind Everywhere: An
Experimentally-Grounded Framework for Understanding Diverse
Bodies and Minds, in Front. Syst. Neurosci. 16:768201. doi: 10.3389/
fnsys.2022.768201
Lewis, T. G. (2009): Network Science. Theory and Applications. New York:
Literature... 333
Wiley & Sons.
Luhmann, N. (1971): Sinn als Grundbegriff der Soziologie, in Habermas/
Luhmann 1971.
Luhmann, N. (1975): Macht. Fink: Stuttgart.
Luhmann, N. (1977): Interpenetration – Zum Verhältnis personal und
sozialer und sozialer Systeme, in Zeitschrift für Soziologie, Jg. 6. Heft
1. Jan.: 62-76.
Luhmann, N. (1982): The World Society as a Social System, in International
Journal of General Systems 8(3): 131–138.
Luhmann, N. (1988): Wie ist Bewusstsein an Kommunikation beteiligt?, in
Gumbrecht/Pfeiffer 1988.
Luhmann, N. 1988: Power, 2nd ed, Stuttgart: Franke.
Luhmann, N. (1989): Ecological Communication. Chicago: University of
Chicago Press.
Luhmann, N. 1990): Essays on Self-Reference. New York: Colombia Univer-
sity Press.
Luhmann, N. (1992): What is Communication?, in Communication Theory,
2(3): 251-259.
Luhmann, N. (1993): Zeichen als Form, in Baecker 1993.
Luhmann, N. (1995): Social Systems. Stanford, CA: Stanford University
Press.
Luhmann, N. (2008): Die Moral der Gesellschaft. Frankfurt a.M.: Suhrkamp.
Luhmann, N. (2012): Theory of Society V. 1. Stanford: Stanford University
Press.
Luhmann, N. (2013): Theory of Society V. 2. Stanford: Stanford University
Press.
Lupton, D. (2014a): Self-tracking Cultures: Towards a Sociology of
Personal Informatics. Paper delivered at the OzHCI Conference 2014.
(https://simplysociology.files.wordpress.com/2014/09/self-tracking-cul-
tures-ozchi-conference-paper.pdf).
Lupton, D. (2014b): Self-Tracking Modes: Reflexive Self-Monitoring and
Data Practices (August 19, 2014). Available at SSRN: http://ssrn.com/
abstract=2483549 or http://dx.doi.org/10.2139/ssrn.2483549.
Lupton, D. (2014c): Apps as Artefacts: Towards a Critical Perspective on
Mobile Health and Medical Apps. Societies 4, 606-622.
From Systems to Actor-Networks
334
Lyotard, J.-F. (1984): The Postmodern Condition. A Report on Knowledge.
University of Minnesota Press.
Malapi-Nelson, A. (2017): The Nature of the Machine and the Collapse
of Cybernetics: A Transhumanist Lesson for Emerging Technologies.
Palgrave Macmillan.
Mann, S. F., Pain, R,. Kirchhoff, M. D. (2022): Free Energy: A User’s Guide.
Biology & Philosophy. 37:33.
Marc, E., Picard, D. 1991: Bateson, Watzlawick und die Schule von Palo
Alto. Frankfurt a.M.: Athenaeum Program.
Maturana, H. R. (1987): Cognition, in: Schmidt 1987.
Maturana, H. R., Varela, F. J. (1980): Autopoiesis and Cognition: The Reali-
zation of the Living. Boston Studies, in the Philosophy of Science.
Maturana, H. R., Varela, F. J (1987): The tree of Knowledge. Biological Basis
of Human Understanding. Boston: Shambala.
Malapi-Nelson, A. (2017): The Nature of the Machine and the Collapse of
Cybernetics, Palgrave Macmillan Cham.
Manicka, S., Levin, M. (2022): Minimal Developmental Computation:
A Causal Network Approach to Understand Morphogenetic Pattern
Formation, in Entropy 2022, 24, 107. https://doi.org/10.3390/e24010107
Miller, J. G., Miller J. L. (1990): The Nature of Life, in Systems Research and
Behavioral Science 35(3):157-163.
Moltmann, J. (1985): Gott in der Schöpfung. Ökologische Schöpfungslehre,
München: Kaiser.
Muller, A., Shipton, C., Clarkson, C. (2022): Stone Toolmaking Difficulty
and the Evolution of Hominin Technological Skills, in Sci Rep 12, 5883.
https://doi.org/10.1038/s41598-022-09914-2.
Mujis, D., Ainscow, M., Chapman, C., West, M. (2011): Collaboration and
Networking, in Education. Dordrecht: Springer.
Nassehi, A. 1992: Wie wirklich sind Systeme? Zum ontologischen und
epistemologischen Status von Luhmanns Theorie selbstreferentieller
Systeme, in: Krawietz/Welker 1992.
OECD (2007): Giving Knowledge for Free: The Emergence of Open Educa-
tional Resources. Center for Educational Research and Innovation.
http://www.oecd.org/edu/ceri/38654317.pdf
Open Educational Resources (2013): The William and Flora Hewlett Foun-
Literature... 335
dation. http://www.hewlett.org/programs/education/open-education-
al-resources.
Orban, G. A., Caruana, F. (2014): The Neural Basis of Human Tool-Use, in
Front. Psychol., 09 April, Sec. Cognition, Volume 5 – 2014. https://doi.
org/10.3389/fpsyg.2014.00310.
Parker, G., Choudary, S. P., Van Alstyne, M.W. (2016): Platform Revolution:
How Networked Markets are Transforming the Economy – And How to
Make them Work For You. New York: Norton.
Peirce, C. S. 1983: Phenomenon and logic of signs. Frankfurt a.M.: Suhrkamp.
Podolny, J. M./Page, K. L. (1998): “Network Forms of Organization,” in:
Annual Review of Sociology, 24:57-76.
Peirce, Ch. S. (1991): Peirce on Signs: Writings on Semiotic by Charles
Sanders Peirce. J. Hoopes (ed). University of North Carolina Press.
Preda, A. (1999): The Turn to Things: Arguments for a Sociological Theory
of Things, in The Sociological Quarterly Vol. 40, No. 2 (Spring, 1999),
pp. 347-366.
Principia Cybernetica Website: http://pespmc1.vub.ac.be/CYBSHIST.html.
Riegas, V., Vetter, Ch. (eds) (1990): Zur Biologie der Kognition. Ein Gespräch
mit Humberto R. Maturana und Beiträge zur Diskussion seines Werkes.
Frankfurt: Suhrkamp.
Roth, G. 1987: Autopoiese und Kognition: Die Theorie H. R. Maturanas und
die Notwendigkeit ihrer Weiterentwicklung, in Schmidt 1987.
Rowlands, M. J. (2010): The New Science of the Mind: From Extended Mind
to Embodied Phenomenology. Cambridge MA: MIT Press.
Saz-Carranza, A., Ospina, S., Vernis, A. (2008): Leadership in Interorganiza-
tional Networks, in Wankel (2008), Vol. 2, 291-300.
Schmidt, S. J. (ed) (1987): Der Diskurs des Radikalen Konstruktivismus.
Frankfurt: Suhrkamp.
Sebeok, Th. (1991): A Sign is Just a Sign, Bloomington: Indiana University Press.
Sebeok, Th. (1994): Signs. An Introduction to Semiotics. Toronto: Univer-
sity of Toronto Press.
Schiewek, W. (1992): On the neglected connection between ‘symbolic
generalization’ and ‘language’ in social systems theory, in: Krawietz /
Welker 1992.
Schmidt, S. J. (ed) (1987): Der Diskurs des radikalen Konstruktivismus.
From Systems to Actor-Networks
336
Frankfurt: Suhrkamp.
Schmidt, S. J. (1994): Kognitive Autonomie und soziale Orientierung.
Konstruktivitische Bermerkungen zum Zusammenhang von Kognition,
Kommunikation, Medien und Kultur. Frankfurt a.M.: Suhrkamp.
Schuilenburg, M., Peeters, R. (eds) (2020): The Algorithmic Society. Tech-
nology, Power, and Knowledge. Routledge: London.
Shannon, C., Weaver, W. (1949): The Mathematical Theory of Communica-
tion, Urbanna: University of Illinois Press.
Shirky, C. (2008): Here Comes Everybody. How Change Happens When
People Come Together. New York: Penguin Press.
Schneider, D. (2014): Personal Learning Environment. Edutech Wiki http://
edutechwiki.unige.ch/en/Personal_learning_environment (retrieved
Jan. 2016)
Siemens, G. (2005): Connectivism: A Learning Theory for the Digital Age,
in International Journal of Instructional Technology and Distance
Learning. Vol. 12. No. 1 Jan.
Simon, H. (2019): The Sciences of the Artificial. MIT Press.
Spencer Brown, G. (1972): Laws of Form. New York: Julian Press.
Stadler, M., Kruse, P. (1992): Zur Emergenz psychischer Qualitäten. Das
psychophysische Problem im Lichte der Selbstorganisationstheorie, in
Krohn/Küppers 1992.
Steffen, W., Richardson, K., Rockström, J. et al. (2020): The Emergence and
Evolution of Earth System Science. in Nat Rev Earth Environ 1, 54–63.
Steier, F. (ed) (1991): Research and Reflexivity, London: Sage.
Surowiecki, J. (2004): The Wisdom of Crowds: Why the Many Are Smarter
Than the Few and How Collective Wisdom Shapes Business, Econo-
mies, Societies and Nations. New York: Randomhouse.
Thomas, G. (1992): World as Relative Unity and as Last Horizon, in: Kraw-
ietz/Welker 1992.
Thorelli, H. B. (1986): Networks: Between Markets and Hierarchies, in Stra-
tegic Management Journal, V. 7, 1 (Jan.-Feb.), 37-51.
Varela, F. J., Maturana, H. R., Uribe, R. (1974): Autopoiesis: The Organiza-
tion of Living Systems, its Characterization and a Model, in BioSystems
5, 187-196.
Wankel, C. (ed) (2008): 21st Century Management: A Reference Handbook.
Literature... 337
2 Vols. UK: Sage Publications.
Watzlawick, P. (1977): How Real is Real?. New York: Vintage.
Weick, K. E. (1979): The Social Psychology of Organizing (2nd ed). New
York: McGraw-Hill.
Weick, K. E. (1995): Sensemaking in Organizations. Foundations for Orga-
nizational Science. London: Sage Publications.
Welker, M. (ed.). (1985): Theologie und funktionale Systemtheorie.
Luhmanns Religionssoziologie in theologischer Diskussion. Frankfurt:
Suhrkamp.
Wiener, N. (1948): Cybernetics: or Control and Communication in the
Animal and the Machine, New York: John Wiley & Sons.
Willke, H. (2004): Systemtheorie I. Eine Einführung in die Grundprobleme.
7th Edition, Stuttgart (UTB)
Willke, H. (1999): Systemtheorie II: Interventionstheorie. 3rd Edition, Stutt-
gart (UTB)
Willke, H. (1995): Systemtheorie III: Steuerungstheorie. 3rd Edition, Stutt-
gart (UTB)
Williams, I. (2020): Contemporary Applications of Actor Network Theory.
Palgrave Macmillan: Singapore.
Wittgenstein, L. (1922): Tractatus logico-philosophicus. Kegan Paul:
London.
Wittgenstein, L. (1953): Philosophical Investigations. Blackwell Publishing:
UK.
Wolfram S. (2002); A New Kind of Science, Wolfram Media; and online
https://www.wolframscience.com/nks/.
Wüthrich, H. A., Osmetz, D., Kaduk, S. (2009): Musterbrecher – Führung
neu leben. Wiesbaden: Gabler.
